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

Diff of /code/trunk/ChangeLog

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

revision 214 by ph10, Wed Aug 15 14:08:10 2007 UTC revision 1566 by ph10, Tue Jun 9 17:45:25 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 09-Aug-07  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 8.38 xx-xxx-xxxx
8    ------------------------
9    
10    1.  If a group that contained a recursive back reference also contained a
11        forward reference subroutine call followed by a non-forward-reference
12        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13        compile correct code, leading to undefined behaviour or an internally
14        detected error. This bug was discovered by the LLVM fuzzer.
15    
16    2.  Quantification of certain items (e.g. atomic back references) could cause
17        incorrect code to be compiled when recursive forward references were
18        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19        This bug was discovered by the LLVM fuzzer.
20    
21    3.  A repeated conditional group whose condition was a reference by name caused
22        a buffer overflow if there was more than one group with the given name.
23        This bug was discovered by the LLVM fuzzer.
24    
25    4.  A recursive back reference by name within a group that had the same name as
26        another group caused a buffer overflow. For example:
27        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28    
29    5.  A forward reference by name to a group whose number is the same as the
30        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31        a buffer overflow at compile time. This bug was discovered by the LLVM
32        fuzzer.
33    
34    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
35        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36    
37    7.  Another buffer overflow bug involved duplicate named groups with a
38        reference between their definition, with a group that reset capture
39        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40        fixed by always allowing for more memory, even if not needed. (A proper fix
41        is implemented in PCRE2, but it involves more refactoring.)
42    
43    8.  There was no check for integer overflow in subroutine calls such as (?123).
44    
45    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
46        being treated as a literal 'l' instead of causing an error.
47    
48    10. There was a buffer overflow if pcre_exec() was called with an ovector of
49        size 1. This bug was found by american fuzzy lop.
50    
51    11. If a non-capturing group containing a conditional group that could match
52        an empty string was repeated, it was not identified as matching an empty
53        string itself. For example: /^(?:(?(1)x|)+)+$()/.
54    
55    
56    Version 8.37 28-April-2015
57    --------------------------
58    
59    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
60        for those parentheses to be closed with whatever has been captured so far.
61        However, it was failing to mark any other groups between the hightest
62        capture so far and the currrent group as "unset". Thus, the ovector for
63        those groups contained whatever was previously there. An example is the
64        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
65    
66    2.  If an assertion condition was quantified with a minimum of zero (an odd
67        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
68    
69    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
70        unrecognized modifier, a crash could occur.
71    
72    4.  An attempt to do global matching in pcretest with a zero-length ovector
73        caused a crash.
74    
75    5.  Fixed a memory leak during matching that could occur for a subpattern
76        subroutine call (recursive or otherwise) if the number of captured groups
77        that had to be saved was greater than ten.
78    
79    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
80        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
81        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
82    
83    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
84        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
85    
86    8. When a pattern is compiled, it remembers the highest back reference so that
87       when matching, if the ovector is too small, extra memory can be obtained to
88       use instead. A conditional subpattern whose condition is a check on a
89       capture having happened, such as, for example in the pattern
90       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
91       setting the highest backreference number. This mattered only if pcre_exec()
92       was called with an ovector that was too small to hold the capture, and there
93       was no other kind of back reference (a situation which is probably quite
94       rare). The effect of the bug was that the condition was always treated as
95       FALSE when the capture could not be consulted, leading to a incorrect
96       behaviour by pcre_exec(). This bug has been fixed.
97    
98    9. A reference to a duplicated named group (either a back reference or a test
99       for being set in a conditional) that occurred in a part of the pattern where
100       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
101       to be incorrectly calculated, leading to overwriting.
102    
103    10. A mutually recursive set of back references such as (\2)(\1) caused a
104        segfault at study time (while trying to find the minimum matching length).
105        The infinite loop is now broken (with the minimum length unset, that is,
106        zero).
107    
108    11. If an assertion that was used as a condition was quantified with a minimum
109        of zero, matching went wrong. In particular, if the whole group had
110        unlimited repetition and could match an empty string, a segfault was
111        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
112        assertions to be quantified, but not if they are being used as conditions,
113        so the above pattern is faulted by Perl. PCRE has now been changed so that
114        it also rejects such patterns.
115    
116    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
117        failed to allow the zero-repeat case if pcre2_exec() was called with an
118        ovector too small to capture the group.
119    
120    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
121        Red Hat Product Security:
122    
123        (a) A crash if /K and /F were both set with the option to save the compiled
124        pattern.
125    
126        (b) Another crash if the option to print captured substrings in a callout
127        was combined with setting a null ovector, for example \O\C+ as a subject
128        string.
129    
130    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
131        forward reference repeated a large (but limited) number of times within a
132        repeated outer group that has a zero minimum quantifier, caused incorrect
133        code to be compiled, leading to the error "internal error:
134        previously-checked referenced subpattern not found" when an incorrect
135        memory address was read. This bug was reported as "heap overflow",
136        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
137        CVE-2015-2325.
138    
139    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
140        call within a group that also contained a recursive back reference caused
141        incorrect code to be compiled. This bug was reported as "heap overflow",
142        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
143        number CVE-2015-2326.
144    
145    24. Computing the size of the JIT read-only data in advance has been a source
146        of various issues, and new ones are still appear unfortunately. To fix
147        existing and future issues, size computation is eliminated from the code,
148        and replaced by on-demand memory allocation.
149    
150    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
151        adjacent to the end of the range, and the range contained characters with
152        more than one other case, caused incorrect behaviour when compiled in UTF
153        mode. In that example, the range a-j was left out of the class.
154    
155    26. Fix JIT compilation of conditional blocks, which assertion
156        is converted to (*FAIL). E.g: /(?(?!))/.
157    
158    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
159        discovered by the LLVM fuzzer.
160    
161    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
162        when this assertion was used as a condition, for example (?(?!)a|b). In
163        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
164        error about an unsupported item.
165    
166    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
167        possessification code could take exponential time to complete. A recursion
168        depth limit of 1000 has been imposed to limit the resources used by this
169        optimization.
170    
171    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
172        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
173        because \S ensures they are all in the class. The code for doing this was
174        interacting badly with the code for computing the amount of space needed to
175        compile the pattern, leading to a buffer overflow. This bug was discovered
176        by the LLVM fuzzer.
177    
178    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
179        other kinds of group caused stack overflow at compile time. This bug was
180        discovered by the LLVM fuzzer.
181    
182    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
183        between a subroutine call and its quantifier was incorrectly compiled,
184        leading to buffer overflow or other errors. This bug was discovered by the
185        LLVM fuzzer.
186    
187    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
188        assertion after (?(. The code was failing to check the character after
189        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
190        was discovered by the LLVM fuzzer.
191    
192    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
193        a fixed maximum following a group that contains a subroutine reference was
194        incorrectly compiled and could trigger buffer overflow. This bug was
195        discovered by the LLVM fuzzer.
196    
197    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
198        caused a stack overflow instead of the diagnosis of a non-fixed length
199        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
200    
201    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
202        (e.g. /(?<=\Ka)/) could make pcregrep loop.
203    
204    37. There was a similar problem to 36 in pcretest for global matches.
205    
206    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
207        and a subsequent item in the pattern caused a non-match, backtracking over
208        the repeated \X did not stop, but carried on past the start of the subject,
209        causing reference to random memory and/or a segfault. There were also some
210        other cases where backtracking after \C could crash. This set of bugs was
211        discovered by the LLVM fuzzer.
212    
213    39. The function for finding the minimum length of a matching string could take
214        a very long time if mutual recursion was present many times in a pattern,
215        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
216        method has been implemented. This infelicity was discovered by the LLVM
217        fuzzer.
218    
219    40. Static linking against the PCRE library using the pkg-config module was
220        failing on missing pthread symbols.
221    
222    
223    Version 8.36 26-September-2014
224    ------------------------------
225    
226    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
227        -Wmissing-field-initializers and -Wunused-parameter.
228    
229    2.  The tests for quantifiers being too big (greater than 65535) were being
230        applied after reading the number, and stupidly assuming that integer
231        overflow would give a negative number. The tests are now applied as the
232        numbers are read.
233    
234    3.  Tidy code in pcre_exec.c where two branches that used to be different are
235        now the same.
236    
237    4.  The JIT compiler did not generate match limit checks for certain
238        bracketed expressions with quantifiers. This may lead to exponential
239        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
240        issue should be resolved now.
241    
242    5.  Fixed an issue, which occures when nested alternatives are optimized
243        with table jumps.
244    
245    6.  Inserted two casts and changed some ints to size_t in the light of some
246        reported 64-bit compiler warnings (Bugzilla 1477).
247    
248    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
249        an empty string, which sometimes were behaving incorrectly in the
250        interpreter (though correctly in the JIT matcher). This pcretest input is
251        an example:
252    
253          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
254          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
255    
256        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
257        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
258        for an empty string was breaking the inner loop and carrying on at a lower
259        level, when possessive repeated groups should always return to a higher
260        level as they have no backtrack points in them. The empty string test now
261        occurs at the outer level.
262    
263    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
264        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
265    
266    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
267        doesn't).
268    
269    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
270        to hold a pattern to be incorrectly computed (too small) when there were
271        named back references to duplicated names. This could cause "internal
272        error: code overflow" or "double free or corruption" or other memory
273        handling errors.
274    
275    11. When named subpatterns had the same prefixes, back references could be
276        confused. For example, in this pattern:
277    
278          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
279    
280        the reference to 'Name' was incorrectly treated as a reference to a
281        duplicate name.
282    
283    12. A pattern such as /^s?c/mi8 where the optional character has more than
284        one "other case" was incorrectly compiled such that it would only try to
285        match starting at "c".
286    
287    13. When a pattern starting with \s was studied, VT was not included in the
288        list of possible starting characters; this should have been part of the
289        8.34/18 patch.
290    
291    14. If a character class started [\Qx]... where x is any character, the class
292        was incorrectly terminated at the ].
293    
294    15. If a pattern that started with a caseless match for a character with more
295        than one "other case" was studied, PCRE did not set up the starting code
296        unit bit map for the list of possible characters. Now it does. This is an
297        optimization improvement, not a bug fix.
298    
299    16. The Unicode data tables have been updated to Unicode 7.0.0.
300    
301    17. Fixed a number of memory leaks in pcregrep.
302    
303    18. Avoid a compiler warning (from some compilers) for a function call with
304        a cast that removes "const" from an lvalue by using an intermediate
305        variable (to which the compiler does not object).
306    
307    19. Incorrect code was compiled if a group that contained an internal recursive
308        back reference was optional (had quantifier with a minimum of zero). This
309        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
310        caused segmentation faults because of stack overflows at compile time.
311    
312    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
313        group that is quantified with an indefinite repeat, caused a compile-time
314        loop which used up all the system stack and provoked a segmentation fault.
315        This was not the same bug as 19 above.
316    
317    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
318        Patch by Mike Frysinger.
319    
320    
321    Version 8.35 04-April-2014
322    --------------------------
323    
324    1.  A new flag is set, when property checks are present in an XCLASS.
325        When this flag is not set, PCRE can perform certain optimizations
326        such as studying these XCLASS-es.
327    
328    2.  The auto-possessification of character sets were improved: a normal
329        and an extended character set can be compared now. Furthermore
330        the JIT compiler optimizes more character set checks.
331    
332    3.  Got rid of some compiler warnings for potentially uninitialized variables
333        that show up only when compiled with -O2.
334    
335    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
336        of a match later then the end of the match. The pcretest program was not
337        handling the case sensibly - it was outputting from the start to the next
338        binary zero. It now reports this situation in a message, and outputs the
339        text from the end to the start.
340    
341    5.  Fast forward search is improved in JIT. Instead of the first three
342        characters, any three characters with fixed position can be searched.
343        Search order: first, last, middle.
344    
345    6.  Improve character range checks in JIT. Characters are read by an inprecise
346        function now, which returns with an unknown value if the character code is
347        above a certain threshold (e.g: 256). The only limitation is that the value
348        must be bigger than the threshold as well. This function is useful when
349        the characters above the threshold are handled in the same way.
350    
351    7.  The macros whose names start with RAWUCHAR are placeholders for a future
352        mode in which only the bottom 21 bits of 32-bit data items are used. To
353        make this more memorable for those maintaining the code, the names have
354        been changed to start with UCHAR21, and an extensive comment has been added
355        to their definition.
356    
357    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
358        to the export list in Makefile.am (they were accidentally omitted from the
359        8.34 tarball).
