/[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 1591 by zherczeg, Tue Aug 11 05:41:09 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    21. Fix infinite recursion in the JIT compiler when certain patterns such as
85        /(?:|a|){100}x/ are analysed.
86    
87    22. Some patterns with character classes involving [: and \\ were incorrectly
88        compiled and could cause reading from uninitialized memory or an incorrect
89        error diagnosis.
90    
91    23. Pathological patterns containing many nested occurrences of [: caused
92        pcre_compile() to run for a very long time.
93    
94    24. A conditional group with only one branch has an implicit empty alternative
95        branch and must therefore be treated as potentially matching an empty
96        string.
97    
98    25. If (?R was followed by - or + incorrect behaviour happened instead of a
99        diagnostic.
100    
101    26. Arrange to give up on finding the minimum matching length for overly
102        complex patterns.
103    
104    27. Similar to (4) above: in a pattern with duplicated named groups and an
105        occurrence of (?| it is possible for an apparently non-recursive back
106        reference to become recursive if a later named group with the relevant
107        number is encountered. This could lead to a buffer overflow. Wen Guanxing
108        from Venustech ADLAB discovered this bug.
109    
110    28. If pcregrep was given the -q option with -c or -l, or when handling a
111        binary file, it incorrectly wrote output to stdout.
112    
113    29. The JIT compiler did not restore the control verb head in case of *THEN
114        control verbs. This issue was found by Karl Skomski with a custom LLVM
115        fuzzer.
116    
117    30. Error messages for syntax errors following \g and \k were giving inaccurate
118        offsets in the pattern.
119    
120    31. Added a check for integer overflow in conditions (?(<digits>) and
121        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122        fuzzer.
123    
124    32. Handling recursive references such as (?2) when the reference is to a group
125        later in the pattern uses code that is very hacked about and error-prone.
126        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
127        give an internal error if it is obvious that compiling has gone wrong.
128    
129    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
130        This issue was found by Karl Skomski with a custom LLVM fuzzer.
131    
132    
133    Version 8.37 28-April-2015
134    --------------------------
135    
136    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
137        for those parentheses to be closed with whatever has been captured so far.
138        However, it was failing to mark any other groups between the hightest
139        capture so far and the currrent group as "unset". Thus, the ovector for
140        those groups contained whatever was previously there. An example is the
141        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
142    
143    2.  If an assertion condition was quantified with a minimum of zero (an odd
144        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
145    
146    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
147        unrecognized modifier, a crash could occur.
148    
149    4.  An attempt to do global matching in pcretest with a zero-length ovector
150        caused a crash.
151    
152    5.  Fixed a memory leak during matching that could occur for a subpattern
153        subroutine call (recursive or otherwise) if the number of captured groups
154        that had to be saved was greater than ten.
155    
156    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
157        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
158        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
159    
160    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
161        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
162    
163    8. When a pattern is compiled, it remembers the highest back reference so that
164       when matching, if the ovector is too small, extra memory can be obtained to
165       use instead. A conditional subpattern whose condition is a check on a
166       capture having happened, such as, for example in the pattern
167       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
168       setting the highest backreference number. This mattered only if pcre_exec()
169       was called with an ovector that was too small to hold the capture, and there
170       was no other kind of back reference (a situation which is probably quite
171       rare). The effect of the bug was that the condition was always treated as
172       FALSE when the capture could not be consulted, leading to a incorrect
173       behaviour by pcre_exec(). This bug has been fixed.
174    
175    9. A reference to a duplicated named group (either a back reference or a test
176       for being set in a conditional) that occurred in a part of the pattern where
177       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
178       to be incorrectly calculated, leading to overwriting.
179    
180    10. A mutually recursive set of back references such as (\2)(\1) caused a
181        segfault at study time (while trying to find the minimum matching length).
182        The infinite loop is now broken (with the minimum length unset, that is,
183        zero).
