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

Diff of /code/trunk/ChangeLog

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

revision 221 by ph10, Fri Aug 17 09:25:08 2007 UTC revision 1576 by ph10, Fri Jul 17 15:15:18 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 17-Aug-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.38 xx-xxx-xxxx
8    ------------------------
9    
10    1.  If a group that contained a recursive back reference also contained a
11        forward reference subroutine call followed by a non-forward-reference
12        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13        compile correct code, leading to undefined behaviour or an internally
14        detected error. This bug was discovered by the LLVM fuzzer.
15    
16    2.  Quantification of certain items (e.g. atomic back references) could cause
17        incorrect code to be compiled when recursive forward references were
18        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19        This bug was discovered by the LLVM fuzzer.
20    
21    3.  A repeated conditional group whose condition was a reference by name caused
22        a buffer overflow if there was more than one group with the given name.
23        This bug was discovered by the LLVM fuzzer.
24    
25    4.  A recursive back reference by name within a group that had the same name as
26        another group caused a buffer overflow. For example:
27        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28    
29    5.  A forward reference by name to a group whose number is the same as the
30        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31        a buffer overflow at compile time. This bug was discovered by the LLVM
32        fuzzer.
33    
34    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
35        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36    
37    7.  Another buffer overflow bug involved duplicate named groups with a
38        reference between their definition, with a group that reset capture
39        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40        fixed by always allowing for more memory, even if not needed. (A proper fix
41        is implemented in PCRE2, but it involves more refactoring.)
42    
43    8.  There was no check for integer overflow in subroutine calls such as (?123).
44    
45    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
46        being treated as a literal 'l' instead of causing an error.
47    
48    10. There was a buffer overflow if pcre_exec() was called with an ovector of
49        size 1. This bug was found by american fuzzy lop.
50    
51    11. If a non-capturing group containing a conditional group that could match
52        an empty string was repeated, it was not identified as matching an empty
53        string itself. For example: /^(?:(?(1)x|)+)+$()/.
54    
55    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    
85    Version 8.37 28-April-2015
86    --------------------------
87    
88    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
89        for those parentheses to be closed with whatever has been captured so far.
90        However, it was failing to mark any other groups between the hightest
91        capture so far and the currrent group as "unset". Thus, the ovector for
92        those groups contained whatever was previously there. An example is the
93        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
94    
95    2.  If an assertion condition was quantified with a minimum of zero (an odd
96        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
97    
98    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
99        unrecognized modifier, a crash could occur.
100    
101    4.  An attempt to do global matching in pcretest with a zero-length ovector
102        caused a crash.
103    
104    5.  Fixed a memory leak during matching that could occur for a subpattern
105        subroutine call (recursive or otherwise) if the number of captured groups
106        that had to be saved was greater than ten.
107    
108    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
109        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
110        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
111    
112    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
113        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
114    
115    8. When a pattern is compiled, it remembers the highest back reference so that
116       when matching, if the ovector is too small, extra memory can be obtained to
117       use instead. A conditional subpattern whose condition is a check on a
118       capture having happened, such as, for example in the pattern
119       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
120       setting the highest backreference number. This mattered only if pcre_exec()
121       was called with an ovector that was too small to hold the capture, and there
122       was no other kind of back reference (a situation which is probably quite
123       rare). The effect of the bug was that the condition was always treated as
124       FALSE when the capture could not be consulted, leading to a incorrect
125       behaviour by pcre_exec(). This bug has been fixed.
126    
127    9. A reference to a duplicated named group (either a back reference or a test
128       for being set in a conditional) that occurred in a part of the pattern where
129       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
130       to be incorrectly calculated, leading to overwriting.
131    
132    10. A mutually recursive set of back references such as (\2)(\1) caused a
133        segfault at study time (while trying to find the minimum matching length).
134        The infinite loop is now broken (with the minimum length unset, that is,
135        zero).
136    
137    11. If an assertion that was used as a condition was quantified with a minimum
138        of zero, matching went wrong. In particular, if the whole group had
139        unlimited repetition and could match an empty string, a segfault was
140        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
141        assertions to be quantified, but not if they are being used as conditions,
142        so the above pattern is faulted by Perl. PCRE has now been changed so that
143        it also rejects such patterns.
144    
145    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
146        failed to allow the zero-repeat case if pcre2_exec() was called with an
147        ovector too small to capture the group.
148    
149    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
150        Red Hat Product Security:
151    
152        (a) A crash if /K and /F were both set with the option to save the compiled
153        pattern.
154    
155        (b) Another crash if the option to print captured substrings in a callout
156        was combined with setting a null ovector, for example \O\C+ as a subject
157        string.
158    
159    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
160        forward reference repeated a large (but limited) number of times within a
161        repeated outer group that has a zero minimum quantifier, caused incorrect
162        code to be compiled, leading to the error "internal error:
163        previously-checked referenced subpattern not found" when an incorrect
164        memory address was read. This bug was reported as "heap overflow",
165        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
166        CVE-2015-2325.
167    
168    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
169        call within a group that also contained a recursive back reference caused
170        incorrect code to be compiled. This bug was reported as "heap overflow",
171        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
172        number CVE-2015-2326.
173    
174    24. Computing the size of the JIT read-only data in advance has been a source
175        of various issues, and new ones are still appear unfortunately. To fix
176        existing and future issues, size computation is eliminated from the code,
177        and replaced by on-demand memory allocation.
178    
179    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
180        adjacent to the end of the range, and the range contained characters with
181        more than one other case, caused incorrect behaviour when compiled in UTF
182        mode. In that example, the range a-j was left out of the class.
183    
184    26. Fix JIT compilation of conditional blocks, which assertion
185        is converted to (*FAIL). E.g: /(?(?!))/.
186    
187    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
188        discovered by the LLVM fuzzer.
189    
190    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
191        when this assertion was used as a condition, for example (?(?!)a|b). In
192        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
193        error about an unsupported item.
194    
195    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
196        possessification code could take exponential time to complete. A recursion
197        depth limit of 1000 has been imposed to limit the resources used by this
198        optimization.
199    
200    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
201        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
202        because \S ensures they are all in the class. The code for doing this was
203        interacting badly with the code for computing the amount of space needed to
204        compile the pattern, leading to a buffer overflow. This bug was discovered
205        by the LLVM fuzzer.
206    
207    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
208        other kinds of group caused stack overflow at compile time. This bug was
209        discovered by the LLVM fuzzer.
210    
211    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
212        between a subroutine call and its quantifier was incorrectly compiled,
213        leading to buffer overflow or other errors. This bug was discovered by the
214        LLVM fuzzer.
215    
216    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
217        assertion after (?(. The code was failing to check the character after
218        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
219        was discovered by the LLVM fuzzer.
220    
221    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
222        a fixed maximum following a group that contains a subroutine reference was
223        incorrectly compiled and could trigger buffer overflow. This bug was
224        discovered by the LLVM fuzzer.
225    
226    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
227        caused a stack overflow instead of the diagnosis of a non-fixed length
228        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
229    
230    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
231        (e.g. /(?<=\Ka)/) could make pcregrep loop.
232    
233    37. There was a similar problem to 36 in pcretest for global matches.
234    
235    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
236        and a subsequent item in the pattern caused a non-match, backtracking over
237        the repeated \X did not stop, but carried on past the start of the subject,
238        causing reference to random memory and/or a segfault. There were also some
239        other cases where backtracking after \C could crash. This set of bugs was
240        discovered by the LLVM fuzzer.
241    
242    39. The function for finding the minimum length of a matching string could take
243        a very long time if mutual recursion was present many times in a pattern,
244        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
245        method has been implemented. This infelicity was discovered by the LLVM
246        fuzzer.
247    
248    40. Static linking against the PCRE library using the pkg-config module was
249        failing on missing pthread symbols.
250    
251    
252    Version 8.36 26-September-2014
253    ------------------------------
254    
255    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
256        -Wmissing-field-initializers and -Wunused-parameter.
257    
258    2.  The tests for quantifiers being too big (greater than 65535) were being
259        applied after reading the number, and stupidly assuming that integer
260        overflow would give a negative number. The tests are now applied as the
261        numbers are read.
262    
263    3.  Tidy code in pcre_exec.c where two branches that used to be different are
264        now the same.
265    
266    4.  The JIT compiler did not generate match limit checks for certain
267        bracketed expressions with quantifiers. This may lead to exponential
268        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
269        issue should be resolved now.
270    
271    5.  Fixed an issue, which occures when nested alternatives are optimized
272        with table jumps.
273    
274    6.  Inserted two casts and changed some ints to size_t in the light of some
275        reported 64-bit compiler warnings (Bugzilla 1477).
276    
277    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
278        an empty string, which sometimes were behaving incorrectly in the
279        interpreter (though correctly in the JIT matcher). This pcretest input is
280        an example:
281    
282          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
283          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
284    
285        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
286        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
287        for an empty string was breaking the inner loop and carrying on at a lower
288        level, when possessive repeated groups should always return to a higher
289        level as they have no backtrack points in them. The empty string test now
290        occurs at the outer level.
291    
292    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
293        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
294    
295    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
296        doesn't).
297    
298    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
299        to hold a pattern to be incorrectly computed (too small) when there were
300        named back references to duplicated names. This could cause "internal
301        error: code overflow" or "double free or corruption" or other memory
302        handling errors.
303    
304    11. When named subpatterns had the same prefixes, back references could be
305        confused. For example, in this pattern:
306    
307          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
308    
309        the reference to 'Name' was incorrectly treated as a reference to a
310        duplicate name.
311    
312    12. A pattern such as /^s?c/mi8 where the optional character has more than
313        one "other case" was incorrectly compiled such that it would only try to
314        match starting at "c".
315    
316    13. When a pattern starting with \s was studied, VT was not included in the
317        list of possible starting characters; this should have been part of the
318        8.34/18 patch.
319    
320    14. If a character class started [\Qx]... where x is any character, the class
321        was incorrectly terminated at the ].
322    
323    15. If a pattern that started with a caseless match for a character with more
324        than one "other case" was studied, PCRE did not set up the starting code
325        unit bit map for the list of possible characters. Now it does. This is an
326        optimization improvement, not a bug fix.
327    
328    16. The Unicode data tables have been updated to Unicode 7.0.0.
329    
330    17. Fixed a number of memory leaks in pcregrep.
331    
332    18. Avoid a compiler warning (from some compilers) for a function call with
333        a cast that removes "const" from an lvalue by using an intermediate
334        variable (to which the compiler does not object).
335    
336    19. Incorrect code was compiled if a group that contained an internal recursive
337        back reference was optional (had quantifier with a minimum of zero). This
338        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
339        caused segmentation faults because of stack overflows at compile time.
340    
341    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
342        group that is quantified with an indefinite repeat, caused a compile-time
343        loop which used up all the system stack and provoked a segmentation fault.
344        This was not the same bug as 19 above.
345    
346    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
347        Patch by Mike Frysinger.
348    
349    
350    Version 8.35 04-April-2014
351    --------------------------
352    
353    1.  A new flag is set, when property checks are present in an XCLASS.
354        When this flag is not set, PCRE can perform certain optimizations
355        such as studying these XCLASS-es.
356    
357    2.  The auto-possessification of character sets were improved: a normal
358        and an extended character set can be compared now. Furthermore
359        the JIT compiler optimizes more character set checks.
360    
361    3.  Got rid of some compiler warnings for potentially uninitialized variables
362        that show up only when compiled with -O2.
363    
364    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
365        of a match later then the end of the match. The pcretest program was not
366        handling the case sensibly - it was outputting from the start to the next
367        binary zero. It now reports this situation in a message, and outputs the
368        text from the end to the start.
369    
370    5.  Fast forward search is improved in JIT. Instead of the first three
371        characters, any three characters with fixed position can be searched.
372        Search order: first, last, middle.
373    
374    6.  Improve character range checks in JIT. Characters are read by an inprecise
375        function now, which returns with an unknown value if the character code is
376        above a certain threshold (e.g: 256). The only limitation is that the value
377        must be bigger than the threshold as well. This function is useful when
378        the characters above the threshold are handled in the same way.
379    
380    7.  The macros whose names start with RAWUCHAR are placeholders for a future
381        mode in which only the bottom 21 bits of 32-bit data items are used. To
382        make this more memorable for those maintaining the code, the names have
383        been changed to start with UCHAR21, and an extensive comment has been added
384        to their definition.