360    
361    9.  The informational output from pcretest used the phrase "starting byte set"
362        which is inappropriate for the 16-bit and 32-bit libraries. As the output
363        for "first char" and "need char" really means "non-UTF-char", I've changed
364        "byte" to "char", and slightly reworded the output. The documentation about
365        these values has also been (I hope) clarified.
366    
367    10. Another JIT related optimization: use table jumps for selecting the correct
368        backtracking path, when more than four alternatives are present inside a
369        bracket.
370    
371    11. Empty match is not possible, when the minimum length is greater than zero,
372        and there is no \K in the pattern. JIT should avoid empty match checks in
373        such cases.
374    
375    12. In a caseless character class with UCP support, when a character with more
376        than one alternative case was not the first character of a range, not all
377        the alternative cases were added to the class. For example, s and \x{17f}
378        are both alternative cases for S: the class [RST] was handled correctly,
379        but [R-T] was not.
380    
381    13. The configure.ac file always checked for pthread support when JIT was
382        enabled. This is not used in Windows, so I have put this test inside a
383        check for the presence of windows.h (which was already tested for).
384    
385    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
386        The algorithm provides a way to skip certain starting offsets, and usually
387        faster than linear prefix searches.
388    
389    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
390        as for 'fr_FR' and 'french'. For some reason, however, it then used the
391        Windows-specific input and output files, which have 'french' screwed in.
392        So this could never have worked. One of the problems with locales is that
393        they aren't always the same. I have now updated RunTest so that it checks
394        the output of the locale test (test 3) against three different output
395        files, and it allows the test to pass if any one of them matches. With luck
396        this should make the test pass on some versions of Solaris where it was
397        failing. Because of the uncertainty, the script did not used to stop if
398        test 3 failed; it now does. If further versions of a French locale ever
399        come to light, they can now easily be added.
400    
401    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
402        there was a message during ./configure, but it did not stop. This now
403        provokes an error. The invalid example in README has been corrected.
404        If a value less than the minimum is given, the minimum value has always
405        been used, but now a warning is given.
406    
407    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
408        was a bug in the test system, which is now fixed. Also, the list of various
409        configurations that are tested for each release did not have one with both
410        16/32 bits and --enable-bar-anycrlf. It now does.
411    
412    18. pcretest was missing "-C bsr" for displaying the \R default setting.
413    
414    19. Little endian PowerPC systems are supported now by the JIT compiler.
415    
416    20. The fast forward newline mechanism could enter to an infinite loop on
417        certain invalid UTF-8 input. Although we don't support these cases
418        this issue can be fixed by a performance optimization.
419    
420    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
421        not take account if existing stack usage. There is now a new global
422        variable called pcre_stack_guard that can be set to point to an external
423        function to check stack availability. It is called at the start of
424        processing every parenthesized group.
425    
426    22. A typo in the code meant that in ungreedy mode the max/min qualifier
427        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
428        not match "ab".
429    
430    23. When UTF was disabled, the JIT program reported some incorrect compile
431        errors. These messages are silenced now.
432    
433    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
434        compiler.
435    
436    25. Change all the temporary files used in RunGrepTest to be different to those
437        used by RunTest so that the tests can be run simultaneously, for example by
438        "make -j check".
439    
440    
441    Version 8.34 15-December-2013
442    -----------------------------
443    
444    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
445        executable memory. Patch inspired by Carsten Klein.
446    
447    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
448        this macro is never tested and has no effect, because the work to support
449        coverage involves only compiling and linking options and special targets in
450        the Makefile. The comment in config.h implied that defining the macro would
451        enable coverage support, which is totally false. There was also support for
452        setting this macro in the CMake files (my fault, I just copied it from
453        configure). SUPPORT_GCOV has now been removed.
454    
455    3.  Make a small performance improvement in strlen16() and strlen32() in
456        pcretest.
457    
458    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
459        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
460        these cases). There was also one in pcretest.c.
461    
462    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
463    
464    6.  In UTF mode, the code for checking whether a group could match an empty
465        string (which is used for indefinitely repeated groups to allow for
466        breaking an infinite loop) was broken when the group contained a repeated
467        negated single-character class with a character that occupied more than one
468        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
469        UTF-8 mode). The effect was undefined: the group might or might not be
470        deemed as matching an empty string, or the program might have crashed.
471    
472    7.  The code for checking whether a group could match an empty string was not
473        recognizing that \h, \H, \v, \V, and \R must match a character.
474    
475    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
476        an empty string. If it can, pcretest shows this in its information output.
477    
478    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
479        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
480        matched by pcre_exec() without using JIT:
481    
482        (a) If the rest of the pattern did not match after a maximal run of
483            grapheme clusters, the code for backing up to try with fewer of them
484            did not always back up over a full grapheme when characters that do not
485            have the modifier quality were involved, e.g. Hangul syllables.
486    
487        (b) If the match point in a subject started with modifier character, and
488            there was no match, the code could incorrectly back up beyond the match
489            point, and potentially beyond the first character in the subject,
490            leading to a segfault or an incorrect match result.
491    
492    10. A conditional group with an assertion condition could lead to PCRE
493        recording an incorrect first data item for a match if no other first data
494        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
495        first data item, and therefore matched "ca" after "c" instead of at the
496        start.
497    
498    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
499        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
500    
501    12. The source of pcregrep now includes z/OS-specific code so that it can be
502        compiled for z/OS as part of the special z/OS distribution.
503    
504    13. Added the -T and -TM options to pcretest.
505    
506    14. The code in pcre_compile.c for creating the table of named capturing groups
507        has been refactored. Instead of creating the table dynamically during the
508        actual compiling pass, the information is remembered during the pre-compile
509        pass (on the stack unless there are more than 20 named groups, in which
510        case malloc() is used) and the whole table is created before the actual
511        compile happens. This has simplified the code (it is now nearly 150 lines
512        shorter) and prepared the way for better handling of references to groups
513        with duplicate names.
514    
515    15. A back reference to a named subpattern when there is more than one of the
516        same name now checks them in the order in which they appear in the pattern.
517        The first one that is set is used for the reference. Previously only the
518        first one was inspected. This change makes PCRE more compatible with Perl.
519    
520    16. Unicode character properties were updated from Unicode 6.3.0.
521    
522    17. The compile-time code for auto-possessification has been refactored, based
523        on a patch by Zoltan Herczeg. It now happens after instead of during
524        compilation. The code is cleaner, and more cases are handled. The option
525        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
526        options in pcretest are provided to set it. It can also be set by
527        (*NO_AUTO_POSSESS) at the start of a pattern.
528    
529    18. The character VT has been added to the default ("C" locale) set of
530        characters that match \s and are generally treated as white space,
531        following this same change in Perl 5.18. There is now no difference between
532        "Perl space" and "POSIX space". Whether VT is treated as white space in
533        other locales depends on the locale.
534    
535    19. The code for checking named groups as conditions, either for being set or
536        for being recursed, has been refactored (this is related to 14 and 15
537        above). Processing unduplicated named groups should now be as fast at
538        numerical groups, and processing duplicated groups should be faster than
539        before.
540    
541    20. Two patches to the CMake build system, by Alexander Barkov:
542    
543          (1) Replace the "source" command by "." in CMakeLists.txt because
544              "source" is a bash-ism.
545    
546          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
547              without these the CMake build does not work on Solaris.
548    
549    21. Perl has changed its handling of \8 and \9. If there is no previously
550        encountered capturing group of those numbers, they are treated as the
551        literal characters 8 and 9 instead of a binary zero followed by the
552        literals. PCRE now does the same.
553    
554    22. Following Perl, added \o{} to specify codepoints in octal, making it
555        possible to specify values greater than 0777 and also making them
556        unambiguous.
557    
558    23. Perl now gives an error for missing closing braces after \x{... instead of
559        treating the string as literal. PCRE now does the same.
560    
561    24. RunTest used to grumble if an inappropriate test was selected explicitly,
562        but just skip it when running all tests. This make it awkward to run ranges
563        of tests when one of them was inappropriate. Now it just skips any
564        inappropriate tests, as it always did when running all tests.
565    
566    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
567        character types such as \d or \w, too many callouts were inserted, and the
568        data that they returned was rubbish.
569    
570    26. In UCP mode, \s was not matching two of the characters that Perl matches,
571        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
572        were matched by \h. The code has now been refactored so that the lists of
573        the horizontal and vertical whitespace characters used for \h and \v (which
574        are defined only in one place) are now also used for \s.
575    
576    27. Add JIT support for the 64 bit TileGX architecture.
577        Patch by Jiong Wang (Tilera Corporation).
578    
579    28. Possessive quantifiers for classes (both explicit and automatically
580        generated) now use special opcodes instead of wrapping in ONCE brackets.
581    
582    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
583        (because it's meaningless), this was not happening when PCRE_CASELESS was
584        set. Not wrong, but inefficient.
585    
586    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
587        properties for \w, \d, etc) is present in a test regex. Otherwise if the
588        test contains no characters greater than 255, Perl doesn't realise it
589        should be using Unicode semantics.
590    
591    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
592        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
593        does in Unicode mode.
594    
595    32. Added the "forbid" facility to pcretest so that putting tests into the
596        wrong test files can sometimes be quickly detected.
597    
598    33. There is now a limit (default 250) on the depth of nesting of parentheses.
599        This limit is imposed to control the amount of system stack used at compile
600        time. It can be changed at build time by --with-parens-nest-limit=xxx or
601        the equivalent in CMake.
602    
603    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
604        errors. Perl warns for these when in warning mode, but PCRE has no facility
605        for giving warnings.
606    
607    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
608        However, this was not working for (?!) because it is optimized to (*FAIL),
609        for which PCRE does not allow quantifiers. The optimization is now disabled
610        when a quantifier follows (?!). I can't see any use for this, but it makes
611        things uniform.
612    
613    36. Perl no longer allows group names to start with digits, so I have made this
614        change also in PCRE. It simplifies the code a bit.
615    
616    37. In extended mode, Perl ignores spaces before a + that indicates a
617        possessive quantifier. PCRE allowed a space before the quantifier, but not
618        before the possessive +. It now does.
619    
620    38. The use of \K (reset reported match start) within a repeated possessive
621        group such as (a\Kb)*+ was not working.
622    
623    40. Document that the same character tables must be used at compile time and
624        run time, and that the facility to pass tables to pcre_exec() and
625        pcre_dfa_exec() is for use only with saved/restored patterns.
626    
627    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
628        features for Builds with MSVC:
629    
630        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
631           builds). The .rc files can be used to set FileDescription and many other
632           attributes.
633    
634        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
635           This allows higher-level build scripts which want .pdb files to avoid
636           hard-coding the exact files needed."
637    
638    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
639        mean "start of word" and "end of word", respectively, as a transition aid.
640    
641    43. A minimizing repeat of a class containing codepoints greater than 255 in
642        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
643        compiled to use the heap for recursion.
644    
645    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
646        is configured.
647    
648    
649    Version 8.33 28-May-2013
650    ------------------------
651    
652    1.  Added 'U' to some constants that are compared to unsigned integers, to
653        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
654        variables that are added to signed variables, to ensure the result is
655        signed and can be negated.
656    
657    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
658        CMake config files.
659    
660    3.  Revise the creation of config.h.generic so that all boolean macros are
661        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
662        overriding via -D on the command line possible.
663    
664    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
665        to unsigned int is reported to make a quite noticeable speed difference in
666        a specific Windows environment. Testing on Linux did also appear to show
667        some benefit (and it is clearly not harmful). Also fixed the definition of
668        Xop which should be unsigned.
669    
670    5.  Related to (4), changing the definition of the intermediate variable cc
671        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
672        improvements.
673    
674    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
675        unnecessary spaces.
676    
677    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
678        and later.
679    
680    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
681    
682    9.  Optimizing fast_forward_start_bits in JIT.
683    
684    10. Adding support for callouts in JIT, and fixing some issues revealed
685        during this work. Namely:
686    
687        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
688    
689        (b) Minimum length was not checked before the matching is started.
690    
691    11. The value of capture_last that is passed to callouts was incorrect in some
692        cases when there was a capture on one path that was subsequently abandoned
693        after a backtrack. Also, the capture_last value is now reset after a
694        recursion, since all captures are also reset in this case.
695    
696    12. The interpreter no longer returns the "too many substrings" error in the
697        case when an overflowing capture is in a branch that is subsequently
698        abandoned after a backtrack.
699    
700    13. In the pathological case when an offset vector of size 2 is used, pcretest
701        now prints out the matched string after a yield of 0 or 1.
702    
703    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
704        Only supported by the JIT compiler at the moment.
705    
706    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
707    
708    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
709        offset of the starting point of the matching process, provided the offsets
710        vector is large enough.