184    
185    11. If an assertion that was used as a condition was quantified with a minimum
186        of zero, matching went wrong. In particular, if the whole group had
187        unlimited repetition and could match an empty string, a segfault was
188        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
189        assertions to be quantified, but not if they are being used as conditions,
190        so the above pattern is faulted by Perl. PCRE has now been changed so that
191        it also rejects such patterns.
192    
193    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
194        failed to allow the zero-repeat case if pcre2_exec() was called with an
195        ovector too small to capture the group.
196    
197    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
198        Red Hat Product Security:
199    
200        (a) A crash if /K and /F were both set with the option to save the compiled
201        pattern.
202    
203        (b) Another crash if the option to print captured substrings in a callout
204        was combined with setting a null ovector, for example \O\C+ as a subject
205        string.
206    
207    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
208        forward reference repeated a large (but limited) number of times within a
209        repeated outer group that has a zero minimum quantifier, caused incorrect
210        code to be compiled, leading to the error "internal error:
211        previously-checked referenced subpattern not found" when an incorrect
212        memory address was read. This bug was reported as "heap overflow",
213        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
214        CVE-2015-2325.
215    
216    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
217        call within a group that also contained a recursive back reference caused
218        incorrect code to be compiled. This bug was reported as "heap overflow",
219        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
220        number CVE-2015-2326.
221    
222    24. Computing the size of the JIT read-only data in advance has been a source
223        of various issues, and new ones are still appear unfortunately. To fix
224        existing and future issues, size computation is eliminated from the code,
225        and replaced by on-demand memory allocation.
226    
227    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
228        adjacent to the end of the range, and the range contained characters with
229        more than one other case, caused incorrect behaviour when compiled in UTF
230        mode. In that example, the range a-j was left out of the class.
231    
232    26. Fix JIT compilation of conditional blocks, which assertion
233        is converted to (*FAIL). E.g: /(?(?!))/.
234    
235    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
236        discovered by the LLVM fuzzer.
237    
238    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
239        when this assertion was used as a condition, for example (?(?!)a|b). In
240        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
241        error about an unsupported item.
242    
243    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
244        possessification code could take exponential time to complete. A recursion
245        depth limit of 1000 has been imposed to limit the resources used by this
246        optimization.
247    
248    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
249        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
250        because \S ensures they are all in the class. The code for doing this was
251        interacting badly with the code for computing the amount of space needed to
252        compile the pattern, leading to a buffer overflow. This bug was discovered
253        by the LLVM fuzzer.
254    
255    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
256        other kinds of group caused stack overflow at compile time. This bug was
257        discovered by the LLVM fuzzer.
258    
259    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
260        between a subroutine call and its quantifier was incorrectly compiled,
261        leading to buffer overflow or other errors. This bug was discovered by the
262        LLVM fuzzer.
263    
264    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
265        assertion after (?(. The code was failing to check the character after
266        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
267        was discovered by the LLVM fuzzer.
268    
269    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
270        a fixed maximum following a group that contains a subroutine reference was
271        incorrectly compiled and could trigger buffer overflow. This bug was
272        discovered by the LLVM fuzzer.
273    
274    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
275        caused a stack overflow instead of the diagnosis of a non-fixed length
276        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
277    
278    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
279        (e.g. /(?<=\Ka)/) could make pcregrep loop.
280    
281    37. There was a similar problem to 36 in pcretest for global matches.
282    
283    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
284        and a subsequent item in the pattern caused a non-match, backtracking over
285        the repeated \X did not stop, but carried on past the start of the subject,
286        causing reference to random memory and/or a segfault. There were also some
287        other cases where backtracking after \C could crash. This set of bugs was
288        discovered by the LLVM fuzzer.
289    
290    39. The function for finding the minimum length of a matching string could take
291        a very long time if mutual recursion was present many times in a pattern,
292        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
293        method has been implemented. This infelicity was discovered by the LLVM
294        fuzzer.
295    
296    40. Static linking against the PCRE library using the pkg-config module was
297        failing on missing pthread symbols.