385    
386    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
387        to the export list in Makefile.am (they were accidentally omitted from the
388        8.34 tarball).
389    
390    9.  The informational output from pcretest used the phrase "starting byte set"
391        which is inappropriate for the 16-bit and 32-bit libraries. As the output
392        for "first char" and "need char" really means "non-UTF-char", I've changed
393        "byte" to "char", and slightly reworded the output. The documentation about
394        these values has also been (I hope) clarified.
395    
396    10. Another JIT related optimization: use table jumps for selecting the correct
397        backtracking path, when more than four alternatives are present inside a
398        bracket.
399    
400    11. Empty match is not possible, when the minimum length is greater than zero,
401        and there is no \K in the pattern. JIT should avoid empty match checks in
402        such cases.
403    
404    12. In a caseless character class with UCP support, when a character with more
405        than one alternative case was not the first character of a range, not all
406        the alternative cases were added to the class. For example, s and \x{17f}
407        are both alternative cases for S: the class [RST] was handled correctly,
408        but [R-T] was not.
409    
410    13. The configure.ac file always checked for pthread support when JIT was
411        enabled. This is not used in Windows, so I have put this test inside a
412        check for the presence of windows.h (which was already tested for).
413    
414    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
415        The algorithm provides a way to skip certain starting offsets, and usually
416        faster than linear prefix searches.
417    
418    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
419        as for 'fr_FR' and 'french'. For some reason, however, it then used the
420        Windows-specific input and output files, which have 'french' screwed in.
421        So this could never have worked. One of the problems with locales is that
422        they aren't always the same. I have now updated RunTest so that it checks
423        the output of the locale test (test 3) against three different output
424        files, and it allows the test to pass if any one of them matches. With luck
425        this should make the test pass on some versions of Solaris where it was
426        failing. Because of the uncertainty, the script did not used to stop if
427        test 3 failed; it now does. If further versions of a French locale ever
428        come to light, they can now easily be added.
429    
430    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
431        there was a message during ./configure, but it did not stop. This now
432        provokes an error. The invalid example in README has been corrected.
433        If a value less than the minimum is given, the minimum value has always
434        been used, but now a warning is given.
435    
436    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
437        was a bug in the test system, which is now fixed. Also, the list of various
438        configurations that are tested for each release did not have one with both
439        16/32 bits and --enable-bar-anycrlf. It now does.
440    
441    18. pcretest was missing "-C bsr" for displaying the \R default setting.
442    
443    19. Little endian PowerPC systems are supported now by the JIT compiler.
444    
445    20. The fast forward newline mechanism could enter to an infinite loop on
446        certain invalid UTF-8 input. Although we don't support these cases
447        this issue can be fixed by a performance optimization.
448    
449    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
450        not take account if existing stack usage. There is now a new global
451        variable called pcre_stack_guard that can be set to point to an external
452        function to check stack availability. It is called at the start of
453        processing every parenthesized group.
454    
455    22. A typo in the code meant that in ungreedy mode the max/min qualifier
456        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
457        not match "ab".
458    
459    23. When UTF was disabled, the JIT program reported some incorrect compile
460        errors. These messages are silenced now.
461    
462    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
463        compiler.
464    
465    25. Change all the temporary files used in RunGrepTest to be different to those
466        used by RunTest so that the tests can be run simultaneously, for example by
467        "make -j check".
468    
469    
470    Version 8.34 15-December-2013
471    -----------------------------
472    
473    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
474        executable memory. Patch inspired by Carsten Klein.
475    
476    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
477        this macro is never tested and has no effect, because the work to support
478        coverage involves only compiling and linking options and special targets in
479        the Makefile. The comment in config.h implied that defining the macro would
480        enable coverage support, which is totally false. There was also support for
481        setting this macro in the CMake files (my fault, I just copied it from
482        configure). SUPPORT_GCOV has now been removed.
483    
484    3.  Make a small performance improvement in strlen16() and strlen32() in
485        pcretest.
486    
487    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
488        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
489        these cases). There was also one in pcretest.c.
490    
491    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
492    
493    6.  In UTF mode, the code for checking whether a group could match an empty
494        string (which is used for indefinitely repeated groups to allow for
495        breaking an infinite loop) was broken when the group contained a repeated
496        negated single-character class with a character that occupied more than one
497        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
498        UTF-8 mode). The effect was undefined: the group might or might not be
499        deemed as matching an empty string, or the program might have crashed.
500    
501    7.  The code for checking whether a group could match an empty string was not
502        recognizing that \h, \H, \v, \V, and \R must match a character.
503    
504    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
505        an empty string. If it can, pcretest shows this in its information output.
506    
507    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
508        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
509        matched by pcre_exec() without using JIT:
510    
511        (a) If the rest of the pattern did not match after a maximal run of
512            grapheme clusters, the code for backing up to try with fewer of them
513            did not always back up over a full grapheme when characters that do not
514            have the modifier quality were involved, e.g. Hangul syllables.
515    
516        (b) If the match point in a subject started with modifier character, and
517            there was no match, the code could incorrectly back up beyond the match
518            point, and potentially beyond the first character in the subject,
519            leading to a segfault or an incorrect match result.
520    
521    10. A conditional group with an assertion condition could lead to PCRE
522        recording an incorrect first data item for a match if no other first data
523        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
524        first data item, and therefore matched "ca" after "c" instead of at the
525        start.
526    
527    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
528        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
529    
530    12. The source of pcregrep now includes z/OS-specific code so that it can be
531        compiled for z/OS as part of the special z/OS distribution.
532    
533    13. Added the -T and -TM options to pcretest.
534    
535    14. The code in pcre_compile.c for creating the table of named capturing groups
536        has been refactored. Instead of creating the table dynamically during the
537        actual compiling pass, the information is remembered during the pre-compile
538        pass (on the stack unless there are more than 20 named groups, in which
539        case malloc() is used) and the whole table is created before the actual
540        compile happens. This has simplified the code (it is now nearly 150 lines
541        shorter) and prepared the way for better handling of references to groups
542        with duplicate names.
543    
544    15. A back reference to a named subpattern when there is more than one of the
545        same name now checks them in the order in which they appear in the pattern.
546        The first one that is set is used for the reference. Previously only the
547        first one was inspected. This change makes PCRE more compatible with Perl.
548    
549    16. Unicode character properties were updated from Unicode 6.3.0.
550    
551    17. The compile-time code for auto-possessification has been refactored, based
552        on a patch by Zoltan Herczeg. It now happens after instead of during
553        compilation. The code is cleaner, and more cases are handled. The option
554        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
555        options in pcretest are provided to set it. It can also be set by
556        (*NO_AUTO_POSSESS) at the start of a pattern.
557    
558    18. The character VT has been added to the default ("C" locale) set of
559        characters that match \s and are generally treated as white space,
560        following this same change in Perl 5.18. There is now no difference between
561        "Perl space" and "POSIX space". Whether VT is treated as white space in
562        other locales depends on the locale.
563    
564    19. The code for checking named groups as conditions, either for being set or
565        for being recursed, has been refactored (this is related to 14 and 15
566        above). Processing unduplicated named groups should now be as fast at
567        numerical groups, and processing duplicated groups should be faster than
568        before.
569    
570    20. Two patches to the CMake build system, by Alexander Barkov:
571    
572          (1) Replace the "source" command by "." in CMakeLists.txt because
573              "source" is a bash-ism.
574    
575          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
576              without these the CMake build does not work on Solaris.
577    
578    21. Perl has changed its handling of \8 and \9. If there is no previously
579        encountered capturing group of those numbers, they are treated as the
580        literal characters 8 and 9 instead of a binary zero followed by the
581        literals. PCRE now does the same.
582    
583    22. Following Perl, added \o{} to specify codepoints in octal, making it
584        possible to specify values greater than 0777 and also making them
585        unambiguous.
586    
587    23. Perl now gives an error for missing closing braces after \x{... instead of
588        treating the string as literal. PCRE now does the same.
589    
590    24. RunTest used to grumble if an inappropriate test was selected explicitly,
591        but just skip it when running all tests. This make it awkward to run ranges
592        of tests when one of them was inappropriate. Now it just skips any
593        inappropriate tests, as it always did when running all tests.
594    
595    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
596        character types such as \d or \w, too many callouts were inserted, and the
597        data that they returned was rubbish.
598    
599    26. In UCP mode, \s was not matching two of the characters that Perl matches,
600        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
601        were matched by \h. The code has now been refactored so that the lists of
602        the horizontal and vertical whitespace characters used for \h and \v (which
603        are defined only in one place) are now also used for \s.
604    
605    27. Add JIT support for the 64 bit TileGX architecture.
606        Patch by Jiong Wang (Tilera Corporation).
607    
608    28. Possessive quantifiers for classes (both explicit and automatically
609        generated) now use special opcodes instead of wrapping in ONCE brackets.
610    
611    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
612        (because it's meaningless), this was not happening when PCRE_CASELESS was
613        set. Not wrong, but inefficient.
614    
615    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
616        properties for \w, \d, etc) is present in a test regex. Otherwise if the
617        test contains no characters greater than 255, Perl doesn't realise it
618        should be using Unicode semantics.
619    
620    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
621        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
622        does in Unicode mode.
623    
624    32. Added the "forbid" facility to pcretest so that putting tests into the
625        wrong test files can sometimes be quickly detected.
626    
627    33. There is now a limit (default 250) on the depth of nesting of parentheses.
628        This limit is imposed to control the amount of system stack used at compile
629        time. It can be changed at build time by --with-parens-nest-limit=xxx or
630        the equivalent in CMake.
631    
632    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
633        errors. Perl warns for these when in warning mode, but PCRE has no facility
634        for giving warnings.
635    
636    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
637        However, this was not working for (?!) because it is optimized to (*FAIL),
638        for which PCRE does not allow quantifiers. The optimization is now disabled
639        when a quantifier follows (?!). I can't see any use for this, but it makes
640        things uniform.
641    
642    36. Perl no longer allows group names to start with digits, so I have made this
643        change also in PCRE. It simplifies the code a bit.
644    
645    37. In extended mode, Perl ignores spaces before a + that indicates a
646        possessive quantifier. PCRE allowed a space before the quantifier, but not
647        before the possessive +. It now does.
648    
649    38. The use of \K (reset reported match start) within a repeated possessive
650        group such as (a\Kb)*+ was not working.
651    
652    40. Document that the same character tables must be used at compile time and
653        run time, and that the facility to pass tables to pcre_exec() and
654        pcre_dfa_exec() is for use only with saved/restored patterns.
655    
656    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
657        features for Builds with MSVC:
658    
659        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
660           builds). The .rc files can be used to set FileDescription and many other
661           attributes.
662    
663        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
664           This allows higher-level build scripts which want .pdb files to avoid
665           hard-coding the exact files needed."
666    
667    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
668        mean "start of word" and "end of word", respectively, as a transition aid.
669    
670    43. A minimizing repeat of a class containing codepoints greater than 255 in
671        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
672        compiled to use the heap for recursion.
673    
674    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
675        is configured.
676    
677    
678    Version 8.33 28-May-2013
679    ------------------------
680    
681    1.  Added 'U' to some constants that are compared to unsigned integers, to
682        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
683        variables that are added to signed variables, to ensure the result is
684        signed and can be negated.
685    
686    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
687        CMake config files.
688    
689    3.  Revise the creation of config.h.generic so that all boolean macros are
690        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
691        overriding via -D on the command line possible.
692    
693    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
694        to unsigned int is reported to make a quite noticeable speed difference in
695        a specific Windows environment. Testing on Linux did also appear to show
696        some benefit (and it is clearly not harmful). Also fixed the definition of
697        Xop which should be unsigned.
698    
699    5.  Related to (4), changing the definition of the intermediate variable cc
700        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
701        improvements.
702    
703    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
704        unnecessary spaces.
705    
706    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
707        and later.
708    
709    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
710    
711    9.  Optimizing fast_forward_start_bits in JIT.
712    
713    10. Adding support for callouts in JIT, and fixing some issues revealed
714        during this work. Namely:
715    
716        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
717    
718        (b) Minimum length was not checked before the matching is started.
719    
720    11. The value of capture_last that is passed to callouts was incorrect in some
721        cases when there was a capture on one path that was subsequently abandoned
722        after a backtrack. Also, the capture_last value is now reset after a
723        recursion, since all captures are also reset in this case.
724    
725    12. The interpreter no longer returns the "too many substrings" error in the
726        case when an overflowing capture is in a branch that is subsequently
727        abandoned after a backtrack.