711    
712    17. The \A escape now records a lookbehind value of 1, though its execution
713        does not actually inspect the previous character. This is to ensure that,
714        in partial multi-segment matching, at least one character from the old
715        segment is retained when a new segment is processed. Otherwise, if there
716        are no lookbehinds in the pattern, \A might match incorrectly at the start
717        of a new segment.
718    
719    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
720    
721    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
722        gives some modest performance improvement in 8-bit mode.
723    
724    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
725        be expressed in certain programming languages using Universal Character
726        Names.
727    
728    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
729        which points out that "non characters" are not "characters that may not
730        appear in Unicode strings" but rather "characters that are reserved for
731        internal use and have only local meaning".
732    
733    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
734        there was a conditional group that depended on an assertion, if the
735        assertion was false, the callout that immediately followed the alternation
736        in the condition was skipped when pcre_exec() was used for matching.
737    
738    23. Allow an explicit callout to be inserted before an assertion that is the
739        condition for a conditional group, for compatibility with automatic
740        callouts, which always insert a callout at this point.
741    
742    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
743        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
744    
745    25. (*PRUNE) is now supported by the JIT compiler.
746    
747    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
748    
749    27. Fix the case where there are two or more SKIPs with arguments that may be
750        ignored.
751    
752    28. (*SKIP) is now supported by the JIT compiler.
753    
754    29. (*THEN) is now supported by the JIT compiler.
755    
756    30. Update RunTest with additional test selector options.
757    
758    31. The way PCRE handles backtracking verbs has been changed in two ways.
759    
760        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
761        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
762        backtracking. In some cases this makes it more Perl-compatible, but Perl's
763        rather obscure rules do not always do the same thing.
764    
765        (2) Previously, backtracking verbs were confined within assertions. This is
766        no longer the case for positive assertions, except for (*ACCEPT). Again,
767        this sometimes improves Perl compatibility, and sometimes does not.
768    
769    32. A number of tests that were in test 2 because Perl did things differently
770        have been moved to test 1, because either Perl or PCRE has changed, and
771        these tests are now compatible.
772    
773    32. Backtracking control verbs are now handled in the same way in JIT and
774        interpreter.
775    
776    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
777        contained a forward subroutine reference caused a compile error.
778    
779    34. Auto-detect and optimize limited repetitions in JIT.
780    
781    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
782        blocking (*UTF) etc.
783    
784    36. In the interpreter, maximizing pattern repetitions for characters and
785        character types now use tail recursion, which reduces stack usage.
786    
787    37. The value of the max lookbehind was not correctly preserved if a compiled
788        and saved regex was reloaded on a host of different endianness.
789    
790    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
791        of the compiled pattern block, expand the flags field from 16 to 32 bits
792        because it was almost full.
793    
794    39. Try madvise first before posix_madvise.
795    
796    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
797        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
798        original change locked out all matches of empty strings. This has been
799        changed so that one match of an empty string per line is recognized.
800        Subsequent searches on the same line (for colouring or for --only-matching,
801        for example) do not recognize empty strings.
802    
803    41. Applied a user patch to fix a number of spelling mistakes in comments.
804    
805    42. Data lines longer than 65536 caused pcretest to crash.
806    
807    43. Clarified the data type for length and startoffset arguments for pcre_exec
808        and pcre_dfa_exec in the function-specific man pages, where they were
809        explicitly stated to be in bytes, never having been updated. I also added
810        some clarification to the pcreapi man page.
811    
812    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
813        a segmentation fault.
814    
815    
816    Version 8.32 30-November-2012
817    -----------------------------
818    
819    1.  Improved JIT compiler optimizations for first character search and single
820        character iterators.
821    
822    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
823        Patch by Daniel Richard G.
824    
825    3.  Single character iterator optimizations in the JIT compiler.
826    
827    4.  Improved JIT compiler optimizations for character ranges.
828    
829    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
830        Reported by Giuseppe D'Angelo.
831    
832    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
833        of a line, was being set incorrectly in cases where .* appeared inside
834        atomic brackets at the start of a pattern, or where there was a subsequent
835        *PRUNE or *SKIP.
836    
837    7.  Improved instruction cache flush for POWER/PowerPC.
838        Patch by Daniel Richard G.
839    
840    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
841        grep:
842    
843        (a) There is now no limit to the number of patterns to be matched.
844    
845        (b) An error is given if a pattern is too long.
846    
847        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
848            are now supported.
849    
850        (d) --exclude-from and --include-from (multiple use) have been added.
851    
852        (e) Exclusions and inclusions now apply to all files and directories, not
853            just to those obtained from scanning a directory recursively.
854    
855        (f) Multiple uses of -f and --file-list are now supported.
856    
857        (g) In a Windows environment, the default for -d has been changed from
858            "read" (the GNU grep default) to "skip", because otherwise the presence
859            of a directory in the file list provokes an error.
860    
861        (h) The documentation has been revised and clarified in places.
862    
863    9.  Improve the matching speed of capturing brackets.
864    
865    10. Changed the meaning of \X so that it now matches a Unicode extended
866        grapheme cluster.
867    
868    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
869        out POSIX threads when JIT support is configured.
870    
871    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
872    
873    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
874        structure could go wrong in environments where size_t is not the same size
875        as int.
876    
877    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
878        set.
879    
880    15. The EBCDIC support had decayed; later updates to the code had included
881        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
882        general tidy up of EBCDIC-related issues, and the documentation was also
883        not quite right. There is now a test that can be run on ASCII systems to
884        check some of the EBCDIC-related things (but is it not a full test).
885    
886    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
887        in a small tidy to the code.
888    
889    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
890    
891    18. If the --only-matching (-o) option in pcregrep is specified multiple
892        times, each one causes appropriate output. For example, -o1 -o2 outputs the
893        substrings matched by the 1st and 2nd capturing parentheses. A separating
894        string can be specified by --om-separator (default empty).
895    
896    19. Improving the first n character searches.
897    
898    20. Turn case lists for horizontal and vertical white space into macros so that
899        they are defined only once.
900    
901    21. This set of changes together give more compatible Unicode case-folding
902        behaviour for characters that have more than one other case when UCP
903        support is available.
904    
905        (a) The Unicode property table now has offsets into a new table of sets of
906            three or more characters that are case-equivalent. The MultiStage2.py
907            script that generates these tables (the pcre_ucd.c file) now scans
908            CaseFolding.txt instead of UnicodeData.txt for character case
909            information.
910    
911        (b) The code for adding characters or ranges of characters to a character
912            class has been abstracted into a generalized function that also handles
913            case-independence. In UTF-mode with UCP support, this uses the new data
914            to handle characters with more than one other case.
915    
916        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
917            whose other case is greater than 256 are now correctly matched
918            caselessly. Previously, the high codepoint matched the low one, but not
919            vice versa.
920    
921        (d) The processing of \h, \H, \v, and \ in character classes now makes use
922            of the new class addition function, using character lists defined as
923            macros alongside the case definitions of 20 above.
924    
925        (e) Caseless back references now work with characters that have more than
926            one other case.
927    
928        (f) General caseless matching of characters with more than one other case
929            is supported.
930    
931    22. Unicode character properties were updated from Unicode 6.2.0
932    
933    23. Improved CMake support under Windows. Patch by Daniel Richard G.
934    
935    24. Add support for 32-bit character strings, and UTF-32
936    
937    25. Major JIT compiler update (code refactoring and bugfixing).
938        Experimental Sparc 32 support is added.
939    
940    26. Applied a modified version of Daniel Richard G's patch to create
941        pcre.h.generic and config.h.generic by "make" instead of in the
942        PrepareRelease script.
943    
944    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
945        pcre_compile.c when checking for a zero character.
946    
947    28. Introducing a native interface for JIT. Through this interface, the compiled
948        machine code can be directly executed. The purpose of this interface is to
949        provide fast pattern matching, so several sanity checks are not performed.
950        However, feature tests are still performed. The new interface provides
951        1.4x speedup compared to the old one.
952    
953    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
954        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
955        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
956    
957    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
958        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
959        the "old" RFC 2279). Instead, it ended up passing a negative length to
960        pcre_exec().
961    
962    31. Add support for GCC's visibility feature to hide internal functions.
963    
964    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
965        "unknown -C option" after outputting 0 or 1.
966    
967    33. There is now support for generating a code coverage report for the test
968        suite in environments where gcc is the compiler and lcov is installed. This
969        is mainly for the benefit of the developers.
970    
971    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
972        unaddressable using valgrind annotations, allowing valgrind to detect
973        invalid memory accesses. This is mainly for the benefit of the developers.
974    
975    25. (*UTF) can now be used to start a pattern in any of the three libraries.
976    
977    26. Give configure error if --enable-cpp but no C++ compiler found.
978    
979    
980    Version 8.31 06-July-2012
981    -------------------------
982    
983    1.  Fixing a wrong JIT test case and some compiler warnings.
984    
985    2.  Removed a bashism from the RunTest script.
986    
987    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
988        to unsigned type, result still unsigned" that was given by an MS compiler
989        on encountering the code "-sizeof(xxx)".
990    
991    4.  Partial matching support is added to the JIT compiler.
992    
993    5.  Fixed several bugs concerned with partial matching of items that consist
994        of more than one character:
995    
996        (a) /^(..)\1/ did not partially match "aba" because checking references was
997            done on an "all or nothing" basis. This also applied to repeated
998            references.
999    
1000        (b) \R did not give a hard partial match if \r was found at the end of the
1001            subject.
1002    
1003        (c) \X did not give a hard partial match after matching one or more
1004            characters at the end of the subject.
1005    
1006        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1007            a partial match for the string "\r".
1008    
1009        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1010            a partial match for a CR character at the end of the subject string.
1011    
1012    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1013        running pcretest, the text "(JIT)" added to the output whenever JIT is
1014        actually used to run the match.
1015    
1016    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1017        or /S+[+] with a digit between 1 and 7.
1018    
1019    8.  OP_NOT now supports any UTF character not just single-byte ones.
1020    
1021    9.  (*MARK) control verb is now supported by the JIT compiler.
1022    
1023    10. The command "./RunTest list" lists the available tests without actually
1024        running any of them. (Because I keep forgetting what they all are.)
1025    
1026    11. Add PCRE_INFO_MAXLOOKBEHIND.
1027    
1028    12. Applied a (slightly modified) user-supplied patch that improves performance
1029        when the heap is used for recursion (compiled with --disable-stack-for-
1030        recursion). Instead of malloc and free for each heap frame each time a
1031        logical recursion happens, frames are retained on a chain and re-used where
1032        possible. This sometimes gives as much as 30% improvement.
1033    
1034    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1035        call.
1036    
1037    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1038    
1039    15. It is now possible to link pcretest with libedit as an alternative to
1040        libreadline.
1041    
1042    16. (*COMMIT) control verb is now supported by the JIT compiler.
1043    
1044    17. The Unicode data tables have been updated to Unicode 6.1.0.
1045    
1046    18. Added --file-list option to pcregrep.
1047    
1048    19. Added binary file support to pcregrep, including the -a, --binary-files,
1049        -I, and --text options.
1050    
1051    20. The madvise function is renamed for posix_madvise for QNX compatibility
1052        reasons. Fixed by Giuseppe D'Angelo.
1053    
1054    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1055    
1056    22. Changed the option for creating symbolic links for 16-bit man pages from
1057        -s to -sf so that re-installing does not cause issues.
1058    
1059    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1060    
1061    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1062        matches in certain environments (the workspace was not being correctly
1063        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1064        some of the workspace data at the beginning of a restart.
1065    
1066    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1067        was not doing so when it should - probably a typo introduced by SVN 528
1068        (change 8.10/14).
1069    
1070    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1071        \w+ when the character tables indicated that \x{c4} was a word character.
1072        There were several related cases, all because the tests for doing a table
1073        lookup were testing for characters less than 127 instead of 255.
1074    
1075    27. If a pattern contains capturing parentheses that are not used in a match,
1076        their slots in the ovector are set to -1. For those that are higher than
1077        any matched groups, this happens at the end of processing. In the case when
1078        there were back references that the ovector was too small to contain
1079        (causing temporary malloc'd memory to be used during matching), and the
1080        highest capturing number was not used, memory off the end of the ovector
1081        was incorrectly being set to -1. (It was using the size of the temporary
1082        memory instead of the true size.)
1083    
1084    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1085        ovector size, it uses memory at the end of the block that it has got.
1086    
1087    29. Check for an overlong MARK name and give an error at compile time. The
1088        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1089    
1090    30. JIT compiler update.
1091    
1092    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1093        Rill for the patch.
1094    
1095    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1096    
1097    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1098    
1099    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1100        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1101        was enabled.