298    
299    
300    Version 8.36 26-September-2014
301    ------------------------------
302    
303    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
304        -Wmissing-field-initializers and -Wunused-parameter.
305    
306    2.  The tests for quantifiers being too big (greater than 65535) were being
307        applied after reading the number, and stupidly assuming that integer
308        overflow would give a negative number. The tests are now applied as the
309        numbers are read.
310    
311    3.  Tidy code in pcre_exec.c where two branches that used to be different are
312        now the same.
313    
314    4.  The JIT compiler did not generate match limit checks for certain
315        bracketed expressions with quantifiers. This may lead to exponential
316        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
317        issue should be resolved now.
318    
319    5.  Fixed an issue, which occures when nested alternatives are optimized
320        with table jumps.
321    
322    6.  Inserted two casts and changed some ints to size_t in the light of some
323        reported 64-bit compiler warnings (Bugzilla 1477).
324    
325    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
326        an empty string, which sometimes were behaving incorrectly in the
327        interpreter (though correctly in the JIT matcher). This pcretest input is
328        an example:
329    
330          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
331          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
332    
333        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
334        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
335        for an empty string was breaking the inner loop and carrying on at a lower
336        level, when possessive repeated groups should always return to a higher
337        level as they have no backtrack points in them. The empty string test now
338        occurs at the outer level.
339    
340    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
341        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
342    
343    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
344        doesn't).
345    
346    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
347        to hold a pattern to be incorrectly computed (too small) when there were
348        named back references to duplicated names. This could cause "internal
349        error: code overflow" or "double free or corruption" or other memory
350        handling errors.
351    
352    11. When named subpatterns had the same prefixes, back references could be
353        confused. For example, in this pattern:
354    
355          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
356    
357        the reference to 'Name' was incorrectly treated as a reference to a
358        duplicate name.
359    
360    12. A pattern such as /^s?c/mi8 where the optional character has more than
361        one "other case" was incorrectly compiled such that it would only try to
362        match starting at "c".
363    
364    13. When a pattern starting with \s was studied, VT was not included in the
365        list of possible starting characters; this should have been part of the
366        8.34/18 patch.
367    
368    14. If a character class started [\Qx]... where x is any character, the class
369        was incorrectly terminated at the ].
370    
371    15. If a pattern that started with a caseless match for a character with more
372        than one "other case" was studied, PCRE did not set up the starting code
373        unit bit map for the list of possible characters. Now it does. This is an
374        optimization improvement, not a bug fix.
375    
376    16. The Unicode data tables have been updated to Unicode 7.0.0.
377    
378    17. Fixed a number of memory leaks in pcregrep.
379    
380    18. Avoid a compiler warning (from some compilers) for a function call with
381        a cast that removes "const" from an lvalue by using an intermediate
382        variable (to which the compiler does not object).
383    
384    19. Incorrect code was compiled if a group that contained an internal recursive
385        back reference was optional (had quantifier with a minimum of zero). This
386        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
387        caused segmentation faults because of stack overflows at compile time.
388    
389    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
390        group that is quantified with an indefinite repeat, caused a compile-time
391        loop which used up all the system stack and provoked a segmentation fault.
392        This was not the same bug as 19 above.
393    
394    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
395        Patch by Mike Frysinger.
396    
397    
398    Version 8.35 04-April-2014
399    --------------------------
400    
401    1.  A new flag is set, when property checks are present in an XCLASS.
402        When this flag is not set, PCRE can perform certain optimizations
403        such as studying these XCLASS-es.
404    
405    2.  The auto-possessification of character sets were improved: a normal
406        and an extended character set can be compared now. Furthermore
407        the JIT compiler optimizes more character set checks.
408    
409    3.  Got rid of some compiler warnings for potentially uninitialized variables
410        that show up only when compiled with -O2.
411    
412    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
413        of a match later then the end of the match. The pcretest program was not
414        handling the case sensibly - it was outputting from the start to the next
415        binary zero. It now reports this situation in a message, and outputs the
416        text from the end to the start.
