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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.447  
changed lines
  Added in v.1584

  ViewVC Help
Powered by ViewVC 1.1.5