728    
729    13. In the pathological case when an offset vector of size 2 is used, pcretest
730        now prints out the matched string after a yield of 0 or 1.
731    
732    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
733        Only supported by the JIT compiler at the moment.
734    
735    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
736    
737    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
738        offset of the starting point of the matching process, provided the offsets
739        vector is large enough.
740    
741    17. The \A escape now records a lookbehind value of 1, though its execution
742        does not actually inspect the previous character. This is to ensure that,
743        in partial multi-segment matching, at least one character from the old
744        segment is retained when a new segment is processed. Otherwise, if there
745        are no lookbehinds in the pattern, \A might match incorrectly at the start
746        of a new segment.
747    
748    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
749    
750    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
751        gives some modest performance improvement in 8-bit mode.
752    
753    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
754        be expressed in certain programming languages using Universal Character
755        Names.
756    
757    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
758        which points out that "non characters" are not "characters that may not
759        appear in Unicode strings" but rather "characters that are reserved for
760        internal use and have only local meaning".
761    
762    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
763        there was a conditional group that depended on an assertion, if the
764        assertion was false, the callout that immediately followed the alternation
765        in the condition was skipped when pcre_exec() was used for matching.
766    
767    23. Allow an explicit callout to be inserted before an assertion that is the
768        condition for a conditional group, for compatibility with automatic
769        callouts, which always insert a callout at this point.
770    
771    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
772        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
773    
774    25. (*PRUNE) is now supported by the JIT compiler.
775    
776    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
777    
778    27. Fix the case where there are two or more SKIPs with arguments that may be
779        ignored.
780    
781    28. (*SKIP) is now supported by the JIT compiler.
782    
783    29. (*THEN) is now supported by the JIT compiler.
784    
785    30. Update RunTest with additional test selector options.
786    
787    31. The way PCRE handles backtracking verbs has been changed in two ways.
788    
789        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
790        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
791        backtracking. In some cases this makes it more Perl-compatible, but Perl's
792        rather obscure rules do not always do the same thing.
793    
794        (2) Previously, backtracking verbs were confined within assertions. This is
795        no longer the case for positive assertions, except for (*ACCEPT). Again,
796        this sometimes improves Perl compatibility, and sometimes does not.
797    
798    32. A number of tests that were in test 2 because Perl did things differently
799        have been moved to test 1, because either Perl or PCRE has changed, and
800        these tests are now compatible.
801    
802    32. Backtracking control verbs are now handled in the same way in JIT and
803        interpreter.
804    
805    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
806        contained a forward subroutine reference caused a compile error.
807    
808    34. Auto-detect and optimize limited repetitions in JIT.
809    
810    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
811        blocking (*UTF) etc.
812    
813    36. In the interpreter, maximizing pattern repetitions for characters and
814        character types now use tail recursion, which reduces stack usage.
815    
816    37. The value of the max lookbehind was not correctly preserved if a compiled
817        and saved regex was reloaded on a host of different endianness.
818    
819    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
820        of the compiled pattern block, expand the flags field from 16 to 32 bits
821        because it was almost full.
822    
823    39. Try madvise first before posix_madvise.
824    
825    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
826        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
827        original change locked out all matches of empty strings. This has been
828        changed so that one match of an empty string per line is recognized.
829        Subsequent searches on the same line (for colouring or for --only-matching,
830        for example) do not recognize empty strings.
831    
832    41. Applied a user patch to fix a number of spelling mistakes in comments.
833    
834    42. Data lines longer than 65536 caused pcretest to crash.
835    
836    43. Clarified the data type for length and startoffset arguments for pcre_exec
837        and pcre_dfa_exec in the function-specific man pages, where they were
838        explicitly stated to be in bytes, never having been updated. I also added
839        some clarification to the pcreapi man page.
840    
841    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
842        a segmentation fault.
843    
844    
845    Version 8.32 30-November-2012
846    -----------------------------
847    
848    1.  Improved JIT compiler optimizations for first character search and single
849        character iterators.
850    
851    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
852        Patch by Daniel Richard G.
853    
854    3.  Single character iterator optimizations in the JIT compiler.
855    
856    4.  Improved JIT compiler optimizations for character ranges.
857    
858    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
859        Reported by Giuseppe D'Angelo.
860    
861    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
862        of a line, was being set incorrectly in cases where .* appeared inside
863        atomic brackets at the start of a pattern, or where there was a subsequent
864        *PRUNE or *SKIP.
865    
866    7.  Improved instruction cache flush for POWER/PowerPC.
867        Patch by Daniel Richard G.
868    
869    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
870        grep:
871    
872        (a) There is now no limit to the number of patterns to be matched.
873    
874        (b) An error is given if a pattern is too long.
875    
876        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
877            are now supported.
878    
879        (d) --exclude-from and --include-from (multiple use) have been added.
880    
881        (e) Exclusions and inclusions now apply to all files and directories, not
882            just to those obtained from scanning a directory recursively.
883    
884        (f) Multiple uses of -f and --file-list are now supported.
885    
886        (g) In a Windows environment, the default for -d has been changed from
887            "read" (the GNU grep default) to "skip", because otherwise the presence
888            of a directory in the file list provokes an error.
889    
890        (h) The documentation has been revised and clarified in places.
891    
892    9.  Improve the matching speed of capturing brackets.
893    
894    10. Changed the meaning of \X so that it now matches a Unicode extended
895        grapheme cluster.
896    
897    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
898        out POSIX threads when JIT support is configured.
899    
900    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
901    
902    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
903        structure could go wrong in environments where size_t is not the same size
904        as int.
905    
906    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
907        set.
908    
909    15. The EBCDIC support had decayed; later updates to the code had included
910        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
911        general tidy up of EBCDIC-related issues, and the documentation was also
912        not quite right. There is now a test that can be run on ASCII systems to
913        check some of the EBCDIC-related things (but is it not a full test).
914    
915    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
916        in a small tidy to the code.
917    
918    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
919    
920    18. If the --only-matching (-o) option in pcregrep is specified multiple
921        times, each one causes appropriate output. For example, -o1 -o2 outputs the
922        substrings matched by the 1st and 2nd capturing parentheses. A separating
923        string can be specified by --om-separator (default empty).
924    
925    19. Improving the first n character searches.
926    
927    20. Turn case lists for horizontal and vertical white space into macros so that
928        they are defined only once.
929    
930    21. This set of changes together give more compatible Unicode case-folding
931        behaviour for characters that have more than one other case when UCP
932        support is available.
933    
934        (a) The Unicode property table now has offsets into a new table of sets of
935            three or more characters that are case-equivalent. The MultiStage2.py
936            script that generates these tables (the pcre_ucd.c file) now scans
937            CaseFolding.txt instead of UnicodeData.txt for character case
938            information.
939    
940        (b) The code for adding characters or ranges of characters to a character
941            class has been abstracted into a generalized function that also handles
942            case-independence. In UTF-mode with UCP support, this uses the new data
943            to handle characters with more than one other case.
944    
945        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
946            whose other case is greater than 256 are now correctly matched
947            caselessly. Previously, the high codepoint matched the low one, but not
948            vice versa.
949    
950        (d) The processing of \h, \H, \v, and \ in character classes now makes use
951            of the new class addition function, using character lists defined as
952            macros alongside the case definitions of 20 above.
953    
954        (e) Caseless back references now work with characters that have more than
955            one other case.
956    
957        (f) General caseless matching of characters with more than one other case
958            is supported.
959    
960    22. Unicode character properties were updated from Unicode 6.2.0
961    
962    23. Improved CMake support under Windows. Patch by Daniel Richard G.
963    
964    24. Add support for 32-bit character strings, and UTF-32
965    
966    25. Major JIT compiler update (code refactoring and bugfixing).
967        Experimental Sparc 32 support is added.
968    
969    26. Applied a modified version of Daniel Richard G's patch to create
970        pcre.h.generic and config.h.generic by "make" instead of in the
971        PrepareRelease script.
972    
973    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
974        pcre_compile.c when checking for a zero character.
975    
976    28. Introducing a native interface for JIT. Through this interface, the compiled
977        machine code can be directly executed. The purpose of this interface is to
978        provide fast pattern matching, so several sanity checks are not performed.
979        However, feature tests are still performed. The new interface provides
980        1.4x speedup compared to the old one.
981    
982    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
983        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
984        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
985    
986    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
987        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
988        the "old" RFC 2279). Instead, it ended up passing a negative length to
989        pcre_exec().
990    
991    31. Add support for GCC's visibility feature to hide internal functions.
992    
993    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
994        "unknown -C option" after outputting 0 or 1.
995    
996    33. There is now support for generating a code coverage report for the test
997        suite in environments where gcc is the compiler and lcov is installed. This
998        is mainly for the benefit of the developers.
999    
1000    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1001        unaddressable using valgrind annotations, allowing valgrind to detect
1002        invalid memory accesses. This is mainly for the benefit of the developers.
1003    
1004    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1005    
1006    26. Give configure error if --enable-cpp but no C++ compiler found.
1007    
1008    
1009    Version 8.31 06-July-2012
1010    -------------------------
1011    
1012    1.  Fixing a wrong JIT test case and some compiler warnings.
1013    
1014    2.  Removed a bashism from the RunTest script.
1015    
1016    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1017        to unsigned type, result still unsigned" that was given by an MS compiler
1018        on encountering the code "-sizeof(xxx)".
1019    
1020    4.  Partial matching support is added to the JIT compiler.
1021    
1022    5.  Fixed several bugs concerned with partial matching of items that consist
1023        of more than one character:
1024    
1025        (a) /^(..)\1/ did not partially match "aba" because checking references was
1026            done on an "all or nothing" basis. This also applied to repeated
1027            references.
1028    
1029        (b) \R did not give a hard partial match if \r was found at the end of the
1030            subject.
1031    
1032        (c) \X did not give a hard partial match after matching one or more
1033            characters at the end of the subject.
1034    
1035        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1036            a partial match for the string "\r".
1037    
1038        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1039            a partial match for a CR character at the end of the subject string.
1040    
1041    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1042        running pcretest, the text "(JIT)" added to the output whenever JIT is
1043        actually used to run the match.
1044    
1045    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1046        or /S+[+] with a digit between 1 and 7.
1047    
1048    8.  OP_NOT now supports any UTF character not just single-byte ones.
1049    
1050    9.  (*MARK) control verb is now supported by the JIT compiler.
1051    
1052    10. The command "./RunTest list" lists the available tests without actually
1053        running any of them. (Because I keep forgetting what they all are.)
1054    
1055    11. Add PCRE_INFO_MAXLOOKBEHIND.
1056    
1057    12. Applied a (slightly modified) user-supplied patch that improves performance
1058        when the heap is used for recursion (compiled with --disable-stack-for-
1059        recursion). Instead of malloc and free for each heap frame each time a
1060        logical recursion happens, frames are retained on a chain and re-used where
1061        possible. This sometimes gives as much as 30% improvement.
1062    
1063    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1064        call.
1065    
1066    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1067    
1068    15. It is now possible to link pcretest with libedit as an alternative to
1069        libreadline.
1070    
1071    16. (*COMMIT) control verb is now supported by the JIT compiler.
1072    
1073    17. The Unicode data tables have been updated to Unicode 6.1.0.
1074    
1075    18. Added --file-list option to pcregrep.
1076    
1077    19. Added binary file support to pcregrep, including the -a, --binary-files,
1078        -I, and --text options.
1079    
1080    20. The madvise function is renamed for posix_madvise for QNX compatibility
1081        reasons. Fixed by Giuseppe D'Angelo.
1082    
1083    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1084    
1085    22. Changed the option for creating symbolic links for 16-bit man pages from
1086        -s to -sf so that re-installing does not cause issues.
1087    
1088    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1089    
1090    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1091        matches in certain environments (the workspace was not being correctly
1092        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1093        some of the workspace data at the beginning of a restart.
1094    
1095    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1096        was not doing so when it should - probably a typo introduced by SVN 528
1097        (change 8.10/14).
1098    
1099    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1100        \w+ when the character tables indicated that \x{c4} was a word character.
1101        There were several related cases, all because the tests for doing a table
1102        lookup were testing for characters less than 127 instead of 255.
1103    
1104    27. If a pattern contains capturing parentheses that are not used in a match,
1105        their slots in the ovector are set to -1. For those that are higher than
1106        any matched groups, this happens at the end of processing. In the case when
1107        there were back references that the ovector was too small to contain
1108        (causing temporary malloc'd memory to be used during matching), and the
1109        highest capturing number was not used, memory off the end of the ovector
1110        was incorrectly being set to -1. (It was using the size of the temporary
1111        memory instead of the true size.)