417    
418    5.  Fast forward search is improved in JIT. Instead of the first three
419        characters, any three characters with fixed position can be searched.
420        Search order: first, last, middle.
421    
422    6.  Improve character range checks in JIT. Characters are read by an inprecise
423        function now, which returns with an unknown value if the character code is
424        above a certain threshold (e.g: 256). The only limitation is that the value
425        must be bigger than the threshold as well. This function is useful when
426        the characters above the threshold are handled in the same way.
427    
428    7.  The macros whose names start with RAWUCHAR are placeholders for a future
429        mode in which only the bottom 21 bits of 32-bit data items are used. To
430        make this more memorable for those maintaining the code, the names have
431        been changed to start with UCHAR21, and an extensive comment has been added
432        to their definition.
433    
434    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
435        to the export list in Makefile.am (they were accidentally omitted from the
436        8.34 tarball).
437    
438    9.  The informational output from pcretest used the phrase "starting byte set"
439        which is inappropriate for the 16-bit and 32-bit libraries. As the output
440        for "first char" and "need char" really means "non-UTF-char", I've changed
441        "byte" to "char", and slightly reworded the output. The documentation about
442        these values has also been (I hope) clarified.
443    
444    10. Another JIT related optimization: use table jumps for selecting the correct
445        backtracking path, when more than four alternatives are present inside a
446        bracket.
447    
448    11. Empty match is not possible, when the minimum length is greater than zero,
449        and there is no \K in the pattern. JIT should avoid empty match checks in
450        such cases.
451    
452    12. In a caseless character class with UCP support, when a character with more
453        than one alternative case was not the first character of a range, not all
454        the alternative cases were added to the class. For example, s and \x{17f}
455        are both alternative cases for S: the class [RST] was handled correctly,
456        but [R-T] was not.
457    
458    13. The configure.ac file always checked for pthread support when JIT was
459        enabled. This is not used in Windows, so I have put this test inside a
460        check for the presence of windows.h (which was already tested for).
461    
462    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
463        The algorithm provides a way to skip certain starting offsets, and usually
464        faster than linear prefix searches.
465    
466    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
467        as for 'fr_FR' and 'french'. For some reason, however, it then used the
468        Windows-specific input and output files, which have 'french' screwed in.
469        So this could never have worked. One of the problems with locales is that
470        they aren't always the same. I have now updated RunTest so that it checks
471        the output of the locale test (test 3) against three different output
472        files, and it allows the test to pass if any one of them matches. With luck
473        this should make the test pass on some versions of Solaris where it was
474        failing. Because of the uncertainty, the script did not used to stop if
475        test 3 failed; it now does. If further versions of a French locale ever
476        come to light, they can now easily be added.
477    
478    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
479        there was a message during ./configure, but it did not stop. This now
480        provokes an error. The invalid example in README has been corrected.
481        If a value less than the minimum is given, the minimum value has always
482        been used, but now a warning is given.
483    
484    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
485        was a bug in the test system, which is now fixed. Also, the list of various
486        configurations that are tested for each release did not have one with both
487        16/32 bits and --enable-bar-anycrlf. It now does.
488    
489    18. pcretest was missing "-C bsr" for displaying the \R default setting.
490    
491    19. Little endian PowerPC systems are supported now by the JIT compiler.
492    
493    20. The fast forward newline mechanism could enter to an infinite loop on
494        certain invalid UTF-8 input. Although we don't support these cases
495        this issue can be fixed by a performance optimization.
496    
497    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
498        not take account if existing stack usage. There is now a new global
499        variable called pcre_stack_guard that can be set to point to an external
500        function to check stack availability. It is called at the start of
501        processing every parenthesized group.
502    
503    22. A typo in the code meant that in ungreedy mode the max/min qualifier
504        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
505        not match "ab".
506    
507    23. When UTF was disabled, the JIT program reported some incorrect compile
508        errors. These messages are silenced now.
509    
510    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
511        compiler.
512    
513    25. Change all the temporary files used in RunGrepTest to be different to those
514        used by RunTest so that the tests can be run simultaneously, for example by
515        "make -j check".