1102    
1103    35. Improve JIT code generation for greedy plus quantifier.
1104    
1105    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1106        1 to "aa" instead of to an empty string. The bug affected repeated groups
1107        that could potentially match an empty string.
1108    
1109    37. Optimizing single character iterators in JIT.
1110    
1111    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1112        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1113        codepoints that are too big for the mode are faulted, and in a UTF mode,
1114        disallowed codepoints are also faulted.
1115    
1116    39. If PCRE was compiled with UTF support, in three places in the DFA
1117        matcher there was code that should only have been obeyed in UTF mode, but
1118        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1119        processing when bytes with values greater than 127 were present. In 16-bit
1120        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1121        both cases the values are those that cannot be the first data item in a UTF
1122        character. The three items that might have provoked this were recursions,
1123        possessively repeated groups, and atomic groups.
1124    
1125    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1126        and pcregrep, because some OS require shared objects to be explicitly
1127        passed to ld, causing the link step to fail if they are not.
1128    
1129    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1130        mode, patterns that started with \h* or \R* might be incorrectly matched.
1131    
1132    
1133    Version 8.30 04-February-2012
1134    -----------------------------
1135    
1136    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1137        name is defined in ctype.h.
1138    
1139    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1140        only in quite long subpatterns.
1141    
1142    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1143        since it was replaced by pcre_fullinfo() in February 2000.
1144    
1145    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1146        match a (*MARK), and the match failed at the start of the subject, a
1147        reference to memory before the start of the subject could occur. This bug
1148        was introduced by fix 17 of release 8.21.
1149    
1150    5.  A reference to an unset group with zero minimum repetition was giving
1151        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1152        /(another)?(\1?)test/ matched against "hello world test". This bug was
1153        introduced in release 8.13.
1154    
1155    6.  Add support for 16-bit character strings (a large amount of work involving
1156        many changes and refactorings).
1157    
1158    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1159        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1160        from a file.
1161    
1162    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1163        rounding is not applied in this particular case).
1164    
1165    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1166        if they appear, or are escaped, in patterns.
1167    
1168    10. Get rid of a number of -Wunused-but-set-variable warnings.
1169    
1170    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1171        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1172        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1173        also returns the mark "x". This bug applied to capturing parentheses,
1174        non-capturing parentheses, and atomic parentheses. It also applied to some
1175        assertions.
1176    
1177    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1178        information out of configure.ac instead of relying on pcre.h.generic, which
1179        is not stored in the repository.
1180    
1181    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1182        -lreadline.
1183    
1184    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1185    
1186    15. Applied Graycode's patch to put the top-level frame on the stack rather
1187        than the heap when not using the stack for recursion. This gives a
1188        performance improvement in many cases when recursion is not deep.
1189    
1190    16. Experimental code added to "pcretest -C" to output the stack frame size.
1191    
1192    
1193    Version 8.21 12-Dec-2011
1194    ------------------------
1195    
1196    1.  Updating the JIT compiler.
1197    
1198    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1199        are added as well.
1200    
1201    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1202        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1203        calling _pcre_jit_exec. Some extra comments are added.
1204    
1205    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1206        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1207        was introduced by change 18 for 8.20.
1208    
1209    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1210        ECMA-262 standard.
1211    
1212    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1213        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1214        This bug was probably introduced by change 9 of 8.13.
1215    
1216    7.  While fixing 6 above, I noticed that a number of other items were being
1217        incorrectly rejected as "not fixed length". This arose partly because newer
1218        opcodes had not been added to the fixed-length checking code. I have (a)
1219        corrected the bug and added tests for these items, and (b) arranged for an
1220        error to occur if an unknown opcode is encountered while checking for fixed
1221        length instead of just assuming "not fixed length". The items that were
1222        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1223        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1224        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1225    
1226    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1227        being incorrectly compiled and would have given unpredicatble results.
1228    
1229    9.  A possessively repeated subpattern with minimum repeat count greater than
1230        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1231        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1232        the first (A) could occur when it should not.
1233    
1234    10. Add a cast and remove a redundant test from the code.
1235    
1236    11. JIT should use pcre_malloc/pcre_free for allocation.
1237    
1238    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1239        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1240        is anything other than /usr, -L is still shown).
1241    
1242    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1243    
1244    14. Perl does not support \N without a following name in a [] class; PCRE now
1245        also gives an error.
1246    
1247    15. If a forward reference was repeated with an upper limit of around 2000,
1248        it caused the error "internal error: overran compiling workspace". The
1249        maximum number of forward references (including repeats) was limited by the
1250        internal workspace, and dependent on the LINK_SIZE. The code has been
1251        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1252        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1253        of around 200,000 forward references. While doing this, I also speeded up
1254        the filling in of repeated forward references.
1255    
1256    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1257        incorrectly expecting the subject to contain another "a" after the start.
1258    
1259    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1260        in the match, the SKIP should be ignored. This was not happening; instead
1261        the SKIP was being treated as NOMATCH. For patterns such as
1262        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1263        tested.
1264    
1265    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1266        now much more compatible with Perl, in particular in cases where the result
1267        is a non-match for a non-anchored pattern. For example, if
1268        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1269        "m", where previously it did not return a name. A side effect of this
1270        change is that for partial matches, the last encountered mark name is
1271        returned, as for non matches. A number of tests that were previously not
1272        Perl-compatible have been moved into the Perl-compatible test files. The
1273        refactoring has had the pleasing side effect of removing one argument from
1274        the match() function, thus reducing its stack requirements.
1275    
1276    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1277        subsequent uses of /S (without +) incorrectly behaved like /S+.
1278    
1279    21. Retrieve executable code size support for the JIT compiler and fixing
1280        some warnings.
1281    
1282    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1283        not work when the shorter character appeared right at the end of the
1284        subject string.
1285    
1286    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1287        systems.
1288    
1289    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1290        output it when the /M option is used in pcretest.
1291    
1292    25. The CheckMan script was not being included in the distribution. Also, added
1293        an explicit "perl" to run Perl scripts from the PrepareRelease script
1294        because this is reportedly needed in Windows.
1295    
1296    26. If study data was being save in a file and studying had not found a set of
1297        "starts with" bytes for the pattern, the data written to the file (though
1298        never used) was taken from uninitialized memory and so caused valgrind to
1299        complain.
1300    
1301    27. Updated RunTest.bat as provided by Sheri Pierce.
1302    
1303    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1304    
1305    29. Computation of memory usage for the table of capturing group names was
1306        giving an unnecessarily large value.
1307    
1308    
1309    Version 8.20 21-Oct-2011
1310    ------------------------
1311    
1312    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1313        a POSIX class. After further experiments with Perl, which convinced me that
1314        Perl has bugs and confusions, a closing square bracket is no longer allowed
1315        in a POSIX name. This bug also affected patterns with classes that started
1316        with full stops.
1317    
1318    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1319        captured substring, but while checking the failing first alternative,
1320        substring 1 is temporarily captured. If the output vector supplied to
1321        pcre_exec() was not big enough for this capture, the yield of the function
1322        was still zero ("insufficient space for captured substrings"). This cannot
1323        be totally fixed without adding another stack variable, which seems a lot
1324        of expense for a edge case. However, I have improved the situation in cases
1325        such as /(a)(b)x|abc/ matched against "abc", where the return code
1326        indicates that fewer than the maximum number of slots in the ovector have
1327        been set.
1328    
1329    3.  Related to (2) above: when there are more back references in a pattern than
1330        slots in the output vector, pcre_exec() uses temporary memory during
1331        matching, and copies in the captures as far as possible afterwards. It was
1332        using the entire output vector, but this conflicts with the specification
1333        that only 2/3 is used for passing back captured substrings. Now it uses
1334        only the first 2/3, for compatibility. This is, of course, another edge
1335        case.
1336    
1337    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1338        main code base, and can be used by building with --enable-jit. When this is
1339        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1340        runtime --no-jit option is given.
1341    
1342    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1343        ovector, the return from the function was zero, implying that there were
1344        other matches that did not fit. The correct "exactly full" value is now
1345        returned.
1346    
1347    6.  If a subpattern that was called recursively or as a subroutine contained
1348        (*PRUNE) or any other control that caused it to give a non-standard return,
1349        invalid errors such as "Error -26 (nested recursion at the same subject
1350        position)" or even infinite loops could occur.
1351    
1352    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1353        computing the minimum length on reaching *ACCEPT, and so ended up with the
1354        wrong value of 1 rather than 0. Further investigation indicates that
1355        computing a minimum subject length in the presence of *ACCEPT is difficult
1356        (think back references, subroutine calls), and so I have changed the code
1357        so that no minimum is registered for a pattern that contains *ACCEPT.
1358    
1359    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1360        it was not handled as intended. [But see 16 below.]
1361    
1362    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1363        Sheri Pierce.
1364    
1365    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1366        the first byte in a match must be "a".
1367    
1368    11. Change 17 for 8.13 increased the recursion depth for patterns like
1369        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1370        pattern contains any instances of (*THEN). If it does not, the old
1371        optimizations are restored. It would be nice to do this on a per-group
1372        basis, but at the moment that is not feasible.
1373    
1374    12. In some environments, the output of pcretest -C is CRLF terminated. This
1375        broke RunTest's code that checks for the link size. A single white space
1376        character after the value is now allowed for.
1377    
1378    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1379        For "fr", it uses the Windows-specific input and output files.
1380    
1381    14. If (*THEN) appeared in a group that was called recursively or as a
1382        subroutine, it did not work as intended. [But see next item.]
1383    
1384    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1385        pattern fragments (but not containing any | characters). If A and B are
1386        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1387        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1388        D. In other words, Perl considers parentheses that do not contain any |
1389        characters to be part of a surrounding alternative, whereas PCRE was
1390        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1391        differently. PCRE now behaves in the same way as Perl, except in the case
1392        of subroutine/recursion calls such as (?1) which have in any case always
1393        been different (but PCRE had them first :-).
1394    
1395    16. Related to 15 above: Perl does not treat the | in a conditional group as
1396        creating alternatives. Such a group is treated in the same way as an
1397        ordinary group without any | characters when processing (*THEN). PCRE has
1398        been changed to match Perl's behaviour.
1399    
1400    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1401        RunGrepTest script failed.
1402    
1403    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1404        inevitable for groups that contain captures, but it can lead to a lot of
1405        stack use in large patterns. The old behaviour has been restored for atomic
1406        groups that do not contain any capturing parentheses.
1407    
1408    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1409        suppress the check for a minimum subject length at run time. (If it was
1410        given to pcre_exec() or pcre_dfa_exec() it did work.)
1411    
1412    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1413        fail to work when decoding hex characters in data strings in EBCDIC
1414        environments.
1415    
1416    21. It appears that in at least one Mac OS environment, the isxdigit() function
1417        is implemented as a macro that evaluates to its argument more than once,
1418        contravening the C 90 Standard (I haven't checked a later standard). There
1419        was an instance in pcretest which caused it to go wrong when processing
1420        \x{...} escapes in subject strings. The has been rewritten to avoid using
1421        things like p++ in the argument of isxdigit().
1422    
1423    
1424    Version 8.13 16-Aug-2011
1425    ------------------------
1426    
1427    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1428    
1429    2.  Two minor typos in pcre_internal.h have been fixed.
1430    
1431    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1432        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1433        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1434    
1435    4.  There were a number of related bugs in the code for matching backrefences
1436        caselessly in UTF-8 mode when codes for the characters concerned were
1437        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1438        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1439        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1440        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1441        2-byte code at the end of the subject (it thought there wasn't enough data
1442        left).
1443    
1444    5.  Comprehensive information about what went wrong is now returned by
1445        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1446        as the output vector has at least 2 elements. The offset of the start of
1447        the failing character and a reason code are placed in the vector.
1448    
1449    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1450        now returned is for the first byte of the failing character, instead of the
1451        last byte inspected. This is an incompatible change, but I hope it is small
1452        enough not to be a problem. It makes the returned offset consistent with
1453        pcre_exec() and pcre_dfa_exec().
1454    
1455    7.  pcretest now gives a text phrase as well as the error number when
1456        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1457        failure, the offset and reason code are output.
1458    
1459    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1460        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1461        back over a single character (\n). This seems wrong (because it treated the
1462        two characters as a single entity when going forwards), conflicts with the
1463        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1464        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1465        has been changed.
1466    
1467    9.  Some internal refactoring has changed the processing so that the handling
1468        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1469        time (the PCRE_DOTALL option was changed this way some time ago: version
1470        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1471        which was always a bit of a fudge. It also means that there is one less
1472        argument for the match() function, which reduces its stack requirements
1473        slightly. This change also fixes an incompatibility with Perl: the pattern
1474        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1475    
1476    10. More internal refactoring has drastically reduced the number of recursive
1477        calls to match() for possessively repeated groups such as (abc)++ when
1478        using pcre_exec().