1112    
1113    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1114        ovector size, it uses memory at the end of the block that it has got.
1115    
1116    29. Check for an overlong MARK name and give an error at compile time. The
1117        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1118    
1119    30. JIT compiler update.
1120    
1121    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1122        Rill for the patch.
1123    
1124    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1125    
1126    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1127    
1128    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1129        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1130        was enabled.
1131    
1132    35. Improve JIT code generation for greedy plus quantifier.
1133    
1134    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1135        1 to "aa" instead of to an empty string. The bug affected repeated groups
1136        that could potentially match an empty string.
1137    
1138    37. Optimizing single character iterators in JIT.
1139    
1140    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1141        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1142        codepoints that are too big for the mode are faulted, and in a UTF mode,
1143        disallowed codepoints are also faulted.
1144    
1145    39. If PCRE was compiled with UTF support, in three places in the DFA
1146        matcher there was code that should only have been obeyed in UTF mode, but
1147        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1148        processing when bytes with values greater than 127 were present. In 16-bit
1149        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1150        both cases the values are those that cannot be the first data item in a UTF
1151        character. The three items that might have provoked this were recursions,
1152        possessively repeated groups, and atomic groups.
1153    
1154    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1155        and pcregrep, because some OS require shared objects to be explicitly
1156        passed to ld, causing the link step to fail if they are not.
1157    
1158    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1159        mode, patterns that started with \h* or \R* might be incorrectly matched.
1160    
1161    
1162    Version 8.30 04-February-2012
1163    -----------------------------
1164    
1165    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1166        name is defined in ctype.h.
1167    
1168    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1169        only in quite long subpatterns.
1170    
1171    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1172        since it was replaced by pcre_fullinfo() in February 2000.
1173    
1174    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1175        match a (*MARK), and the match failed at the start of the subject, a
1176        reference to memory before the start of the subject could occur. This bug
1177        was introduced by fix 17 of release 8.21.
1178    
1179    5.  A reference to an unset group with zero minimum repetition was giving
1180        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1181        /(another)?(\1?)test/ matched against "hello world test". This bug was
1182        introduced in release 8.13.
1183    
1184    6.  Add support for 16-bit character strings (a large amount of work involving
1185        many changes and refactorings).
1186    
1187    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1188        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1189        from a file.
1190    
1191    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1192        rounding is not applied in this particular case).
1193    
1194    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1195        if they appear, or are escaped, in patterns.
1196    
1197    10. Get rid of a number of -Wunused-but-set-variable warnings.
1198    
1199    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1200        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1201        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1202        also returns the mark "x". This bug applied to capturing parentheses,
1203        non-capturing parentheses, and atomic parentheses. It also applied to some
1204        assertions.
1205    
1206    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1207        information out of configure.ac instead of relying on pcre.h.generic, which
1208        is not stored in the repository.
1209    
1210    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1211        -lreadline.
1212    
1213    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1214    
1215    15. Applied Graycode's patch to put the top-level frame on the stack rather
1216        than the heap when not using the stack for recursion. This gives a
1217        performance improvement in many cases when recursion is not deep.
1218    
1219    16. Experimental code added to "pcretest -C" to output the stack frame size.
1220    
1221    
1222    Version 8.21 12-Dec-2011
1223    ------------------------
1224    
1225    1.  Updating the JIT compiler.
1226    
1227    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1228        are added as well.
1229    
1230    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1231        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1232        calling _pcre_jit_exec. Some extra comments are added.
1233    
1234    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1235        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1236        was introduced by change 18 for 8.20.
1237    
1238    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1239        ECMA-262 standard.
1240    
1241    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1242        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1243        This bug was probably introduced by change 9 of 8.13.
1244    
1245    7.  While fixing 6 above, I noticed that a number of other items were being
1246        incorrectly rejected as "not fixed length". This arose partly because newer
1247        opcodes had not been added to the fixed-length checking code. I have (a)
1248        corrected the bug and added tests for these items, and (b) arranged for an
1249        error to occur if an unknown opcode is encountered while checking for fixed
1250        length instead of just assuming "not fixed length". The items that were
1251        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1252        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1253        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1254    
1255    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1256        being incorrectly compiled and would have given unpredicatble results.
1257    
1258    9.  A possessively repeated subpattern with minimum repeat count greater than
1259        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1260        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1261        the first (A) could occur when it should not.
1262    
1263    10. Add a cast and remove a redundant test from the code.
1264    
1265    11. JIT should use pcre_malloc/pcre_free for allocation.
1266    
1267    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1268        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1269        is anything other than /usr, -L is still shown).
1270    
1271    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1272    
1273    14. Perl does not support \N without a following name in a [] class; PCRE now
1274        also gives an error.
1275    
1276    15. If a forward reference was repeated with an upper limit of around 2000,
1277        it caused the error "internal error: overran compiling workspace". The
1278        maximum number of forward references (including repeats) was limited by the
1279        internal workspace, and dependent on the LINK_SIZE. The code has been
1280        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1281        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1282        of around 200,000 forward references. While doing this, I also speeded up
1283        the filling in of repeated forward references.
1284    
1285    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1286        incorrectly expecting the subject to contain another "a" after the start.
1287    
1288    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1289        in the match, the SKIP should be ignored. This was not happening; instead
1290        the SKIP was being treated as NOMATCH. For patterns such as
1291        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1292        tested.
1293    
1294    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1295        now much more compatible with Perl, in particular in cases where the result
1296        is a non-match for a non-anchored pattern. For example, if
1297        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1298        "m", where previously it did not return a name. A side effect of this
1299        change is that for partial matches, the last encountered mark name is
1300        returned, as for non matches. A number of tests that were previously not
1301        Perl-compatible have been moved into the Perl-compatible test files. The
1302        refactoring has had the pleasing side effect of removing one argument from
1303        the match() function, thus reducing its stack requirements.
1304    
1305    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1306        subsequent uses of /S (without +) incorrectly behaved like /S+.
1307    
1308    21. Retrieve executable code size support for the JIT compiler and fixing
1309        some warnings.
1310    
1311    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1312        not work when the shorter character appeared right at the end of the
1313        subject string.
1314    
1315    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1316        systems.
1317    
1318    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1319        output it when the /M option is used in pcretest.
1320    
1321    25. The CheckMan script was not being included in the distribution. Also, added
1322        an explicit "perl" to run Perl scripts from the PrepareRelease script
1323        because this is reportedly needed in Windows.
1324    
1325    26. If study data was being save in a file and studying had not found a set of
1326        "starts with" bytes for the pattern, the data written to the file (though
1327        never used) was taken from uninitialized memory and so caused valgrind to
1328        complain.
1329    
1330    27. Updated RunTest.bat as provided by Sheri Pierce.
1331    
1332    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1333    
1334    29. Computation of memory usage for the table of capturing group names was
1335        giving an unnecessarily large value.
1336    
1337    
1338    Version 8.20 21-Oct-2011
1339    ------------------------
1340    
1341    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1342        a POSIX class. After further experiments with Perl, which convinced me that
1343        Perl has bugs and confusions, a closing square bracket is no longer allowed
1344        in a POSIX name. This bug also affected patterns with classes that started
1345        with full stops.
1346    
1347    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1348        captured substring, but while checking the failing first alternative,
1349        substring 1 is temporarily captured. If the output vector supplied to
1350        pcre_exec() was not big enough for this capture, the yield of the function
1351        was still zero ("insufficient space for captured substrings"). This cannot
1352        be totally fixed without adding another stack variable, which seems a lot
1353        of expense for a edge case. However, I have improved the situation in cases
1354        such as /(a)(b)x|abc/ matched against "abc", where the return code
1355        indicates that fewer than the maximum number of slots in the ovector have
1356        been set.
1357    
1358    3.  Related to (2) above: when there are more back references in a pattern than
1359        slots in the output vector, pcre_exec() uses temporary memory during
1360        matching, and copies in the captures as far as possible afterwards. It was
1361        using the entire output vector, but this conflicts with the specification
1362        that only 2/3 is used for passing back captured substrings. Now it uses
1363        only the first 2/3, for compatibility. This is, of course, another edge
1364        case.
1365    
1366    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1367        main code base, and can be used by building with --enable-jit. When this is
1368        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1369        runtime --no-jit option is given.
1370    
1371    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1372        ovector, the return from the function was zero, implying that there were
1373        other matches that did not fit. The correct "exactly full" value is now
1374        returned.
1375    
1376    6.  If a subpattern that was called recursively or as a subroutine contained
1377        (*PRUNE) or any other control that caused it to give a non-standard return,
1378        invalid errors such as "Error -26 (nested recursion at the same subject
1379        position)" or even infinite loops could occur.
1380    
1381    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1382        computing the minimum length on reaching *ACCEPT, and so ended up with the
1383        wrong value of 1 rather than 0. Further investigation indicates that
1384        computing a minimum subject length in the presence of *ACCEPT is difficult
1385        (think back references, subroutine calls), and so I have changed the code
1386        so that no minimum is registered for a pattern that contains *ACCEPT.
1387    
1388    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1389        it was not handled as intended. [But see 16 below.]
1390    
1391    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1392        Sheri Pierce.
1393    
1394    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1395        the first byte in a match must be "a".
1396    
1397    11. Change 17 for 8.13 increased the recursion depth for patterns like
1398        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1399        pattern contains any instances of (*THEN). If it does not, the old
1400        optimizations are restored. It would be nice to do this on a per-group
1401        basis, but at the moment that is not feasible.
1402    
1403    12. In some environments, the output of pcretest -C is CRLF terminated. This
1404        broke RunTest's code that checks for the link size. A single white space
1405        character after the value is now allowed for.
1406    
1407    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1408        For "fr", it uses the Windows-specific input and output files.
1409    
1410    14. If (*THEN) appeared in a group that was called recursively or as a
1411        subroutine, it did not work as intended. [But see next item.]
1412    
1413    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1414        pattern fragments (but not containing any | characters). If A and B are
1415        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1416        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1417        D. In other words, Perl considers parentheses that do not contain any |
1418        characters to be part of a surrounding alternative, whereas PCRE was
1419        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1420        differently. PCRE now behaves in the same way as Perl, except in the case
1421        of subroutine/recursion calls such as (?1) which have in any case always
1422        been different (but PCRE had them first :-).
1423    
1424    16. Related to 15 above: Perl does not treat the | in a conditional group as
1425        creating alternatives. Such a group is treated in the same way as an
1426        ordinary group without any | characters when processing (*THEN). PCRE has
1427        been changed to match Perl's behaviour.
1428    
1429    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1430        RunGrepTest script failed.
1431    
1432    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1433        inevitable for groups that contain captures, but it can lead to a lot of
1434        stack use in large patterns. The old behaviour has been restored for atomic
1435        groups that do not contain any capturing parentheses.
1436    
1437    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1438        suppress the check for a minimum subject length at run time. (If it was
1439        given to pcre_exec() or pcre_dfa_exec() it did work.)
1440    
1441    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1442        fail to work when decoding hex characters in data strings in EBCDIC
1443        environments.
1444    
1445    21. It appears that in at least one Mac OS environment, the isxdigit() function
1446        is implemented as a macro that evaluates to its argument more than once,
1447        contravening the C 90 Standard (I haven't checked a later standard). There
1448        was an instance in pcretest which caused it to go wrong when processing
1449        \x{...} escapes in subject strings. The has been rewritten to avoid using
1450        things like p++ in the argument of isxdigit().
1451    
1452    
1453    Version 8.13 16-Aug-2011
1454    ------------------------
1455    
1456    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1457    
1458    2.  Two minor typos in pcre_internal.h have been fixed.
1459    
1460    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1461        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1462        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1463    
1464    4.  There were a number of related bugs in the code for matching backrefences
1465        caselessly in UTF-8 mode when codes for the characters concerned were
1466        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1467        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1468        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1469        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1470        2-byte code at the end of the subject (it thought there wasn't enough data
1471        left).
1472    
1473    5.  Comprehensive information about what went wrong is now returned by
1474        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1475        as the output vector has at least 2 elements. The offset of the start of
1476        the failing character and a reason code are placed in the vector.
1477    
1478    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1479        now returned is for the first byte of the failing character, instead of the
1480        last byte inspected. This is an incompatible change, but I hope it is small
1481        enough not to be a problem. It makes the returned offset consistent with
1482        pcre_exec() and pcre_dfa_exec().