516    
517    
518    Version 8.34 15-December-2013
519    -----------------------------
520    
521    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
522        executable memory. Patch inspired by Carsten Klein.
523    
524    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
525        this macro is never tested and has no effect, because the work to support
526        coverage involves only compiling and linking options and special targets in
527        the Makefile. The comment in config.h implied that defining the macro would
528        enable coverage support, which is totally false. There was also support for
529        setting this macro in the CMake files (my fault, I just copied it from
530        configure). SUPPORT_GCOV has now been removed.
531    
532    3.  Make a small performance improvement in strlen16() and strlen32() in
533        pcretest.
534    
535    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
536        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
537        these cases). There was also one in pcretest.c.
538    
539    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
540    
541    6.  In UTF mode, the code for checking whether a group could match an empty
542        string (which is used for indefinitely repeated groups to allow for
543        breaking an infinite loop) was broken when the group contained a repeated
544        negated single-character class with a character that occupied more than one
545        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
546        UTF-8 mode). The effect was undefined: the group might or might not be
547        deemed as matching an empty string, or the program might have crashed.
548    
549    7.  The code for checking whether a group could match an empty string was not
550        recognizing that \h, \H, \v, \V, and \R must match a character.
551    
552    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
553        an empty string. If it can, pcretest shows this in its information output.
554    
555    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
556        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
557        matched by pcre_exec() without using JIT:
558    
559        (a) If the rest of the pattern did not match after a maximal run of
560            grapheme clusters, the code for backing up to try with fewer of them
561            did not always back up over a full grapheme when characters that do not
562            have the modifier quality were involved, e.g. Hangul syllables.
563    
564        (b) If the match point in a subject started with modifier character, and
565            there was no match, the code could incorrectly back up beyond the match
566            point, and potentially beyond the first character in the subject,
567            leading to a segfault or an incorrect match result.
568    
569    10. A conditional group with an assertion condition could lead to PCRE
570        recording an incorrect first data item for a match if no other first data
571        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
572        first data item, and therefore matched "ca" after "c" instead of at the
573        start.
574    
575    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
576        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
577    
578    12. The source of pcregrep now includes z/OS-specific code so that it can be
579        compiled for z/OS as part of the special z/OS distribution.
580    
581    13. Added the -T and -TM options to pcretest.
582    
583    14. The code in pcre_compile.c for creating the table of named capturing groups
584        has been refactored. Instead of creating the table dynamically during the
585        actual compiling pass, the information is remembered during the pre-compile
586        pass (on the stack unless there are more than 20 named groups, in which
587        case malloc() is used) and the whole table is created before the actual
588        compile happens. This has simplified the code (it is now nearly 150 lines
589        shorter) and prepared the way for better handling of references to groups
590        with duplicate names.
591    
592    15. A back reference to a named subpattern when there is more than one of the
593        same name now checks them in the order in which they appear in the pattern.
594        The first one that is set is used for the reference. Previously only the
595        first one was inspected. This change makes PCRE more compatible with Perl.
596    
597    16. Unicode character properties were updated from Unicode 6.3.0.
598    
599    17. The compile-time code for auto-possessification has been refactored, based
600        on a patch by Zoltan Herczeg. It now happens after instead of during
601        compilation. The code is cleaner, and more cases are handled. The option
602        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
603        options in pcretest are provided to set it. It can also be set by
604        (*NO_AUTO_POSSESS) at the start of a pattern.
605    
606    18. The character VT has been added to the default ("C" locale) set of
607        characters that match \s and are generally treated as white space,
608        following this same change in Perl 5.18. There is now no difference between
609        "Perl space" and "POSIX space". Whether VT is treated as white space in
610        other locales depends on the locale.
611    
612    19. The code for checking named groups as conditions, either for being set or
613        for being recursed, has been refactored (this is related to 14 and 15
614        above). Processing unduplicated named groups should now be as fast at
615        numerical groups, and processing duplicated groups should be faster than
616        before.