1479    
1480    11. While implementing 10, a number of bugs in the handling of groups were
1481        discovered and fixed:
1482    
1483        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1484        (a|)*(?1) gave a compile-time internal error.
1485        ((a|)+)+  did not notice that the outer group could match an empty string.
1486        (^a|^)+   was not marked as anchored.
1487        (.*a|.*)+ was not marked as matching at start or after a newline.
1488    
1489    12. Yet more internal refactoring has removed another argument from the match()
1490        function. Special calls to this function are now indicated by setting a
1491        value in a variable in the "match data" data block.
1492    
1493    13. Be more explicit in pcre_study() instead of relying on "default" for
1494        opcodes that mean there is no starting character; this means that when new
1495        ones are added and accidentally left out of pcre_study(), testing should
1496        pick them up.
1497    
1498    14. The -s option of pcretest has been documented for ages as being an old
1499        synonym of -m (show memory usage). I have changed it to mean "force study
1500        for every regex", that is, assume /S for every regex. This is similar to -i
1501        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1502        using it. It makes it easier to run collections of tests with and without
1503        study enabled, and thereby test pcre_study() more easily. All the standard
1504        tests are now run with and without -s (but some patterns can be marked as
1505        "never study" - see 20 below).
1506    
1507    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1508        restoration of the capturing data to the outer values was not happening
1509        correctly.
1510    
1511    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1512        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1513        pattern had matched an empty string, and so incorrectly returned a no
1514        match.
1515    
1516    17. There was optimizing code for the last branch of non-capturing parentheses,
1517        and also for the obeyed branch of a conditional subexpression, which used
1518        tail recursion to cut down on stack usage. Unfortunately, now that there is
1519        the possibility of (*THEN) occurring in these branches, tail recursion is
1520        no longer possible because the return has to be checked for (*THEN). These
1521        two optimizations have therefore been removed. [But see 8.20/11 above.]
1522    
1523    18. If a pattern containing \R was studied, it was assumed that \R always
1524        matched two bytes, thus causing the minimum subject length to be
1525        incorrectly computed because \R can also match just one byte.
1526    
1527    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1528        was incorrectly computed.
1529    
1530    20. If /S is present twice on a test pattern in pcretest input, it now
1531        *disables* studying, thereby overriding the use of -s on the command line
1532        (see 14 above). This is necessary for one or two tests to keep the output
1533        identical in both cases.
1534    
1535    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1536        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1537    
1538    22. When an atomic group that contained a capturing parenthesis was
1539        successfully matched, but the branch in which it appeared failed, the
1540        capturing was not being forgotten if a higher numbered group was later
1541        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1542        group 1 to "a", when in fact it should be unset. This applied to multi-
1543        branched capturing and non-capturing groups, repeated or not, and also to
1544        positive assertions (capturing in negative assertions does not happen
1545        in PCRE) and also to nested atomic groups.
1546    
1547    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1548        subject after a captured substring, to make it easier to tell which of a
1549        number of identical substrings has been captured.
1550    
1551    24. The way atomic groups are processed by pcre_exec() has been changed so that
1552        if they are repeated, backtracking one repetition now resets captured
1553        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1554        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1555        "aaa". Previously, it would have been "a". As part of this code
1556        refactoring, the way recursive calls are handled has also been changed.
1557    
1558    25. If an assertion condition captured any substrings, they were not passed
1559        back unless some other capturing happened later. For example, if
1560        (?(?=(a))a) was matched against "a", no capturing was returned.
1561    
1562    26. When studying a pattern that contained subroutine calls or assertions,
1563        the code for finding the minimum length of a possible match was handling
1564        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1565        group 1 called group 2 while simultaneously a separate group 2 called group
1566        1). A stack overflow occurred in this case. I have fixed this by limiting
1567        the recursion depth to 10.
1568    
1569    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1570        Fortmann. This supports explicit test numbers on the command line, and has
1571        argument validation and error reporting.
1572    
1573    28. An instance of \X with an unlimited repeat could fail if at any point the
1574        first character it looked at was a mark character.
1575    
1576    29. Some minor code refactoring concerning Unicode properties and scripts
1577        should reduce the stack requirement of match() slightly.
1578    
1579    30. Added the '=' option to pcretest to check the setting of unused capturing
1580        slots at the end of the pattern, which are documented as being -1, but are
1581        not included in the return count.
1582    
1583    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1584        compiled something random. Now it gives a compile-time error (as does
1585        Perl).
1586    
1587    32. A *MARK encountered during the processing of a positive assertion is now
1588        recorded and passed back (compatible with Perl).
1589    
1590    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1591        had alternative anchored branches, the search for a second match in a line
1592        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1593        matched the line "0102" twice. The same bug affected patterns that started
1594        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1595        twice.
1596    
1597    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1598        does, and because of capturing effects, quantifying parenthesized
1599        assertions may at times be useful. Quantifiers are now allowed for
1600        parenthesized assertions.
1601    
1602    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1603    
1604    36. \g was being checked for fancy things in a character class, when it should
1605        just be a literal "g".
1606    
1607    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1608        appearance of a nested POSIX class supersedes an apparent external class.
1609        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1610        unescaped square brackets may also appear as part of class names. For
1611        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1612        more like Perl. (But see 8.20/1 above.)
1613    
1614    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1615        was because it thought it was \N{name}, which is not supported).
1616    
1617    39. Add minix to OS list not supporting the -S option in pcretest.
1618    
1619    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1620        cannot analyze patterns in sufficient detail to catch mutual recursions
1621        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1622        subgroup is called recursively as a subpattern for a second time at the
1623        same position in the subject string. In previous releases this might have
1624        been caught by the recursion limit, or it might have run out of stack.
1625    
1626    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1627        happen only once. PCRE was, however incorrectly giving a compile time error
1628        "recursive call could loop indefinitely" because it cannot analyze the
1629        pattern in sufficient detail. The compile time test no longer happens when
1630        PCRE is compiling a conditional subpattern, but actual runaway loops are
1631        now caught at runtime (see 40 above).
1632    
1633    42. It seems that Perl allows any characters other than a closing parenthesis
1634        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1635        has been changed to be the same.
1636    
1637    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1638        as not to get warnings when autogen.sh is called. Also changed
1639        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1640    
1641    44. To help people who use pcregrep to scan files containing exceedingly long
1642        lines, the following changes have been made:
1643    
1644        (a) The default value of the buffer size parameter has been increased from
1645            8K to 20K. (The actual buffer used is three times this size.)
1646    
1647        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1648            PCRE is built.
1649    
1650        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1651            to be set at run time.
1652    
1653        (d) Numerical values in pcregrep options can be followed by K or M, for
1654            example --buffer-size=50K.
1655    
1656        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1657            given and the return code is set to 2.
1658    
1659    45. Add a pointer to the latest mark to the callout data block.
1660    
1661    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1662        partial match of an empty string instead of no match. This was specific to
1663        the use of ".".
1664    
1665    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1666        complete match instead of a partial match. This bug was dependent on both
1667        the PCRE_UTF8 and PCRE_DOTALL options being set.
1668    
1669    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1670        starting byte set, because \b was not being ignored.
1671    
1672    
1673    Version 8.12 15-Jan-2011
1674    ------------------------
1675    
1676    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1677        checks for such things as part of the documentation building process.
1678    
1679    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1680        --match-limit and --recursion-limit options (added for 8.11). In
1681        particular, this made one of the standard tests fail. (The integer value
1682        went into the wrong half of a long int.)
1683    
1684    3.  If the --colour option was given to pcregrep with -v (invert match), it
1685        did strange things, either producing crazy output, or crashing. It should,
1686        of course, ignore a request for colour when reporting lines that do not
1687        match.
1688    
1689    4.  Another pcregrep bug caused similar problems if --colour was specified with
1690        -M (multiline) and the pattern match finished with a line ending.
1691    
1692    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1693        matched in multiline mode, the following line was shown as part of the
1694        match. This seems wrong, so I have changed it.
1695    
1696    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1697        the check for further matches in the same line (so they could be coloured)
1698        to overrun the end of the current line. If another match was found, it was
1699        incorrectly shown (and then shown again when found in the next line).
1700    
1701    7.  If pcregrep was compiled under Windows, there was a reference to the
1702        function pcregrep_exit() before it was defined. I am assuming this was
1703        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1704        reported by a user. I've moved the definition above the reference.
1705    
1706    
1707    Version 8.11 10-Dec-2010
1708    ------------------------
1709    
1710    1.  (*THEN) was not working properly if there were untried alternatives prior
1711        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1712        backtracked to try for "b" instead of moving to the next alternative branch
1713        at the same level (in this case, to look for "c"). The Perl documentation
1714        is clear that when (*THEN) is backtracked onto, it goes to the "next
1715        alternative in the innermost enclosing group".
1716    
1717    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1718        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1719        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1720        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1721        (*THEN).
1722    
1723    3.  If \s appeared in a character class, it removed the VT character from
1724        the class, even if it had been included by some previous item, for example
1725        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1726        of \s, but is part of the POSIX "space" class.)
1727    
1728    4.  A partial match never returns an empty string (because you can always
1729        match an empty string at the end of the subject); however the checking for
1730        an empty string was starting at the "start of match" point. This has been
1731        changed to the "earliest inspected character" point, because the returned
1732        data for a partial match starts at this character. This means that, for
1733        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1734        (previously it gave "no match").
1735    
1736    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1737        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1738        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1739        has an implication that the given string is incomplete (because a partial
1740        match is preferred over a full match). For this reason, these items now
1741        give a partial match in this situation. [Aside: previously, the one case
1742        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1743        match rather than a full match, which was wrong by the old rules, but is
1744        now correct.]
1745    
1746    6.  There was a bug in the handling of #-introduced comments, recognized when
1747        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1748        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1749        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1750        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1751        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1752        places in pcre_compile().
1753    
1754    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1755        comments when looking ahead for named forward references to subpatterns,
1756        the only newline sequence it recognized was NL. It now handles newlines
1757        according to the set newline convention.
1758    
1759    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1760        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1761        cater for a lack of strerror(). These oversights have been fixed.
1762    
1763    9.  Added --match-limit and --recursion-limit to pcregrep.
1764    
1765    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1766    
1767    11. When the -o option was used, pcregrep was setting a return code of 1, even
1768        when matches were found, and --line-buffered was not being honoured.
1769    
1770    12. Added an optional parentheses number to the -o and --only-matching options
1771        of pcregrep.
1772    
1773    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1774        can match an empty string. The code to do it in pcretest and pcredemo
1775        needed fixing:
1776    
1777        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1778            only one byte after an empty string match just before CRLF (this case
1779            just got forgotten; "any" and "anycrlf" were OK).
1780    
1781        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1782            mode when an empty string match preceded an ASCII character followed by
1783            a non-ASCII character. (The code for advancing by one character rather
1784            than one byte was nonsense.)
1785    
1786        (c) The pcredemo.c sample program did not have any code at all to handle
1787            the cases when CRLF is a valid newline sequence.
1788    
1789    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1790        as a starting offset was within the subject string. There is now a new
1791        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1792        negative or greater than the length of the string. In order to test this,
1793        pcretest is extended to allow the setting of negative starting offsets.
1794    
1795    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1796        starting offset points to the beginning of a UTF-8 character was
1797        unnecessarily clumsy. I tidied it up.
1798    
1799    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1800        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1801    
1802    17. Nobody had reported that the --include_dir option, which was added in
1803        release 7.7 should have been called --include-dir (hyphen, not underscore)
1804        for compatibility with GNU grep. I have changed it to --include-dir, but
1805        left --include_dir as an undocumented synonym, and the same for
1806        --exclude-dir, though that is not available in GNU grep, at least as of
1807        release 2.5.4.
1808    
1809    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1810        characters from a string of bytes) have been redefined so as not to use
1811        loops, in order to improve performance in some environments. At the same
1812        time, I abstracted some of the common code into auxiliary macros to save
1813        repetition (this should not affect the compiled code).
1814    
1815    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1816        compile-time error is now given if \c is not followed by an ASCII
1817        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1818        different, and any byte value is allowed.)
1819    
1820    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1821        START_OPTIMIZE option, which is now allowed at compile time - but just
1822        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1823        to pcregrep and other applications that have no direct access to PCRE
1824        options. The new /Y option in pcretest sets this option when calling
1825        pcre_compile().