1483    
1484    7.  pcretest now gives a text phrase as well as the error number when
1485        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1486        failure, the offset and reason code are output.
1487    
1488    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1489        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1490        back over a single character (\n). This seems wrong (because it treated the
1491        two characters as a single entity when going forwards), conflicts with the
1492        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1493        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1494        has been changed.
1495    
1496    9.  Some internal refactoring has changed the processing so that the handling
1497        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1498        time (the PCRE_DOTALL option was changed this way some time ago: version
1499        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1500        which was always a bit of a fudge. It also means that there is one less
1501        argument for the match() function, which reduces its stack requirements
1502        slightly. This change also fixes an incompatibility with Perl: the pattern
1503        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1504    
1505    10. More internal refactoring has drastically reduced the number of recursive
1506        calls to match() for possessively repeated groups such as (abc)++ when
1507        using pcre_exec().
1508    
1509    11. While implementing 10, a number of bugs in the handling of groups were
1510        discovered and fixed:
1511    
1512        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1513        (a|)*(?1) gave a compile-time internal error.
1514        ((a|)+)+  did not notice that the outer group could match an empty string.
1515        (^a|^)+   was not marked as anchored.
1516        (.*a|.*)+ was not marked as matching at start or after a newline.
1517    
1518    12. Yet more internal refactoring has removed another argument from the match()
1519        function. Special calls to this function are now indicated by setting a
1520        value in a variable in the "match data" data block.
1521    
1522    13. Be more explicit in pcre_study() instead of relying on "default" for
1523        opcodes that mean there is no starting character; this means that when new
1524        ones are added and accidentally left out of pcre_study(), testing should
1525        pick them up.
1526    
1527    14. The -s option of pcretest has been documented for ages as being an old
1528        synonym of -m (show memory usage). I have changed it to mean "force study
1529        for every regex", that is, assume /S for every regex. This is similar to -i
1530        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1531        using it. It makes it easier to run collections of tests with and without
1532        study enabled, and thereby test pcre_study() more easily. All the standard
1533        tests are now run with and without -s (but some patterns can be marked as
1534        "never study" - see 20 below).
1535    
1536    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1537        restoration of the capturing data to the outer values was not happening
1538        correctly.
1539    
1540    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1541        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1542        pattern had matched an empty string, and so incorrectly returned a no
1543        match.
1544    
1545    17. There was optimizing code for the last branch of non-capturing parentheses,
1546        and also for the obeyed branch of a conditional subexpression, which used
1547        tail recursion to cut down on stack usage. Unfortunately, now that there is
1548        the possibility of (*THEN) occurring in these branches, tail recursion is
1549        no longer possible because the return has to be checked for (*THEN). These
1550        two optimizations have therefore been removed. [But see 8.20/11 above.]
1551    
1552    18. If a pattern containing \R was studied, it was assumed that \R always
1553        matched two bytes, thus causing the minimum subject length to be
1554        incorrectly computed because \R can also match just one byte.
1555    
1556    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1557        was incorrectly computed.
1558    
1559    20. If /S is present twice on a test pattern in pcretest input, it now
1560        *disables* studying, thereby overriding the use of -s on the command line
1561        (see 14 above). This is necessary for one or two tests to keep the output
1562        identical in both cases.
1563    
1564    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1565        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1566    
1567    22. When an atomic group that contained a capturing parenthesis was
1568        successfully matched, but the branch in which it appeared failed, the
1569        capturing was not being forgotten if a higher numbered group was later
1570        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1571        group 1 to "a", when in fact it should be unset. This applied to multi-
1572        branched capturing and non-capturing groups, repeated or not, and also to
1573        positive assertions (capturing in negative assertions does not happen
1574        in PCRE) and also to nested atomic groups.
1575    
1576    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1577        subject after a captured substring, to make it easier to tell which of a
1578        number of identical substrings has been captured.
1579    
1580    24. The way atomic groups are processed by pcre_exec() has been changed so that
1581        if they are repeated, backtracking one repetition now resets captured
1582        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1583        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1584        "aaa". Previously, it would have been "a". As part of this code
1585        refactoring, the way recursive calls are handled has also been changed.
1586    
1587    25. If an assertion condition captured any substrings, they were not passed
1588        back unless some other capturing happened later. For example, if
1589        (?(?=(a))a) was matched against "a", no capturing was returned.
1590    
1591    26. When studying a pattern that contained subroutine calls or assertions,
1592        the code for finding the minimum length of a possible match was handling
1593        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1594        group 1 called group 2 while simultaneously a separate group 2 called group
1595        1). A stack overflow occurred in this case. I have fixed this by limiting
1596        the recursion depth to 10.
1597    
1598    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1599        Fortmann. This supports explicit test numbers on the command line, and has
1600        argument validation and error reporting.
1601    
1602    28. An instance of \X with an unlimited repeat could fail if at any point the
1603        first character it looked at was a mark character.
1604    
1605    29. Some minor code refactoring concerning Unicode properties and scripts
1606        should reduce the stack requirement of match() slightly.
1607    
1608    30. Added the '=' option to pcretest to check the setting of unused capturing
1609        slots at the end of the pattern, which are documented as being -1, but are
1610        not included in the return count.
1611    
1612    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1613        compiled something random. Now it gives a compile-time error (as does
1614        Perl).
1615    
1616    32. A *MARK encountered during the processing of a positive assertion is now
1617        recorded and passed back (compatible with Perl).
1618    
1619    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1620        had alternative anchored branches, the search for a second match in a line
1621        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1622        matched the line "0102" twice. The same bug affected patterns that started
1623        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1624        twice.
1625    
1626    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1627        does, and because of capturing effects, quantifying parenthesized
1628        assertions may at times be useful. Quantifiers are now allowed for
1629        parenthesized assertions.
1630    
1631    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1632    
1633    36. \g was being checked for fancy things in a character class, when it should
1634        just be a literal "g".
1635    
1636    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1637        appearance of a nested POSIX class supersedes an apparent external class.
1638        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1639        unescaped square brackets may also appear as part of class names. For
1640        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1641        more like Perl. (But see 8.20/1 above.)
1642    
1643    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1644        was because it thought it was \N{name}, which is not supported).
1645    
1646    39. Add minix to OS list not supporting the -S option in pcretest.
1647    
1648    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1649        cannot analyze patterns in sufficient detail to catch mutual recursions
1650        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1651        subgroup is called recursively as a subpattern for a second time at the
1652        same position in the subject string. In previous releases this might have
1653        been caught by the recursion limit, or it might have run out of stack.
1654    
1655    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1656        happen only once. PCRE was, however incorrectly giving a compile time error
1657        "recursive call could loop indefinitely" because it cannot analyze the
1658        pattern in sufficient detail. The compile time test no longer happens when
1659        PCRE is compiling a conditional subpattern, but actual runaway loops are
1660        now caught at runtime (see 40 above).
1661    
1662    42. It seems that Perl allows any characters other than a closing parenthesis
1663        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1664        has been changed to be the same.
1665    
1666    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1667        as not to get warnings when autogen.sh is called. Also changed
1668        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1669    
1670    44. To help people who use pcregrep to scan files containing exceedingly long
1671        lines, the following changes have been made:
1672    
1673        (a) The default value of the buffer size parameter has been increased from
1674            8K to 20K. (The actual buffer used is three times this size.)
1675    
1676        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1677            PCRE is built.
1678    
1679        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1680            to be set at run time.
1681    
1682        (d) Numerical values in pcregrep options can be followed by K or M, for
1683            example --buffer-size=50K.
1684    
1685        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1686            given and the return code is set to 2.
1687    
1688    45. Add a pointer to the latest mark to the callout data block.
1689    
1690    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1691        partial match of an empty string instead of no match. This was specific to
1692        the use of ".".
1693    
1694    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1695        complete match instead of a partial match. This bug was dependent on both
1696        the PCRE_UTF8 and PCRE_DOTALL options being set.
1697    
1698    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1699        starting byte set, because \b was not being ignored.
1700    
1701    
1702    Version 8.12 15-Jan-2011
1703    ------------------------
1704    
1705    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1706        checks for such things as part of the documentation building process.
1707    
1708    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1709        --match-limit and --recursion-limit options (added for 8.11). In
1710        particular, this made one of the standard tests fail. (The integer value
1711        went into the wrong half of a long int.)
1712    
1713    3.  If the --colour option was given to pcregrep with -v (invert match), it
1714        did strange things, either producing crazy output, or crashing. It should,
1715        of course, ignore a request for colour when reporting lines that do not
1716        match.
1717    
1718    4.  Another pcregrep bug caused similar problems if --colour was specified with
1719        -M (multiline) and the pattern match finished with a line ending.
1720    
1721    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1722        matched in multiline mode, the following line was shown as part of the
1723        match. This seems wrong, so I have changed it.
1724    
1725    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1726        the check for further matches in the same line (so they could be coloured)
1727        to overrun the end of the current line. If another match was found, it was
1728        incorrectly shown (and then shown again when found in the next line).
1729    
1730    7.  If pcregrep was compiled under Windows, there was a reference to the
1731        function pcregrep_exit() before it was defined. I am assuming this was
1732        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1733        reported by a user. I've moved the definition above the reference.
1734    
1735    
1736    Version 8.11 10-Dec-2010
1737    ------------------------
1738    
1739    1.  (*THEN) was not working properly if there were untried alternatives prior
1740        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1741        backtracked to try for "b" instead of moving to the next alternative branch
1742        at the same level (in this case, to look for "c"). The Perl documentation
1743        is clear that when (*THEN) is backtracked onto, it goes to the "next
1744        alternative in the innermost enclosing group".
1745    
1746    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1747        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1748        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1749        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1750        (*THEN).
1751    
1752    3.  If \s appeared in a character class, it removed the VT character from
1753        the class, even if it had been included by some previous item, for example
1754        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1755        of \s, but is part of the POSIX "space" class.)
1756    
1757    4.  A partial match never returns an empty string (because you can always
1758        match an empty string at the end of the subject); however the checking for
1759        an empty string was starting at the "start of match" point. This has been
1760        changed to the "earliest inspected character" point, because the returned
1761        data for a partial match starts at this character. This means that, for
1762        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1763        (previously it gave "no match").
1764    
1765    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1766        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1767        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1768        has an implication that the given string is incomplete (because a partial
1769        match is preferred over a full match). For this reason, these items now
1770        give a partial match in this situation. [Aside: previously, the one case
1771        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1772        match rather than a full match, which was wrong by the old rules, but is
1773        now correct.]
1774    
1775    6.  There was a bug in the handling of #-introduced comments, recognized when
1776        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1777        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1778        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1779        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1780        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1781        places in pcre_compile().
1782    
1783    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1784        comments when looking ahead for named forward references to subpatterns,
1785        the only newline sequence it recognized was NL. It now handles newlines
1786        according to the set newline convention.
1787    
1788    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1789        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1790        cater for a lack of strerror(). These oversights have been fixed.
1791    
1792    9.  Added --match-limit and --recursion-limit to pcregrep.
1793    
1794    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1795    
1796    11. When the -o option was used, pcregrep was setting a return code of 1, even
1797        when matches were found, and --line-buffered was not being honoured.
1798    
1799    12. Added an optional parentheses number to the -o and --only-matching options
1800        of pcregrep.
1801    
1802    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1803        can match an empty string. The code to do it in pcretest and pcredemo
1804        needed fixing:
1805    
1806        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1807            only one byte after an empty string match just before CRLF (this case
1808            just got forgotten; "any" and "anycrlf" were OK).
1809    
1810        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1811            mode when an empty string match preceded an ASCII character followed by
1812            a non-ASCII character. (The code for advancing by one character rather
1813            than one byte was nonsense.)
1814    
1815        (c) The pcredemo.c sample program did not have any code at all to handle
1816            the cases when CRLF is a valid newline sequence.
1817    
1818    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1819        as a starting offset was within the subject string. There is now a new
1820        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1821        negative or greater than the length of the string. In order to test this,
1822        pcretest is extended to allow the setting of negative starting offsets.
1823    
1824    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1825        starting offset points to the beginning of a UTF-8 character was
1826        unnecessarily clumsy. I tidied it up.