617    
618    20. Two patches to the CMake build system, by Alexander Barkov:
619    
620          (1) Replace the "source" command by "." in CMakeLists.txt because
621              "source" is a bash-ism.
622    
623          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
624              without these the CMake build does not work on Solaris.
625    
626    21. Perl has changed its handling of \8 and \9. If there is no previously
627        encountered capturing group of those numbers, they are treated as the
628        literal characters 8 and 9 instead of a binary zero followed by the
629        literals. PCRE now does the same.
630    
631    22. Following Perl, added \o{} to specify codepoints in octal, making it
632        possible to specify values greater than 0777 and also making them
633        unambiguous.
634    
635    23. Perl now gives an error for missing closing braces after \x{... instead of
636        treating the string as literal. PCRE now does the same.
637    
638    24. RunTest used to grumble if an inappropriate test was selected explicitly,
639        but just skip it when running all tests. This make it awkward to run ranges
640        of tests when one of them was inappropriate. Now it just skips any
641        inappropriate tests, as it always did when running all tests.
642    
643    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
644        character types such as \d or \w, too many callouts were inserted, and the
645        data that they returned was rubbish.
646    
647    26. In UCP mode, \s was not matching two of the characters that Perl matches,
648        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
649        were matched by \h. The code has now been refactored so that the lists of
650        the horizontal and vertical whitespace characters used for \h and \v (which
651        are defined only in one place) are now also used for \s.
652    
653    27. Add JIT support for the 64 bit TileGX architecture.
654        Patch by Jiong Wang (Tilera Corporation).
655    
656    28. Possessive quantifiers for classes (both explicit and automatically
657        generated) now use special opcodes instead of wrapping in ONCE brackets.
658    
659    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
660        (because it's meaningless), this was not happening when PCRE_CASELESS was
661        set. Not wrong, but inefficient.
662    
663    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
664        properties for \w, \d, etc) is present in a test regex. Otherwise if the
665        test contains no characters greater than 255, Perl doesn't realise it
666        should be using Unicode semantics.
667    
668    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
669        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
670        does in Unicode mode.
671    
672    32. Added the "forbid" facility to pcretest so that putting tests into the
673        wrong test files can sometimes be quickly detected.
674    
675    33. There is now a limit (default 250) on the depth of nesting of parentheses.
676        This limit is imposed to control the amount of system stack used at compile
677        time. It can be changed at build time by --with-parens-nest-limit=xxx or
678        the equivalent in CMake.
679    
680    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
681        errors. Perl warns for these when in warning mode, but PCRE has no facility
682        for giving warnings.
683    
684    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
685        However, this was not working for (?!) because it is optimized to (*FAIL),
686        for which PCRE does not allow quantifiers. The optimization is now disabled
687        when a quantifier follows (?!). I can't see any use for this, but it makes
688        things uniform.
689    
690    36. Perl no longer allows group names to start with digits, so I have made this
691        change also in PCRE. It simplifies the code a bit.
692    
693    37. In extended mode, Perl ignores spaces before a + that indicates a
694        possessive quantifier. PCRE allowed a space before the quantifier, but not
695        before the possessive +. It now does.
696    
697    38. The use of \K (reset reported match start) within a repeated possessive
698        group such as (a\Kb)*+ was not working.
699    
700    40. Document that the same character tables must be used at compile time and
701        run time, and that the facility to pass tables to pcre_exec() and
702        pcre_dfa_exec() is for use only with saved/restored patterns.
703    
704    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
705        features for Builds with MSVC:
706    
707        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
708           builds). The .rc files can be used to set FileDescription and many other
709           attributes.
710    
711        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
712           This allows higher-level build scripts which want .pdb files to avoid
713           hard-coding the exact files needed."
714    
715    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
716        mean "start of word" and "end of word", respectively, as a transition aid.
717    
718    43. A minimizing repeat of a class containing codepoints greater than 255 in
719        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
720        compiled to use the heap for recursion.