1826    
1827    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1828        back references. Groups containing recursive back references were forced to
1829        be atomic by that change, but in the case of named groups, the amount of
1830        memory required was incorrectly computed, leading to "Failed: internal
1831        error: code overflow". This has been fixed.
1832    
1833    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1834        pcretest.c, to avoid build problems in some Borland environments.
1835    
1836    
1837    Version 8.10 25-Jun-2010
1838    ------------------------
1839    
1840    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1841        THEN.
1842    
1843    2.  (*ACCEPT) was not working when inside an atomic group.
1844    
1845    3.  Inside a character class, \B is treated as a literal by default, but
1846        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1847        causes the error). The code is unchanged, but I tidied the documentation.
1848    
1849    4.  Inside a character class, PCRE always treated \R and \X as literals,
1850        whereas Perl faults them if its -w option is set. I have changed PCRE so
1851        that it faults them when PCRE_EXTRA is set.
1852    
1853    5.  Added support for \N, which always matches any character other than
1854        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1855    
1856    6.  When compiling pcregrep with newer versions of gcc which may have
1857        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1858        declared with attribute warn_unused_result" were given. Just casting the
1859        result to (void) does not stop the warnings; a more elaborate fudge is
1860        needed. I've used a macro to implement this.
1861    
1862    7.  Minor change to pcretest.c to avoid a compiler warning.
1863    
1864    8.  Added four artifical Unicode properties to help with an option to make
1865        \s etc use properties (see next item). The new properties are: Xan
1866        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1867    
1868    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1869        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1870        this option. Modified pcretest to add /W to test this facility. Added
1871        REG_UCP to make it available via the POSIX interface.
1872    
1873    10. Added --line-buffered to pcregrep.
1874    
1875    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1876        studied, and the match started with a letter with a code point greater than
1877        127 whose first byte was different to the first byte of the other case of
1878        the letter, the other case of this starting letter was not recognized
1879        (#976).
1880    
1881    12. If a pattern that was studied started with a repeated Unicode property
1882        test, for example, \p{Nd}+, there was the theoretical possibility of
1883        setting up an incorrect bitmap of starting bytes, but fortunately it could
1884        not have actually happened in practice until change 8 above was made (it
1885        added property types that matched character-matching opcodes).
1886    
1887    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1888        possible starting bytes for non-anchored patterns.
1889    
1890    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1891        \R, and also a number of cases that involve Unicode properties, both
1892        explicit and implicit when PCRE_UCP is set.
1893    
1894    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1895        input, it could crash or give wrong results if characters with values
1896        greater than 0xc0 were present in the subject string. (Detail: it assumed
1897        UTF-8 input when processing these items.)
1898    
1899    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1900        size_t is 64-bit (#991).
1901    
1902    17. Added a check for running out of memory when PCRE is compiled with
1903        --disable-stack-for-recursion (#990).
1904    
1905    18. If the last data line in a file for pcretest does not have a newline on
1906        the end, a newline was missing in the output.
1907    
1908    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1909        less than 128) in its various bitmaps. However, there is a facility for
1910        generating tables according to the current locale when PCRE is compiled. It
1911        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1912        characters, are recognized by isspace() and therefore were getting set in
1913        these tables, and indeed these tables seem to approximate to ISO 8859. This
1914        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1915        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1916        which of course cannot start UTF-8 characters. I have changed the code so
1917        that only real ASCII characters (less than 128) and the correct starting
1918        bytes for UTF-8 encodings are set for characters greater than 127 when in
1919        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1920        altogether.)
1921    
1922    20. Added the /T option to pcretest so as to be able to run tests with non-
1923        standard character tables, thus making it possible to include the tests
1924        used for 19 above in the standard set of tests.
1925    
1926    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1927        reference to a subpattern the other side of a comment that contains an
1928        opening parenthesis caused either an internal compiling error, or a
1929        reference to the wrong subpattern.
1930    
1931    
1932    Version 8.02 19-Mar-2010
1933    ------------------------
1934    
1935    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1936    
1937    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1938        configured.
1939    
1940    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1941        original author of that file, following a query about its status.
1942    
1943    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1944        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1945    
1946    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1947        quantifier applied to a forward-referencing subroutine call, could compile
1948        incorrect code or give the error "internal error: previously-checked
1949        referenced subpattern not found".
1950    
1951    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1952        variables to point to external functions. For these systems, therefore,
1953        pcre_malloc etc. are now initialized to local functions that call the
1954        relevant global functions.
1955    
1956    7.  There were two entries missing in the vectors called coptable and poptable
1957        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1958        I've fixed the data, and added a kludgy way of testing at compile time that
1959        the lengths are correct (equal to the number of opcodes).
1960    
1961    8.  Following on from 7, I added a similar kludge to check the length of the
1962        eint vector in pcreposix.c.
1963    
1964    9.  Error texts for pcre_compile() are held as one long string to avoid too
1965        much relocation at load time. To find a text, the string is searched,
1966        counting zeros. There was no check for running off the end of the string,
1967        which could happen if a new error number was added without updating the
1968        string.
1969    
1970    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1971    
1972    11. \K was not working if it appeared in an atomic group or in a group that
1973        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1974        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1975        the assertion is positive, but not if it is negative.
1976    
1977    12. Change 11 fortuitously reduced the size of the stack frame used in the
1978        "match()" function of pcre_exec.c by one pointer. Forthcoming
1979        implementation of support for (*MARK) will need an extra pointer on the
1980        stack; I have reserved it now, so that the stack frame size does not
1981        decrease.
1982    
1983    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1984        item in branch that calls a recursion is a subroutine call - as in the
1985        second branch in the above example - was incorrectly given the compile-
1986        time error "recursive call could loop indefinitely" because pcre_compile()
1987        was not correctly checking the subroutine for matching a non-empty string.
1988    
1989    14. The checks for overrunning compiling workspace could trigger after an
1990        overrun had occurred. This is a "should never occur" error, but it can be
1991        triggered by pathological patterns such as hundreds of nested parentheses.
1992        The checks now trigger 100 bytes before the end of the workspace.
1993    
1994    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1995    
1996    
1997    Version 8.01 19-Jan-2010
1998    ------------------------
1999    
2000    1.  If a pattern contained a conditional subpattern with only one branch (in
2001        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2002        computed the wrong minimum data length (which is of course zero for such
2003        subpatterns). This could cause incorrect "no match" results.
2004    
2005    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2006        the pattern is reset in the first branch, pcre_compile() failed with
2007        "internal error: code overflow at offset...". This happened only when
2008        the reset was to the original external option setting. (An optimization
2009        abstracts leading options settings into an external setting, which was the
2010        cause of this.)
2011    
2012    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2013        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2014        assertion pattern did not match (meaning that the assertion was true), it
2015        was incorrectly treated as false if the SKIP had been reached during the
2016        matching. This also applied to assertions used as conditions.
2017    
2018    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2019        assertion subpattern, including such a pattern used as a condition,
2020        unpredictable results occurred, instead of the error return
2021        PCRE_ERROR_DFA_UITEM.
2022    
2023    5.  The C++ GlobalReplace function was not working like Perl for the special
2024        situation when an empty string is matched. It now does the fancy magic
2025        stuff that is necessary.
2026    
2027    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2028        removed. (These were left over from very, very early versions of PCRE.)
2029    
2030    7.  Some cosmetic changes to the code to make life easier when compiling it
2031        as part of something else:
2032    
2033        (a) Change DEBUG to PCRE_DEBUG.
2034    
2035        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2036            called "current" as "current_branch", to prevent a collision with the
2037            Linux macro when compiled as a kernel module.
2038    
2039        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2040            prevent a collision with the Linux macro when compiled as a kernel
2041            module.
2042    
2043    8.  In pcre_compile() there are some checks for integer overflows that used to
2044        cast potentially large values to (double). This has been changed to that
2045        when building, a check for int64_t is made, and if it is found, it is used
2046        instead, thus avoiding the use of floating point arithmetic. (There is no
2047        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2048        double.
2049    
2050    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2051        2005 (difference between two addresses compared to an unsigned value).
2052    
2053    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2054        custom one, because of the following reported problem in Windows:
2055    
2056          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2057              under Win32.
2058          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2059              therefore missing the function definition.
2060          - The compiler thus generates a "C" signature for the test function.
2061          - The linker fails to find the "C" function.
2062          - PCRE fails to configure if asked to do so against libbz2.
2063    
2064    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2065        messages were output:
2066    
2067          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2068          rerunning libtoolize, to keep the correct libtool macros in-tree.
2069          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2070    
2071        I have done both of these things.
2072    
2073    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2074        most of the time, it *can* run out if it is given a pattern that contains a
2075        runaway infinite recursion. I updated the discussion in the pcrestack man
2076        page.
2077    
2078    13. Now that we have gone to the x.xx style of version numbers, the minor
2079        version may start with zero. Using 08 or 09 is a bad idea because users
2080        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2081        interpreted as invalid octal numbers. I've updated the previous comment in
2082        configure.ac, and also added a check that gives an error if 08 or 09 are
2083        used.
2084    
2085    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2086        causing partial matching to fail when the end of the subject matched \W
2087        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2088    
2089    15. There were some discrepancies between the declarations in pcre_internal.h
2090        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2091        their definitions. The declarations used "const uschar *" and the
2092        definitions used USPTR. Even though USPTR is normally defined as "const
2093        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2094        reported that: "This difference in casting confuses some C++ compilers, for
2095        example, SunCC recognizes above declarations as different functions and
2096        generates broken code for hbpcre." I have changed the declarations to use
2097        USPTR.
2098    
2099    16. GNU libtool is named differently on some systems. The autogen.sh script now
2100        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2101        (FreeBSD).
2102    
2103    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2104        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2105        comment: "Figure out how to create a longlong from a string: strtoll and
2106        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2107        instance, but it only takes 2 args instead of 3!"
2108    
2109    18. A subtle bug concerned with back references has been fixed by a change of
2110        specification, with a corresponding code fix. A pattern such as
2111        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2112        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2113        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2114        same bug. Such groups have to be quantified to be useful, or contained
2115        inside another quantified group. (If there's no repetition, the reference
2116        can never match.) The problem arises because, having left the group and
2117        moved on to the rest of the pattern, a later failure that backtracks into
2118        the group uses the captured value from the final iteration of the group
2119        rather than the correct earlier one. I have fixed this in PCRE by forcing
2120        any group that contains a reference to itself to be an atomic group; that
2121        is, there cannot be any backtracking into it once it has completed. This is
2122        similar to recursive and subroutine calls.
2123    
2124    
2125    Version 8.00 19-Oct-09
2126    ----------------------
2127    
2128    1.  The table for translating pcre_compile() error codes into POSIX error codes
2129        was out-of-date, and there was no check on the pcre_compile() error code
2130        being within the table. This could lead to an OK return being given in
2131        error.
2132    
2133    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2134        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2135        in a Windows environment.
2136    
2137    3.  The pcregrep --count option prints the count for each file even when it is
2138        zero, as does GNU grep. However, pcregrep was also printing all files when
2139        --files-with-matches was added. Now, when both options are given, it prints
2140        counts only for those files that have at least one match. (GNU grep just
2141        prints the file name in this circumstance, but including the count seems
2142        more useful - otherwise, why use --count?) Also ensured that the
2143        combination -clh just lists non-zero counts, with no names.
2144    
2145    4.  The long form of the pcregrep -F option was incorrectly implemented as
2146        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2147        but it seems right to fix it, and I didn't think it was worth preserving
2148        the old behaviour.
2149    
2150    5.  The command line items --regex=pattern and --regexp=pattern were not
2151        recognized by pcregrep, which required --regex pattern or --regexp pattern
2152        (with a space rather than an '='). The man page documented the '=' forms,
2153        which are compatible with GNU grep; these now work.
2154    
2155    6.  No libpcreposix.pc file was created for pkg-config; there was just
2156        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2157    
2158    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2159        when UCP support is not needed, by modifying the Python script that
2160        generates it from Unicode data files. This should not matter if the module
2161        is correctly used as a library, but I received one complaint about 50K of
2162        unwanted data. My guess is that the person linked everything into his
2163        program rather than using a library. Anyway, it does no harm.
2164    
2165    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2166        was a minimum greater than 1 for a wide character in a possessive
2167        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2168        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2169        character. Chaos in the form of incorrect output or a compiling loop could
2170        result.
2171    
2172    9.  The restrictions on what a pattern can contain when partial matching is
2173        requested for pcre_exec() have been removed. All patterns can now be
2174        partially matched by this function. In addition, if there are at least two
2175        slots in the offset vector, the offset of the earliest inspected character
2176        for the match and the offset of the end of the subject are set in them when
2177        PCRE_ERROR_PARTIAL is returned.
2178    
2179    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2180        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2181        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2182        and may be more useful for multi-segment matching.