1827    
1828    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1829        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1830    
1831    17. Nobody had reported that the --include_dir option, which was added in
1832        release 7.7 should have been called --include-dir (hyphen, not underscore)
1833        for compatibility with GNU grep. I have changed it to --include-dir, but
1834        left --include_dir as an undocumented synonym, and the same for
1835        --exclude-dir, though that is not available in GNU grep, at least as of
1836        release 2.5.4.
1837    
1838    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1839        characters from a string of bytes) have been redefined so as not to use
1840        loops, in order to improve performance in some environments. At the same
1841        time, I abstracted some of the common code into auxiliary macros to save
1842        repetition (this should not affect the compiled code).
1843    
1844    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1845        compile-time error is now given if \c is not followed by an ASCII
1846        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1847        different, and any byte value is allowed.)
1848    
1849    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1850        START_OPTIMIZE option, which is now allowed at compile time - but just
1851        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1852        to pcregrep and other applications that have no direct access to PCRE
1853        options. The new /Y option in pcretest sets this option when calling
1854        pcre_compile().
1855    
1856    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1857        back references. Groups containing recursive back references were forced to
1858        be atomic by that change, but in the case of named groups, the amount of
1859        memory required was incorrectly computed, leading to "Failed: internal
1860        error: code overflow". This has been fixed.
1861    
1862    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1863        pcretest.c, to avoid build problems in some Borland environments.
1864    
1865    
1866    Version 8.10 25-Jun-2010
1867    ------------------------
1868    
1869    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1870        THEN.
1871    
1872    2.  (*ACCEPT) was not working when inside an atomic group.
1873    
1874    3.  Inside a character class, \B is treated as a literal by default, but
1875        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1876        causes the error). The code is unchanged, but I tidied the documentation.
1877    
1878    4.  Inside a character class, PCRE always treated \R and \X as literals,
1879        whereas Perl faults them if its -w option is set. I have changed PCRE so
1880        that it faults them when PCRE_EXTRA is set.
1881    
1882    5.  Added support for \N, which always matches any character other than
1883        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1884    
1885    6.  When compiling pcregrep with newer versions of gcc which may have
1886        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1887        declared with attribute warn_unused_result" were given. Just casting the
1888        result to (void) does not stop the warnings; a more elaborate fudge is
1889        needed. I've used a macro to implement this.
1890    
1891    7.  Minor change to pcretest.c to avoid a compiler warning.
1892    
1893    8.  Added four artifical Unicode properties to help with an option to make
1894        \s etc use properties (see next item). The new properties are: Xan
1895        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1896    
1897    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1898        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1899        this option. Modified pcretest to add /W to test this facility. Added
1900        REG_UCP to make it available via the POSIX interface.
1901    
1902    10. Added --line-buffered to pcregrep.
1903    
1904    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1905        studied, and the match started with a letter with a code point greater than
1906        127 whose first byte was different to the first byte of the other case of
1907        the letter, the other case of this starting letter was not recognized
1908        (#976).
1909    
1910    12. If a pattern that was studied started with a repeated Unicode property
1911        test, for example, \p{Nd}+, there was the theoretical possibility of
1912        setting up an incorrect bitmap of starting bytes, but fortunately it could
1913        not have actually happened in practice until change 8 above was made (it
1914        added property types that matched character-matching opcodes).
1915    
1916    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1917        possible starting bytes for non-anchored patterns.
1918    
1919    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1920        \R, and also a number of cases that involve Unicode properties, both
1921        explicit and implicit when PCRE_UCP is set.
1922    
1923    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1924        input, it could crash or give wrong results if characters with values
1925        greater than 0xc0 were present in the subject string. (Detail: it assumed
1926        UTF-8 input when processing these items.)
1927    
1928    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1929        size_t is 64-bit (#991).
1930    
1931    17. Added a check for running out of memory when PCRE is compiled with
1932        --disable-stack-for-recursion (#990).
1933    
1934    18. If the last data line in a file for pcretest does not have a newline on
1935        the end, a newline was missing in the output.
1936    
1937    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1938        less than 128) in its various bitmaps. However, there is a facility for
1939        generating tables according to the current locale when PCRE is compiled. It
1940        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1941        characters, are recognized by isspace() and therefore were getting set in
1942        these tables, and indeed these tables seem to approximate to ISO 8859. This
1943        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1944        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1945        which of course cannot start UTF-8 characters. I have changed the code so
1946        that only real ASCII characters (less than 128) and the correct starting
1947        bytes for UTF-8 encodings are set for characters greater than 127 when in
1948        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1949        altogether.)
1950    
1951    20. Added the /T option to pcretest so as to be able to run tests with non-
1952        standard character tables, thus making it possible to include the tests
1953        used for 19 above in the standard set of tests.
1954    
1955    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1956        reference to a subpattern the other side of a comment that contains an
1957        opening parenthesis caused either an internal compiling error, or a
1958        reference to the wrong subpattern.
1959    
1960    
1961    Version 8.02 19-Mar-2010
1962    ------------------------
1963    
1964    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1965    
1966    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1967        configured.
1968    
1969    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1970        original author of that file, following a query about its status.
1971    
1972    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1973        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1974    
1975    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1976        quantifier applied to a forward-referencing subroutine call, could compile
1977        incorrect code or give the error "internal error: previously-checked
1978        referenced subpattern not found".
1979    
1980    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1981        variables to point to external functions. For these systems, therefore,
1982        pcre_malloc etc. are now initialized to local functions that call the
1983        relevant global functions.
1984    
1985    7.  There were two entries missing in the vectors called coptable and poptable
1986        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1987        I've fixed the data, and added a kludgy way of testing at compile time that
1988        the lengths are correct (equal to the number of opcodes).
1989    
1990    8.  Following on from 7, I added a similar kludge to check the length of the
1991        eint vector in pcreposix.c.
1992    
1993    9.  Error texts for pcre_compile() are held as one long string to avoid too
1994        much relocation at load time. To find a text, the string is searched,
1995        counting zeros. There was no check for running off the end of the string,
1996        which could happen if a new error number was added without updating the
1997        string.
1998    
1999    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2000    
2001    11. \K was not working if it appeared in an atomic group or in a group that
2002        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2003        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2004        the assertion is positive, but not if it is negative.
2005    
2006    12. Change 11 fortuitously reduced the size of the stack frame used in the
2007        "match()" function of pcre_exec.c by one pointer. Forthcoming
2008        implementation of support for (*MARK) will need an extra pointer on the
2009        stack; I have reserved it now, so that the stack frame size does not
2010        decrease.
2011    
2012    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2013        item in branch that calls a recursion is a subroutine call - as in the
2014        second branch in the above example - was incorrectly given the compile-
2015        time error "recursive call could loop indefinitely" because pcre_compile()
2016        was not correctly checking the subroutine for matching a non-empty string.
2017    
2018    14. The checks for overrunning compiling workspace could trigger after an
2019        overrun had occurred. This is a "should never occur" error, but it can be
2020        triggered by pathological patterns such as hundreds of nested parentheses.
2021        The checks now trigger 100 bytes before the end of the workspace.
2022    
2023    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2024    
2025    
2026    Version 8.01 19-Jan-2010
2027    ------------------------
2028    
2029    1.  If a pattern contained a conditional subpattern with only one branch (in
2030        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2031        computed the wrong minimum data length (which is of course zero for such
2032        subpatterns). This could cause incorrect "no match" results.
2033    
2034    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2035        the pattern is reset in the first branch, pcre_compile() failed with
2036        "internal error: code overflow at offset...". This happened only when
2037        the reset was to the original external option setting. (An optimization
2038        abstracts leading options settings into an external setting, which was the
2039        cause of this.)
2040    
2041    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2042        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2043        assertion pattern did not match (meaning that the assertion was true), it
2044        was incorrectly treated as false if the SKIP had been reached during the
2045        matching. This also applied to assertions used as conditions.
2046    
2047    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2048        assertion subpattern, including such a pattern used as a condition,
2049        unpredictable results occurred, instead of the error return
2050        PCRE_ERROR_DFA_UITEM.
2051    
2052    5.  The C++ GlobalReplace function was not working like Perl for the special
2053        situation when an empty string is matched. It now does the fancy magic
2054        stuff that is necessary.
2055    
2056    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2057        removed. (These were left over from very, very early versions of PCRE.)
2058    
2059    7.  Some cosmetic changes to the code to make life easier when compiling it
2060        as part of something else:
2061    
2062        (a) Change DEBUG to PCRE_DEBUG.
2063    
2064        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2065            called "current" as "current_branch", to prevent a collision with the
2066            Linux macro when compiled as a kernel module.
2067    
2068        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2069            prevent a collision with the Linux macro when compiled as a kernel
2070            module.
2071    
2072    8.  In pcre_compile() there are some checks for integer overflows that used to
2073        cast potentially large values to (double). This has been changed to that
2074        when building, a check for int64_t is made, and if it is found, it is used
2075        instead, thus avoiding the use of floating point arithmetic. (There is no
2076        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2077        double.
2078    
2079    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2080        2005 (difference between two addresses compared to an unsigned value).
2081    
2082    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2083        custom one, because of the following reported problem in Windows:
2084    
2085          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2086              under Win32.
2087          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2088              therefore missing the function definition.
2089          - The compiler thus generates a "C" signature for the test function.
2090          - The linker fails to find the "C" function.
2091          - PCRE fails to configure if asked to do so against libbz2.
2092    
2093    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2094        messages were output:
2095    
2096          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2097          rerunning libtoolize, to keep the correct libtool macros in-tree.
2098          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2099    
2100        I have done both of these things.
2101    
2102    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2103        most of the time, it *can* run out if it is given a pattern that contains a
2104        runaway infinite recursion. I updated the discussion in the pcrestack man
2105        page.
2106    
2107    13. Now that we have gone to the x.xx style of version numbers, the minor
2108        version may start with zero. Using 08 or 09 is a bad idea because users
2109        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2110        interpreted as invalid octal numbers. I've updated the previous comment in
2111        configure.ac, and also added a check that gives an error if 08 or 09 are
2112        used.
2113    
2114    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2115        causing partial matching to fail when the end of the subject matched \W
2116        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2117    
2118    15. There were some discrepancies between the declarations in pcre_internal.h
2119        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2120        their definitions. The declarations used "const uschar *" and the
2121        definitions used USPTR. Even though USPTR is normally defined as "const
2122        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2123        reported that: "This difference in casting confuses some C++ compilers, for
2124        example, SunCC recognizes above declarations as different functions and
2125        generates broken code for hbpcre." I have changed the declarations to use
2126        USPTR.
2127    
2128    16. GNU libtool is named differently on some systems. The autogen.sh script now
2129        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2130        (FreeBSD).
2131    
2132    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2133        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2134        comment: "Figure out how to create a longlong from a string: strtoll and
2135        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2136        instance, but it only takes 2 args instead of 3!"
2137    
2138    18. A subtle bug concerned with back references has been fixed by a change of
2139        specification, with a corresponding code fix. A pattern such as
2140        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2141        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2142        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2143        same bug. Such groups have to be quantified to be useful, or contained
2144        inside another quantified group. (If there's no repetition, the reference
2145        can never match.) The problem arises because, having left the group and
2146        moved on to the rest of the pattern, a later failure that backtracks into
2147        the group uses the captured value from the final iteration of the group
2148        rather than the correct earlier one. I have fixed this in PCRE by forcing
2149        any group that contains a reference to itself to be an atomic group; that
2150        is, there cannot be any backtracking into it once it has completed. This is
2151        similar to recursive and subroutine calls.
2152    
2153    
2154    Version 8.00 19-Oct-09
2155    ----------------------
2156    
2157    1.  The table for translating pcre_compile() error codes into POSIX error codes
2158        was out-of-date, and there was no check on the pcre_compile() error code
2159        being within the table. This could lead to an OK return being given in
2160        error.
2161    
2162    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2163        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2164        in a Windows environment.
2165    
2166    3.  The pcregrep --count option prints the count for each file even when it is
2167        zero, as does GNU grep. However, pcregrep was also printing all files when
2168        --files-with-matches was added. Now, when both options are given, it prints
2169        counts only for those files that have at least one match. (GNU grep just
2170        prints the file name in this circumstance, but including the count seems
2171        more useful - otherwise, why use --count?) Also ensured that the
2172        combination -clh just lists non-zero counts, with no names.
2173    
2174    4.  The long form of the pcregrep -F option was incorrectly implemented as
2175        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2176        but it seems right to fix it, and I didn't think it was worth preserving
2177        the old behaviour.
2178    
2179    5.  The command line items --regex=pattern and --regexp=pattern were not
2180        recognized by pcregrep, which required --regex pattern or --regexp pattern
2181        (with a space rather than an '='). The man page documented the '=' forms,
2182        which are compatible with GNU grep; these now work.
2183    
2184    6.  No libpcreposix.pc file was created for pkg-config; there was just
2185        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2186    
2187    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2188        when UCP support is not needed, by modifying the Python script that
2189        generates it from Unicode data files. This should not matter if the module
2190        is correctly used as a library, but I received one complaint about 50K of
2191        unwanted data. My guess is that the person linked everything into his
2192        program rather than using a library. Anyway, it does no harm.
2193    
2194    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2195        was a minimum greater than 1 for a wide character in a possessive
2196        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2197        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2198        character. Chaos in the form of incorrect output or a compiling loop could
2199        result.
2200    
2201    9.  The restrictions on what a pattern can contain when partial matching is
2202        requested for pcre_exec() have been removed. All patterns can now be
2203        partially matched by this function. In addition, if there are at least two
2204        slots in the offset vector, the offset of the earliest inspected character
2205        for the match and the offset of the end of the subject are set in them when
2206        PCRE_ERROR_PARTIAL is returned.
2207    
2208    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2209        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2210        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2211        and may be more useful for multi-segment matching.
2212    
2213    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2214        used to be given if ever the end of the subject was reached; now it is
2215        given only if matching could not proceed because another character was
2216        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2217        string "Z", which now yields "no match" instead of "partial match". In the
2218        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2219        final character ended with (*FAIL).
2220    
2221    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2222        if the pattern had a "must contain" character that was already found in the
2223        earlier partial match, unless partial matching was again requested. For
2224        example, with the pattern /dog.(body)?/, the "must contain" character is
2225        "g". If the first part-match was for the string "dog", restarting with
2226        "sbody" failed. This bug has been fixed.
2227    
2228    13. The string returned by pcre_dfa_exec() after a partial match has been
2229        changed so that it starts at the first inspected character rather than the
2230        first character of the match. This makes a difference only if the pattern
2231        starts with a lookbehind assertion or \b or \B (\K is not supported by
2232        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2233        matching functions compatible, and I think it's the right thing to do.
2234    
2235    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2236        so that the demonstration program is easily available in environments where
2237        PCRE has not been installed from source.
2238    
2239    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2240        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2241        library.
2242    
2243    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2244        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2245        is not the first non-POSIX option to be added. Clearly some people find
2246        these options useful.
2247    
2248    17. If a caller to the POSIX matching function regexec() passes a non-zero
2249        value for nmatch with a NULL value for pmatch, the value of
2250        nmatch is forced to zero.
2251    
2252    18. RunGrepTest did not have a test for the availability of the -u option of
2253        the diff command, as RunTest does. It now checks in the same way as
2254        RunTest, and also checks for the -b option.
2255    
2256    19. If an odd number of negated classes containing just a single character
2257        interposed, within parentheses, between a forward reference to a named
2258        subpattern and the definition of the subpattern, compilation crashed with
2259        an internal error, complaining that it could not find the referenced
2260        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2261        [The bug was that it was starting one character too far in when skipping
2262        over the character class, thus treating the ] as data rather than
2263        terminating the class. This meant it could skip too much.]
2264    
2265    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2266        /g option in pcretest when the pattern contains \K, which makes it possible
2267        to have an empty string match not at the start, even when the pattern is
2268        anchored. Updated pcretest and pcredemo to use this option.
2269    
2270    21. If the maximum number of capturing subpatterns in a recursion was greater
2271        than the maximum at the outer level, the higher number was returned, but
2272        with unset values at the outer level. The correct (outer level) value is
2273        now given.
2274    
2275    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2276        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2277        make it do so. The string so far is captured, making this feature
2278        compatible with Perl.
2279    
2280    23. The tests have been re-organized, adding tests 11 and 12, to make it
2281        possible to check the Perl 5.10 features against Perl 5.10.
2282    
2283    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2284        pattern matches a fixed length string. PCRE did not allow this; now it
2285        does. Neither allows recursion.
2286    
2287    25. I finally figured out how to implement a request to provide the minimum
2288        length of subject string that was needed in order to match a given pattern.
2289        (It was back references and recursion that I had previously got hung up
2290        on.) This code has now been added to pcre_study(); it finds a lower bound
2291        to the length of subject needed. It is not necessarily the greatest lower
2292        bound, but using it to avoid searching strings that are too short does give
2293        some useful speed-ups. The value is available to calling programs via
2294        pcre_fullinfo().
2295    
2296    26. While implementing 25, I discovered to my embarrassment that pcretest had
2297        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2298        study optimizations had never been tested with that matching function.
2299        Oops. What is worse, even when it was passed study data, there was a bug in
2300        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2301        were also very few tests of studied patterns with pcre_dfa_exec().
2302    
2303    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2304        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2305        on the other side of the coin, they are no longer allowed to have different
2306        names, because these cannot be distinguished in PCRE, and this has caused
2307        confusion. (This is a difference from Perl.)
2308    
2309    28. When duplicate subpattern names are present (necessarily with different
2310        numbers, as required by 27 above), and a test is made by name in a
2311        conditional pattern, either for a subpattern having been matched, or for
2312        recursion in such a pattern, all the associated numbered subpatterns are
2313        tested, and the overall condition is true if the condition is true for any
2314        one of them. This is the way Perl works, and is also more like the way
2315        testing by number works.
2316    
2317    
2318    Version 7.9 11-Apr-09
2319    ---------------------
2320    
2321    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2322        (pcretest), all targets were linked against these libraries. This included
2323        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2324        libraries. This caused unwanted dependencies to be created. This problem
2325        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2326        pcretest is linked with readline.
2327    
2328    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2329        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2330        moved outside it again, because FALSE and TRUE are already defined in AIX,
2331        but BOOL is not.
2332    
2333    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2334        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2335    
2336    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2337        hyphens) following file names and line numbers when outputting matching
2338        lines. This is not true; no spaces are inserted. I have also clarified the
2339        wording for the --colour (or --color) option.
2340    
2341    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2342        was not coloured; this is different to GNU grep, so I have changed it to be
2343        the same.
2344    
2345    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2346        each matching line was coloured. Now it goes on to look for further matches
2347        of any of the test patterns, which is the same behaviour as GNU grep.
2348    
2349    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2350        doesn't make sense to accept an empty string match in pcregrep, so I have
2351        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2352        seems to be how GNU grep behaves. [But see later change 40 for release
2353        8.33.]
2354    
2355    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2356        start or after a newline", because the conditional assertion was not being
2357        correctly handled. The rule now is that both the assertion and what follows
2358        in the first alternative must satisfy the test.
2359    
2360    9.  If auto-callout was enabled in a pattern with a conditional group whose
2361        condition was an assertion, PCRE could crash during matching, both with
2362        pcre_exec() and pcre_dfa_exec().
2363    
2364    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2365        used for matching.
2366    
2367    11. Unicode property support in character classes was not working for
2368        characters (bytes) greater than 127 when not in UTF-8 mode.
2369    
2370    12. Added the -M command line option to pcretest.
2371    
2372    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2373    
2374    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2375    
2376    16. Added comments and documentation about mis-use of no_arg in the C++
2377        wrapper.
2378    
2379    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2380        from Martin Jerabek that uses macro names for all relevant character and
2381        string constants.
2382    
2383    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2384        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2385        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2386        these, but not everybody uses configure.
2387    
2388    19. A conditional group that had only one branch was not being correctly
2389        recognized as an item that could match an empty string. This meant that an
2390        enclosing group might also not be so recognized, causing infinite looping
2391        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2392        with the subject "ab", where knowledge that the repeated group can match
2393        nothing is needed in order to break the loop.
2394    
2395    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2396        exec(), but without supplying a callout function, matching went wrong.
2397    
2398    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2399        leak if the size of the offset vector was greater than 30. When the vector
2400        is smaller, the saved offsets during recursion go onto a local stack
2401        vector, but for larger vectors malloc() is used. It was failing to free
2402        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2403        error, in fact).
2404    
2405    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2406        heapframe that is used only when UTF-8 support is enabled. This caused no
2407        problem, but was untidy.
2408    
2409    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2410        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2411        included within another project.
2412    
2413    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2414        slightly modified by me:
2415    
2416          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2417              not building pcregrep.
2418    
2419          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2420              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2421    
2422    25. Forward references, both numeric and by name, in patterns that made use of
2423        duplicate group numbers, could behave incorrectly or give incorrect errors,
2424        because when scanning forward to find the reference group, PCRE was not
2425        taking into account the duplicate group numbers. A pattern such as
2426        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2427    
2428    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2429        the feature of a custom pointer more persuasive (as requested by a user).
2430    
2431    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2432        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2433        pre-defined.
2434    
2435    28. Added support for (*UTF8) at the start of a pattern.
2436    
2437    29. Arrange for flags added by the "release type" setting in CMake to be shown
2438        in the configuration summary.
2439    
2440    
2441    Version 7.8 05-Sep-08
2442    ---------------------
2443    
2444    1.  Replaced UCP searching code with optimized version as implemented for Ad
2445        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2446        stage table and inline lookup instead of a function, giving speed ups of 2
2447        to 5 times on some simple patterns that I tested. Permission was given to
2448        distribute the MultiStage2.py script that generates the tables (it's not in
2449        the tarball, but is in the Subversion repository).
2450    
2451    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2452        scripts.
2453    
2454    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2455        a group with a zero qualifier. The result of the study could be incorrect,
2456        or the function might crash, depending on the pattern.
2457    
2458    4.  Caseless matching was not working for non-ASCII characters in back
2459        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2460        It now works when Unicode Property Support is available.
2461    
2462    5.  In pcretest, an escape such as \x{de} in the data was always generating
2463        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2464        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2465        truncation.
2466    
2467    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2468    
2469    7.  Added two (int) casts to pcregrep when printing the difference of two
2470        pointers, in case they are 64-bit values.
2471    
2472    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2473        test 2 if it fails.
2474    
2475    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2476        and a #define of that name to empty if it is not externally set. This is to
2477        allow users of MSVC to set it if necessary.
2478    
2479    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2480        the convenience functions in the pcre_get.c source file.
2481    
2482    11. An option change at the start of a pattern that had top-level alternatives
2483        could cause overwriting and/or a crash. This command provoked a crash in
2484        some environments:
2485    
2486          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2487    
2488        This potential security problem was recorded as CVE-2008-2371.
2489    
2490    12. For a pattern where the match had to start at the beginning or immediately
2491        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2492        pcre_dfa_exec() could read past the end of the passed subject if there was
2493        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2494        pcretest so that it places the subject at the end of its malloc-ed buffer.
2495    
2496    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2497        exec() might read past the end of the data buffer in UTF-8 mode.
2498    
2499    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2500        the data contained the byte 0x85 as part of a UTF-8 character within its
2501        first line. This applied both to normal and DFA matching.
2502    
2503    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2504        /^[^d]*?$/8 failed to match "abc".
2505    
2506    16. Added a missing copyright notice to pcrecpp_internal.h.
2507    
2508    17. Make it more clear in the documentation that values returned from
2509        pcre_exec() in ovector are byte offsets, not character counts.
2510    
2511    18. Tidied a few places to stop certain compilers from issuing warnings.
2512    
2513    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2514        supplied by Stefan Weber. I made a further small update for 7.8 because
2515        there is a change of source arrangements: the pcre_searchfuncs.c module is
2516        replaced by pcre_ucd.c.
2517    
2518    
2519    Version 7.7 07-May-08
2520    ---------------------
2521    
2522    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2523        a string to a long long, pretend we don't even have a long long." This is
2524        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2525    
2526    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2527        pre-7.6 versions, which defined a global no_arg variable instead of putting
2528        it in the RE class. (See also #8 below.)
2529    
2530    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2531        Lopes.
2532    
2533    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2534    
2535        (1) The include/exclude patterns were being applied to the whole pathnames
2536            of files, instead of just to the final components.
2537    
2538        (2) If there was more than one level of directory, the subdirectories were
2539            skipped unless they satisfied the include/exclude conditions. This is
2540            inconsistent with GNU grep (and could even be seen as contrary to the
2541            pcregrep specification - which I improved to make it absolutely clear).
2542            The action now is always to scan all levels of directory, and just
2543            apply the include/exclude patterns to regular files.
2544    
2545    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2546        --exclude_dir in the tests to avoid scanning .svn directories.