2183    
2184    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2185        used to be given if ever the end of the subject was reached; now it is
2186        given only if matching could not proceed because another character was
2187        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2188        string "Z", which now yields "no match" instead of "partial match". In the
2189        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2190        final character ended with (*FAIL).
2191    
2192    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2193        if the pattern had a "must contain" character that was already found in the
2194        earlier partial match, unless partial matching was again requested. For
2195        example, with the pattern /dog.(body)?/, the "must contain" character is
2196        "g". If the first part-match was for the string "dog", restarting with
2197        "sbody" failed. This bug has been fixed.
2198    
2199    13. The string returned by pcre_dfa_exec() after a partial match has been
2200        changed so that it starts at the first inspected character rather than the
2201        first character of the match. This makes a difference only if the pattern
2202        starts with a lookbehind assertion or \b or \B (\K is not supported by
2203        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2204        matching functions compatible, and I think it's the right thing to do.
2205    
2206    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2207        so that the demonstration program is easily available in environments where
2208        PCRE has not been installed from source.
2209    
2210    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2211        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2212        library.
2213    
2214    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2215        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2216        is not the first non-POSIX option to be added. Clearly some people find
2217        these options useful.
2218    
2219    17. If a caller to the POSIX matching function regexec() passes a non-zero
2220        value for nmatch with a NULL value for pmatch, the value of
2221        nmatch is forced to zero.
2222    
2223    18. RunGrepTest did not have a test for the availability of the -u option of
2224        the diff command, as RunTest does. It now checks in the same way as
2225        RunTest, and also checks for the -b option.
2226    
2227    19. If an odd number of negated classes containing just a single character
2228        interposed, within parentheses, between a forward reference to a named
2229        subpattern and the definition of the subpattern, compilation crashed with
2230        an internal error, complaining that it could not find the referenced
2231        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2232        [The bug was that it was starting one character too far in when skipping
2233        over the character class, thus treating the ] as data rather than
2234        terminating the class. This meant it could skip too much.]
2235    
2236    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2237        /g option in pcretest when the pattern contains \K, which makes it possible
2238        to have an empty string match not at the start, even when the pattern is
2239        anchored. Updated pcretest and pcredemo to use this option.
2240    
2241    21. If the maximum number of capturing subpatterns in a recursion was greater
2242        than the maximum at the outer level, the higher number was returned, but
2243        with unset values at the outer level. The correct (outer level) value is
2244        now given.
2245    
2246    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2247        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2248        make it do so. The string so far is captured, making this feature
2249        compatible with Perl.
2250    
2251    23. The tests have been re-organized, adding tests 11 and 12, to make it
2252        possible to check the Perl 5.10 features against Perl 5.10.
2253    
2254    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2255        pattern matches a fixed length string. PCRE did not allow this; now it
2256        does. Neither allows recursion.
2257    
2258    25. I finally figured out how to implement a request to provide the minimum
2259        length of subject string that was needed in order to match a given pattern.
2260        (It was back references and recursion that I had previously got hung up
2261        on.) This code has now been added to pcre_study(); it finds a lower bound
2262        to the length of subject needed. It is not necessarily the greatest lower
2263        bound, but using it to avoid searching strings that are too short does give
2264        some useful speed-ups. The value is available to calling programs via
2265        pcre_fullinfo().
2266    
2267    26. While implementing 25, I discovered to my embarrassment that pcretest had
2268        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2269        study optimizations had never been tested with that matching function.
2270        Oops. What is worse, even when it was passed study data, there was a bug in
2271        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2272        were also very few tests of studied patterns with pcre_dfa_exec().
2273    
2274    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2275        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2276        on the other side of the coin, they are no longer allowed to have different
2277        names, because these cannot be distinguished in PCRE, and this has caused
2278        confusion. (This is a difference from Perl.)
2279    
2280    28. When duplicate subpattern names are present (necessarily with different
2281        numbers, as required by 27 above), and a test is made by name in a
2282        conditional pattern, either for a subpattern having been matched, or for
2283        recursion in such a pattern, all the associated numbered subpatterns are
2284        tested, and the overall condition is true if the condition is true for any
2285        one of them. This is the way Perl works, and is also more like the way
2286        testing by number works.
2287    
2288    
2289    Version 7.9 11-Apr-09
2290    ---------------------
2291    
2292    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2293        (pcretest), all targets were linked against these libraries. This included
2294        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2295        libraries. This caused unwanted dependencies to be created. This problem
2296        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2297        pcretest is linked with readline.
2298    
2299    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2300        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2301        moved outside it again, because FALSE and TRUE are already defined in AIX,
2302        but BOOL is not.
2303    
2304    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2305        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2306    
2307    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2308        hyphens) following file names and line numbers when outputting matching
2309        lines. This is not true; no spaces are inserted. I have also clarified the
2310        wording for the --colour (or --color) option.
2311    
2312    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2313        was not coloured; this is different to GNU grep, so I have changed it to be
2314        the same.
2315    
2316    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2317        each matching line was coloured. Now it goes on to look for further matches
2318        of any of the test patterns, which is the same behaviour as GNU grep.
2319    
2320    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2321        doesn't make sense to accept an empty string match in pcregrep, so I have
2322        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2323        seems to be how GNU grep behaves. [But see later change 40 for release
2324        8.33.]
2325    
2326    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2327        start or after a newline", because the conditional assertion was not being
2328        correctly handled. The rule now is that both the assertion and what follows
2329        in the first alternative must satisfy the test.
2330    
2331    9.  If auto-callout was enabled in a pattern with a conditional group whose
2332        condition was an assertion, PCRE could crash during matching, both with
2333        pcre_exec() and pcre_dfa_exec().
2334    
2335    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2336        used for matching.
2337    
2338    11. Unicode property support in character classes was not working for
2339        characters (bytes) greater than 127 when not in UTF-8 mode.
2340    
2341    12. Added the -M command line option to pcretest.
2342    
2343    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2344    
2345    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2346    
2347    16. Added comments and documentation about mis-use of no_arg in the C++
2348        wrapper.
2349    
2350    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2351        from Martin Jerabek that uses macro names for all relevant character and
2352        string constants.
2353    
2354    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2355        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2356        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2357        these, but not everybody uses configure.
2358    
2359    19. A conditional group that had only one branch was not being correctly
2360        recognized as an item that could match an empty string. This meant that an
2361        enclosing group might also not be so recognized, causing infinite looping
2362        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2363        with the subject "ab", where knowledge that the repeated group can match
2364        nothing is needed in order to break the loop.
2365    
2366    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2367        exec(), but without supplying a callout function, matching went wrong.
2368    
2369    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2370        leak if the size of the offset vector was greater than 30. When the vector
2371        is smaller, the saved offsets during recursion go onto a local stack
2372        vector, but for larger vectors malloc() is used. It was failing to free
2373        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2374        error, in fact).
2375    
2376    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2377        heapframe that is used only when UTF-8 support is enabled. This caused no
2378        problem, but was untidy.
2379    
2380    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2381        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2382        included within another project.
2383    
2384    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2385        slightly modified by me:
2386    
2387          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2388              not building pcregrep.
2389    
2390          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2391              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2392    
2393    25. Forward references, both numeric and by name, in patterns that made use of
2394        duplicate group numbers, could behave incorrectly or give incorrect errors,
2395        because when scanning forward to find the reference group, PCRE was not
2396        taking into account the duplicate group numbers. A pattern such as
2397        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2398    
2399    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2400        the feature of a custom pointer more persuasive (as requested by a user).
2401    
2402    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2403        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2404        pre-defined.
2405    
2406    28. Added support for (*UTF8) at the start of a pattern.
2407    
2408    29. Arrange for flags added by the "release type" setting in CMake to be shown
2409        in the configuration summary.
2410    
2411    
2412    Version 7.8 05-Sep-08
2413    ---------------------
2414    
2415    1.  Replaced UCP searching code with optimized version as implemented for Ad
2416        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2417        stage table and inline lookup instead of a function, giving speed ups of 2
2418        to 5 times on some simple patterns that I tested. Permission was given to
2419        distribute the MultiStage2.py script that generates the tables (it's not in
2420        the tarball, but is in the Subversion repository).
2421    
2422    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2423        scripts.
2424    
2425    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2426        a group with a zero qualifier. The result of the study could be incorrect,
2427        or the function might crash, depending on the pattern.
2428    
2429    4.  Caseless matching was not working for non-ASCII characters in back
2430        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2431        It now works when Unicode Property Support is available.
2432    
2433    5.  In pcretest, an escape such as \x{de} in the data was always generating
2434        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2435        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2436        truncation.
2437    
2438    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2439    
2440    7.  Added two (int) casts to pcregrep when printing the difference of two
2441        pointers, in case they are 64-bit values.
2442    
2443    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2444        test 2 if it fails.
2445    
2446    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2447        and a #define of that name to empty if it is not externally set. This is to
2448        allow users of MSVC to set it if necessary.
2449    
2450    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2451        the convenience functions in the pcre_get.c source file.
2452    
2453    11. An option change at the start of a pattern that had top-level alternatives
2454        could cause overwriting and/or a crash. This command provoked a crash in
2455        some environments:
2456    
2457          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2458    
2459        This potential security problem was recorded as CVE-2008-2371.
2460    
2461    12. For a pattern where the match had to start at the beginning or immediately
2462        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2463        pcre_dfa_exec() could read past the end of the passed subject if there was
2464        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2465        pcretest so that it places the subject at the end of its malloc-ed buffer.
2466    
2467    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2468        exec() might read past the end of the data buffer in UTF-8 mode.
2469    
2470    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2471        the data contained the byte 0x85 as part of a UTF-8 character within its
2472        first line. This applied both to normal and DFA matching.
2473    
2474    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2475        /^[^d]*?$/8 failed to match "abc".
2476    
2477    16. Added a missing copyright notice to pcrecpp_internal.h.
2478    
2479    17. Make it more clear in the documentation that values returned from
2480        pcre_exec() in ovector are byte offsets, not character counts.
2481    
2482    18. Tidied a few places to stop certain compilers from issuing warnings.
2483    
2484    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2485        supplied by Stefan Weber. I made a further small update for 7.8 because
2486        there is a change of source arrangements: the pcre_searchfuncs.c module is
2487        replaced by pcre_ucd.c.
2488    
2489    
2490    Version 7.7 07-May-08
2491    ---------------------
2492    
2493    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2494        a string to a long long, pretend we don't even have a long long." This is
2495        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2496    
2497    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2498        pre-7.6 versions, which defined a global no_arg variable instead of putting
2499        it in the RE class. (See also #8 below.)
2500    
2501    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2502        Lopes.
2503    
2504    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2505    
2506        (1) The include/exclude patterns were being applied to the whole pathnames
2507            of files, instead of just to the final components.
2508    
2509        (2) If there was more than one level of directory, the subdirectories were
2510            skipped unless they satisfied the include/exclude conditions. This is
2511            inconsistent with GNU grep (and could even be seen as contrary to the
2512            pcregrep specification - which I improved to make it absolutely clear).
2513            The action now is always to scan all levels of directory, and just
2514            apply the include/exclude patterns to regular files.
2515    
2516    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2517        --exclude_dir in the tests to avoid scanning .svn directories.
2518    
2519    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2520        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2521        doesn't support NULs in patterns.
2522    
2523    7.  Added some missing "const"s to declarations of static tables in
2524        pcre_compile.c and pcre_dfa_exec.c.
2525    
2526    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2527        caused by fix #2  above. (Subsequently also a second patch to fix the
2528        first patch. And a third patch - this was a messy problem.)
2529    
2530    9.  Applied Craig's patch to remove the use of push_back().
2531    
2532    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2533        matching function regexec().
2534    
2535    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2536        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2537        references. PCRE supports relative numbers with this syntax (I don't think
2538        Oniguruma does).
2539    
2540    12. Previously, a group with a zero repeat such as (...){0} was completely
2541        omitted from the compiled regex. However, this means that if the group
2542        was called as a subroutine from elsewhere in the pattern, things went wrong
2543        (an internal error was given). Such groups are now left in the compiled
2544        pattern, with a new opcode that causes them to be skipped at execution
2545        time.
2546    
2547    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2548        to the way PCRE behaves:
2549    
2550        (a) A lone ] character is dis-allowed (Perl treats it as data).
2551    
2552        (b) A back reference to an unmatched subpattern matches an empty string
2553            (Perl fails the current match path).