2547    
2548    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2549        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2550        doesn't support NULs in patterns.
2551    
2552    7.  Added some missing "const"s to declarations of static tables in
2553        pcre_compile.c and pcre_dfa_exec.c.
2554    
2555    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2556        caused by fix #2  above. (Subsequently also a second patch to fix the
2557        first patch. And a third patch - this was a messy problem.)
2558    
2559    9.  Applied Craig's patch to remove the use of push_back().
2560    
2561    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2562        matching function regexec().
2563    
2564    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2565        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2566        references. PCRE supports relative numbers with this syntax (I don't think
2567        Oniguruma does).
2568    
2569    12. Previously, a group with a zero repeat such as (...){0} was completely
2570        omitted from the compiled regex. However, this means that if the group
2571        was called as a subroutine from elsewhere in the pattern, things went wrong
2572        (an internal error was given). Such groups are now left in the compiled
2573        pattern, with a new opcode that causes them to be skipped at execution
2574        time.
2575    
2576    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2577        to the way PCRE behaves:
2578    
2579        (a) A lone ] character is dis-allowed (Perl treats it as data).
2580    
2581        (b) A back reference to an unmatched subpattern matches an empty string
2582            (Perl fails the current match path).
2583    
2584        (c) A data ] in a character class must be notated as \] because if the
2585            first data character in a class is ], it defines an empty class. (In
2586            Perl it is not possible to have an empty class.) The empty class []
2587            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2588            The negative empty class [^] matches any one character, independently
2589            of the DOTALL setting.
2590    
2591    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2592        non-existent subpattern following a character class starting with ']' and
2593        containing () gave an internal compiling error instead of "reference to
2594        non-existent subpattern". Fortunately, when the pattern did exist, the
2595        compiled code was correct. (When scanning forwards to check for the
2596        existence of the subpattern, it was treating the data ']' as terminating
2597        the class, so got the count wrong. When actually compiling, the reference
2598        was subsequently set up correctly.)
2599    
2600    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2601        it was being rejected as not supported by pcre_dfa_exec(), even though
2602        other assertions are supported. I have made pcre_dfa_exec() support
2603        (*FAIL).
2604    
2605    16. The implementation of 13c above involved the invention of a new opcode,
2606        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2607        cannot be changed at match time, I realized I could make a small
2608        improvement to matching performance by compiling OP_ALLANY instead of
2609        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2610        on the OP_ANY path.
2611    
2612    17. Compiling pcretest on Windows with readline support failed without the
2613        following two fixes: (1) Make the unistd.h include conditional on
2614        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2615    
2616    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2617        ncurses library to be included for pcretest when ReadLine support is
2618        requested, but also to allow for it to be overridden. This patch came from
2619        Daniel Bergström.
2620    
2621    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2622        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2623        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2624        spotting this.
2625    
2626    
2627    Version 7.6 28-Jan-08
2628    ---------------------
2629    
2630    1.  A character class containing a very large number of characters with
2631        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2632        overflow.
2633    
2634    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2635        HAVE_LONG_LONG is not defined.
2636    
2637    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2638        bring them up to date and include new features. This patch includes:
2639    
2640        - Fixed PH's badly added libz and libbz2 support.
2641        - Fixed a problem with static linking.
2642        - Added pcredemo. [But later removed - see 7 below.]
2643        - Fixed dftables problem and added an option.
2644        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2645            HAVE_LONG_LONG.
2646        - Added readline support for pcretest.
2647        - Added an listing of the option settings after cmake has run.
2648    
2649    4.  A user submitted a patch to Makefile that makes it easy to create
2650        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2651        Makefile.am that cause it to include this special target, without
2652        affecting anything else. Note that the same mingw target plus all
2653        the other distribution libraries and programs are now supported
2654        when configuring with CMake (see 6 below) instead of with
2655        Configure/Make.
2656    
2657    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2658        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2659        exported in the Windows port". It has not yet been confirmed that the patch
2660        solves the problem, but it does no harm.
2661    
2662    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2663        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2664        with CMake, and also correct the comment about stack recursion.
2665    
2666    7.  Remove the automatic building of pcredemo from the ./configure system and
2667        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2668        of a program that users should build themselves after PCRE is installed, so
2669        building it automatically is not really right. What is more, it gave
2670        trouble in some build environments.
2671    
2672    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2673    
2674    
2675    Version 7.5 10-Jan-08
2676    ---------------------
2677    
2678    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2679        values in parens when parsing an RE using the C++ wrapper."
2680    
2681    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2682        Characters greater than 255 were excluded from the class instead of being
2683        included.
2684    
2685    3.  The same bug as (2) above applied to negated POSIX classes such as
2686        [:^space:].
2687    
2688    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2689        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2690        I have changed it.
2691    
2692    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2693        first named subpattern) and a construct such as (?&a) would reference the
2694        first named subpattern whose name started with "a" (in other words, the
2695        length check was missing). Both these problems are fixed. "Subpattern name
2696        expected" is now given for (?&) (a zero-length name), and this patch also
2697        makes it give the same error for \k'' (previously it complained that that
2698        was a reference to a non-existent subpattern).
2699    
2700    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2701        this is right because (?- can be followed by option settings as well as by
2702        digits. I have, however, made the messages clearer.
2703    
2704    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2705        than the number used in the conditional) now cause a compile-time error.
2706        This is actually not compatible with Perl, which accepts such patterns, but
2707        treats the conditional as always being FALSE (as PCRE used to), but it
2708        seems to me that giving a diagnostic is better.
2709    
2710    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2711        and messages.
2712    
2713    9.  Fix two occurrences of "backslash" in comments that should have been
2714        "backspace".
2715    
2716    10. Remove two redundant lines of code that can never be obeyed (their function
2717        was moved elsewhere).
2718    
2719    11. The program that makes PCRE's Unicode character property table had a bug
2720        which caused it to generate incorrect table entries for sequences of
2721        characters that have the same character type, but are in different scripts.
2722        It amalgamated them into a single range, with the script of the first of
2723        them. In other words, some characters were in the wrong script. There were
2724        thirteen such cases, affecting characters in the following ranges:
2725    
2726          U+002b0 - U+002c1
2727          U+0060c - U+0060d
2728          U+0061e - U+00612
2729          U+0064b - U+0065e
2730          U+0074d - U+0076d
2731          U+01800 - U+01805
2732          U+01d00 - U+01d77
2733          U+01d9b - U+01dbf
2734          U+0200b - U+0200f
2735          U+030fc - U+030fe
2736          U+03260 - U+0327f
2737          U+0fb46 - U+0fbb1
2738          U+10450 - U+1049d
2739    
2740    12. The -o option (show only the matching part of a line) for pcregrep was not
2741        compatible with GNU grep in that, if there was more than one match in a
2742        line, it showed only the first of them. It now behaves in the same way as
2743        GNU grep.
2744    
2745    13. If the -o and -v options were combined for pcregrep, it printed a blank
2746        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2747        does the same. The return code can be used to tell if there were any
2748        non-matching lines.
2749    
2750    14. Added --file-offsets and --line-offsets to pcregrep.
2751    
2752    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2753        infinitely looping recursion. The bug was that positive lookaheads were not
2754        being skipped when checking for a possible empty match (negative lookaheads
2755        and both kinds of lookbehind were skipped).
2756    
2757    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2758        inclusion of <windows.h> to before rather than after the definition of
2759        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2760    
2761    17. Specifying a possessive quantifier with a specific limit for a Unicode
2762        character property caused pcre_compile() to compile bad code, which led at
2763        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2764        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2765        caused the error; without that there was no problem.
2766    
2767    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2768    
2769    19. Added --enable-pcretest-libreadline.
2770    
2771    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2772        RE::GlobalReplace(). As a result, the number of replacements returned was
2773        double what it should be. I removed one of the increments, but Craig sent a
2774        later patch that removed the other one (the right fix) and added unit tests
2775        that check the return values (which was not done before).
2776    
2777    21. Several CMake things:
2778    
2779        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2780            the names libpcre and libpcreposix, not just pcre and pcreposix.
2781    
2782        (2) The above change means that pcretest and pcregrep are now correctly
2783            linked with the newly-built libraries, not previously installed ones.
2784    
2785        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2786    
2787    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2788        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2789        UTF-8 newline character). The key issue is that the pattern starts .*;
2790        this means that the match must be either at the beginning, or after a
2791        newline. The bug was in the code for advancing after a failed match and
2792        checking that the new position followed a newline. It was not taking
2793        account of UTF-8 characters correctly.
2794    
2795    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2796        character classes. PCRE was not treating the sequence [:...:] as a
2797        character class unless the ... were all letters. Perl, however, seems to
2798        allow any characters between [: and :], though of course it rejects as
2799        unknown any "names" that contain non-letters, because all the known class
2800        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2801        for example, whereas PCRE did not - it did not recognize a POSIX character
2802        class. This seemed a bit dangerous, so the code has been changed to be
2803        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2804        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2805        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2806        Perl does, and where it didn't before.
2807    
2808    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2809        Windows environments %n is disabled by default.
2810    
2811    
2812    Version 7.4 21-Sep-07
2813    ---------------------
2814    
2815    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2816        means that a class such as [\s] counted as "explicit reference to CR or
2817        LF". That isn't really right - the whole point of the change was to try to
2818        help when there was an actual mention of one of the two characters. So now
2819        the change happens only if \r or \n (or a literal CR or LF) character is
2820        encountered.
2821    
2822    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2823        of both had grown to the point where there were only 3 bits left.
2824        Fortunately, there was spare space in the data structure, and so I have
2825        moved the internal flags into a new 16-bit field to free up more option
2826        bits.
2827    
2828    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2829        but did not set the internal JCHANGED flag - either of these is enough to
2830        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2831        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2832        start sets both bits.
2833    
2834    4.  Added options (at build time, compile time, exec time) to change \R from
2835        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2836    
2837    5.  doc/pcresyntax.html was missing from the distribution.
2838    
2839    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2840        compatibility, even though it is no longer used.
2841    
2842    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2843        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2844        windows.h file is present (where different names are used). [This was
2845        reversed later after testing - see 16 below.]
2846    
2847    8.  Changed all #include <config.h> to #include "config.h". There were also
2848        some further <pcre.h> cases that I changed to "pcre.h".
2849    
2850    9.  When pcregrep was used with the --colour option, it missed the line ending
2851        sequence off the lines that it output.
2852    
2853    10. It was pointed out to me that arrays of string pointers cause lots of
2854        relocations when a shared library is dynamically loaded. A technique of
2855        using a single long string with a table of offsets can drastically reduce
2856        these. I have refactored PCRE in four places to do this. The result is
2857        dramatic:
2858    
2859          Originally:                          290
2860          After changing UCP table:            187
2861          After changing error message table:   43
2862          After changing table of "verbs"       36
2863          After changing table of Posix names   22
2864    
2865        Thanks to the folks working on Gregex for glib for this insight.
2866    
2867    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2868        unicode-properties was also set.
2869    
2870    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2871    
2872    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2873        checked only for CRLF.
2874    
2875    14. Added casts to pcretest.c to avoid compiler warnings.
2876    
2877    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2878    
2879    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2880        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2881        entirely. This removes changes made in 7 above.
2882    
2883    17. The CMake files have been updated, and there is now more information about
2884        building with CMake in the NON-UNIX-USE document.
2885    
2886    
2887    Version 7.3 28-Aug-07
2888  ---------------------  ---------------------
2889    
2890   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 143  Version 7.3 17-Aug-07 Line 3026  Version 7.3 17-Aug-07
3026  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3027      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3028    
3029    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3030    
3031    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3032        character were causing crashes (broken optimization).
3033    
3034    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3035        \p or \P) caused a compile-time loop.
3036    
3037    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3038        break. For example, the unstudied pattern [\r\n]A does not match the string
3039        "\r\nA" because change 7.0/46 below moves the current point on by two
3040        characters after failing to match at the start. However, the pattern \nA
3041        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3042        the same is true. There doesn't seem any very clean way out of this, but
3043        what I have chosen to do makes the common cases work: PCRE now takes note
3044        of whether there can be an explicit match for \r or \n anywhere in the
3045        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3046        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3047        pattern has explicit CR or LF references.
3048    
3049    29. Added (*CR) etc for changing newline setting at start of pattern.
3050    
3051    
3052  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3053  ---------------------  ---------------------

Legend:
Removed from v.221  
changed lines
  Added in v.1576

  ViewVC Help
Powered by ViewVC 1.1.5