2554    
2555        (c) A data ] in a character class must be notated as \] because if the
2556            first data character in a class is ], it defines an empty class. (In
2557            Perl it is not possible to have an empty class.) The empty class []
2558            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2559            The negative empty class [^] matches any one character, independently
2560            of the DOTALL setting.
2561    
2562    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2563        non-existent subpattern following a character class starting with ']' and
2564        containing () gave an internal compiling error instead of "reference to
2565        non-existent subpattern". Fortunately, when the pattern did exist, the
2566        compiled code was correct. (When scanning forwards to check for the
2567        existence of the subpattern, it was treating the data ']' as terminating
2568        the class, so got the count wrong. When actually compiling, the reference
2569        was subsequently set up correctly.)
2570    
2571    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2572        it was being rejected as not supported by pcre_dfa_exec(), even though
2573        other assertions are supported. I have made pcre_dfa_exec() support
2574        (*FAIL).
2575    
2576    16. The implementation of 13c above involved the invention of a new opcode,
2577        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2578        cannot be changed at match time, I realized I could make a small
2579        improvement to matching performance by compiling OP_ALLANY instead of
2580        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2581        on the OP_ANY path.
2582    
2583    17. Compiling pcretest on Windows with readline support failed without the
2584        following two fixes: (1) Make the unistd.h include conditional on
2585        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2586    
2587    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2588        ncurses library to be included for pcretest when ReadLine support is
2589        requested, but also to allow for it to be overridden. This patch came from
2590        Daniel Bergström.
2591    
2592    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2593        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2594        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2595        spotting this.
2596    
2597    
2598    Version 7.6 28-Jan-08
2599    ---------------------
2600    
2601    1.  A character class containing a very large number of characters with
2602        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2603        overflow.
2604    
2605    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2606        HAVE_LONG_LONG is not defined.
2607    
2608    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2609        bring them up to date and include new features. This patch includes:
2610    
2611        - Fixed PH's badly added libz and libbz2 support.
2612        - Fixed a problem with static linking.
2613        - Added pcredemo. [But later removed - see 7 below.]
2614        - Fixed dftables problem and added an option.
2615        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2616            HAVE_LONG_LONG.
2617        - Added readline support for pcretest.
2618        - Added an listing of the option settings after cmake has run.
2619    
2620    4.  A user submitted a patch to Makefile that makes it easy to create
2621        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2622        Makefile.am that cause it to include this special target, without
2623        affecting anything else. Note that the same mingw target plus all
2624        the other distribution libraries and programs are now supported
2625        when configuring with CMake (see 6 below) instead of with
2626        Configure/Make.
2627    
2628    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2629        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2630        exported in the Windows port". It has not yet been confirmed that the patch
2631        solves the problem, but it does no harm.
2632    
2633    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2634        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2635        with CMake, and also correct the comment about stack recursion.
2636    
2637    7.  Remove the automatic building of pcredemo from the ./configure system and
2638        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2639        of a program that users should build themselves after PCRE is installed, so
2640        building it automatically is not really right. What is more, it gave
2641        trouble in some build environments.
2642    
2643    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2644    
2645    
2646    Version 7.5 10-Jan-08
2647    ---------------------
2648    
2649    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2650        values in parens when parsing an RE using the C++ wrapper."
2651    
2652    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2653        Characters greater than 255 were excluded from the class instead of being
2654        included.
2655    
2656    3.  The same bug as (2) above applied to negated POSIX classes such as
2657        [:^space:].
2658    
2659    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2660        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2661        I have changed it.
2662    
2663    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2664        first named subpattern) and a construct such as (?&a) would reference the
2665        first named subpattern whose name started with "a" (in other words, the
2666        length check was missing). Both these problems are fixed. "Subpattern name
2667        expected" is now given for (?&) (a zero-length name), and this patch also
2668        makes it give the same error for \k'' (previously it complained that that
2669        was a reference to a non-existent subpattern).
2670    
2671    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2672        this is right because (?- can be followed by option settings as well as by
2673        digits. I have, however, made the messages clearer.
2674    
2675    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2676        than the number used in the conditional) now cause a compile-time error.
2677        This is actually not compatible with Perl, which accepts such patterns, but
2678        treats the conditional as always being FALSE (as PCRE used to), but it
2679        seems to me that giving a diagnostic is better.
2680    
2681    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2682        and messages.
2683    
2684    9.  Fix two occurrences of "backslash" in comments that should have been
2685        "backspace".
2686    
2687    10. Remove two redundant lines of code that can never be obeyed (their function
2688        was moved elsewhere).
2689    
2690    11. The program that makes PCRE's Unicode character property table had a bug
2691        which caused it to generate incorrect table entries for sequences of
2692        characters that have the same character type, but are in different scripts.
2693        It amalgamated them into a single range, with the script of the first of
2694        them. In other words, some characters were in the wrong script. There were
2695        thirteen such cases, affecting characters in the following ranges:
2696    
2697          U+002b0 - U+002c1
2698          U+0060c - U+0060d
2699          U+0061e - U+00612
2700          U+0064b - U+0065e
2701          U+0074d - U+0076d
2702          U+01800 - U+01805
2703          U+01d00 - U+01d77
2704          U+01d9b - U+01dbf
2705          U+0200b - U+0200f
2706          U+030fc - U+030fe
2707          U+03260 - U+0327f
2708          U+0fb46 - U+0fbb1
2709          U+10450 - U+1049d
2710    
2711    12. The -o option (show only the matching part of a line) for pcregrep was not
2712        compatible with GNU grep in that, if there was more than one match in a
2713        line, it showed only the first of them. It now behaves in the same way as
2714        GNU grep.
2715    
2716    13. If the -o and -v options were combined for pcregrep, it printed a blank
2717        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2718        does the same. The return code can be used to tell if there were any
2719        non-matching lines.
2720    
2721    14. Added --file-offsets and --line-offsets to pcregrep.
2722    
2723    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2724        infinitely looping recursion. The bug was that positive lookaheads were not
2725        being skipped when checking for a possible empty match (negative lookaheads
2726        and both kinds of lookbehind were skipped).
2727    
2728    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2729        inclusion of <windows.h> to before rather than after the definition of
2730        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2731    
2732    17. Specifying a possessive quantifier with a specific limit for a Unicode
2733        character property caused pcre_compile() to compile bad code, which led at
2734        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2735        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2736        caused the error; without that there was no problem.
2737    
2738    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2739    
2740    19. Added --enable-pcretest-libreadline.
2741    
2742    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2743        RE::GlobalReplace(). As a result, the number of replacements returned was
2744        double what it should be. I removed one of the increments, but Craig sent a
2745        later patch that removed the other one (the right fix) and added unit tests
2746        that check the return values (which was not done before).
2747    
2748    21. Several CMake things:
2749    
2750        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2751            the names libpcre and libpcreposix, not just pcre and pcreposix.
2752    
2753        (2) The above change means that pcretest and pcregrep are now correctly
2754            linked with the newly-built libraries, not previously installed ones.
2755    
2756        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2757    
2758    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2759        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2760        UTF-8 newline character). The key issue is that the pattern starts .*;
2761        this means that the match must be either at the beginning, or after a
2762        newline. The bug was in the code for advancing after a failed match and
2763        checking that the new position followed a newline. It was not taking
2764        account of UTF-8 characters correctly.
2765    
2766    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2767        character classes. PCRE was not treating the sequence [:...:] as a
2768        character class unless the ... were all letters. Perl, however, seems to
2769        allow any characters between [: and :], though of course it rejects as
2770        unknown any "names" that contain non-letters, because all the known class
2771        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2772        for example, whereas PCRE did not - it did not recognize a POSIX character
2773        class. This seemed a bit dangerous, so the code has been changed to be
2774        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2775        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2776        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2777        Perl does, and where it didn't before.
2778    
2779    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2780        Windows environments %n is disabled by default.
2781    
2782    
2783    Version 7.4 21-Sep-07
2784    ---------------------
2785    
2786    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2787        means that a class such as [\s] counted as "explicit reference to CR or
2788        LF". That isn't really right - the whole point of the change was to try to
2789        help when there was an actual mention of one of the two characters. So now
2790        the change happens only if \r or \n (or a literal CR or LF) character is
2791        encountered.
2792    
2793    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2794        of both had grown to the point where there were only 3 bits left.
2795        Fortunately, there was spare space in the data structure, and so I have
2796        moved the internal flags into a new 16-bit field to free up more option
2797        bits.
2798    
2799    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2800        but did not set the internal JCHANGED flag - either of these is enough to
2801        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2802        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2803        start sets both bits.
2804    
2805    4.  Added options (at build time, compile time, exec time) to change \R from
2806        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2807    
2808    5.  doc/pcresyntax.html was missing from the distribution.
2809    
2810    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2811        compatibility, even though it is no longer used.
2812    
2813    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2814        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2815        windows.h file is present (where different names are used). [This was
2816        reversed later after testing - see 16 below.]
2817    
2818    8.  Changed all #include <config.h> to #include "config.h". There were also
2819        some further <pcre.h> cases that I changed to "pcre.h".
2820    
2821    9.  When pcregrep was used with the --colour option, it missed the line ending
2822        sequence off the lines that it output.
2823    
2824    10. It was pointed out to me that arrays of string pointers cause lots of
2825        relocations when a shared library is dynamically loaded. A technique of
2826        using a single long string with a table of offsets can drastically reduce
2827        these. I have refactored PCRE in four places to do this. The result is
2828        dramatic:
2829    
2830          Originally:                          290
2831          After changing UCP table:            187
2832          After changing error message table:   43
2833          After changing table of "verbs"       36
2834          After changing table of Posix names   22
2835    
2836        Thanks to the folks working on Gregex for glib for this insight.
2837    
2838    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2839        unicode-properties was also set.
2840    
2841    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2842    
2843    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2844        checked only for CRLF.
2845    
2846    14. Added casts to pcretest.c to avoid compiler warnings.
2847    
2848    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2849    
2850    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2851        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2852        entirely. This removes changes made in 7 above.
2853    
2854    17. The CMake files have been updated, and there is now more information about
2855        building with CMake in the NON-UNIX-USE document.
2856    
2857    
2858    Version 7.3 28-Aug-07
2859  ---------------------  ---------------------
2860    
2861   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 98  Version 7.3 09-Aug-07 Line 2952  Version 7.3 09-Aug-07
2952      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2953      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2954      does: it's just the validity check that is more restrictive.      does: it's just the validity check that is more restrictive.
2955    
2956  16. Inserted checks for integer overflows during escape sequence (backslash)  16. Inserted checks for integer overflows during escape sequence (backslash)
2957      processing, and also fixed erroneous offset values for syntax errors during      processing, and also fixed erroneous offset values for syntax errors during
2958      backslash processing.      backslash processing.
2959    
2960  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2961      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2962    
2963    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2964        caused an overrun.
2965    
2966    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2967        something other than just ASCII characters) inside a group that had an
2968        unlimited repeat caused a loop at compile time (while checking to see
2969        whether the group could match an empty string).
2970    
2971    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2972        [\P{Any}] did so. (Error in the code for printing property names.)
2973    
2974    21. An orphan \E inside a character class could cause a crash.
2975    
2976    22. A repeated capturing bracket such as (A)? could cause a wild memory
2977        reference during compilation.
2978    
2979    23. There are several functions in pcre_compile() that scan along a compiled
2980        expression for various reasons (e.g. to see if it's fixed length for look
2981        behind). There were bugs in these functions when a repeated \p or \P was
2982        present in the pattern. These operators have additional parameters compared
2983        with \d, etc, and these were not being taken into account when moving along
2984        the compiled data. Specifically:
2985    
2986        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2987            length.
2988    
2989        (b) An item such as \pL+ within a repeated group could cause crashes or
2990            loops.
2991    
2992        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2993            "reference to non-existent subpattern" error.
2994    
2995        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2996    
2997    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2998        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2999    
3000    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3001    
3002    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3003        character were causing crashes (broken optimization).
3004    
3005    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3006        \p or \P) caused a compile-time loop.
3007    
3008    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3009        break. For example, the unstudied pattern [\r\n]A does not match the string
3010        "\r\nA" because change 7.0/46 below moves the current point on by two
3011        characters after failing to match at the start. However, the pattern \nA
3012        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3013        the same is true. There doesn't seem any very clean way out of this, but
3014        what I have chosen to do makes the common cases work: PCRE now takes note
3015        of whether there can be an explicit match for \r or \n anywhere in the
3016        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3017        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3018        pattern has explicit CR or LF references.
3019    
3020    29. Added (*CR) etc for changing newline setting at start of pattern.
3021    
3022    
3023  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.214  
changed lines
  Added in v.1566

  ViewVC Help
Powered by ViewVC 1.1.5