721    
722    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
723        is configured.
724    
725    
726    Version 8.33 28-May-2013
727    ------------------------
728    
729    1.  Added 'U' to some constants that are compared to unsigned integers, to
730        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
731        variables that are added to signed variables, to ensure the result is
732        signed and can be negated.
733    
734    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
735        CMake config files.
736    
737    3.  Revise the creation of config.h.generic so that all boolean macros are
738        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
739        overriding via -D on the command line possible.
740    
741    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
742        to unsigned int is reported to make a quite noticeable speed difference in
743        a specific Windows environment. Testing on Linux did also appear to show
744        some benefit (and it is clearly not harmful). Also fixed the definition of
745        Xop which should be unsigned.
746    
747    5.  Related to (4), changing the definition of the intermediate variable cc
748        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
749        improvements.
750    
751    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
752        unnecessary spaces.
753    
754    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
755        and later.
756    
757    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
758    
759    9.  Optimizing fast_forward_start_bits in JIT.
760    
761    10. Adding support for callouts in JIT, and fixing some issues revealed
762        during this work. Namely:
763    
764        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
765    
766        (b) Minimum length was not checked before the matching is started.
767    
768    11. The value of capture_last that is passed to callouts was incorrect in some
769        cases when there was a capture on one path that was subsequently abandoned
770        after a backtrack. Also, the capture_last value is now reset after a
771        recursion, since all captures are also reset in this case.
772    
773    12. The interpreter no longer returns the "too many substrings" error in the
774        case when an overflowing capture is in a branch that is subsequently
775        abandoned after a backtrack.
776    
777    13. In the pathological case when an offset vector of size 2 is used, pcretest
778        now prints out the matched string after a yield of 0 or 1.
779    
780    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
781        Only supported by the JIT compiler at the moment.
782    
783    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
784    
785    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
786        offset of the starting point of the matching process, provided the offsets
787        vector is large enough.
788    
789    17. The \A escape now records a lookbehind value of 1, though its execution
790        does not actually inspect the previous character. This is to ensure that,
791        in partial multi-segment matching, at least one character from the old
792        segment is retained when a new segment is processed. Otherwise, if there
793        are no lookbehinds in the pattern, \A might match incorrectly at the start
794        of a new segment.
795    
796    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
797    
798    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
799        gives some modest performance improvement in 8-bit mode.
800    
801    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
802        be expressed in certain programming languages using Universal Character
803        Names.
804    
805    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
806        which points out that "non characters" are not "characters that may not
807        appear in Unicode strings" but rather "characters that are reserved for
808        internal use and have only local meaning".
809    
810    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
811        there was a conditional group that depended on an assertion, if the
812        assertion was false, the callout that immediately followed the alternation
813        in the condition was skipped when pcre_exec() was used for matching.
814    
815    23. Allow an explicit callout to be inserted before an assertion that is the
816        condition for a conditional group, for compatibility with automatic
817        callouts, which always insert a callout at this point.
818    
819    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
820        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
821    
822    25. (*PRUNE) is now supported by the JIT compiler.
823    
824    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
825    
826    27. Fix the case where there are two or more SKIPs with arguments that may be
827        ignored.
828    
829    28. (*SKIP) is now supported by the JIT compiler.
830    
831    29. (*THEN) is now supported by the JIT compiler.
832    
833    30. Update RunTest with additional test selector options.
834    
835    31. The way PCRE handles backtracking verbs has been changed in two ways.
836    
837        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
838        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
839        backtracking. In some cases this makes it more Perl-compatible, but Perl's
840        rather obscure rules do not always do the same thing.
841    
842        (2) Previously, backtracking verbs were confined within assertions. This is
843        no longer the case for positive assertions, except for (*ACCEPT). Again,
844        this sometimes improves Perl compatibility, and sometimes does not.
845    
846    32. A number of tests that were in test 2 because Perl did things differently
847        have been moved to test 1, because either Perl or PCRE has changed, and
848        these tests are now compatible.
849    
850    32. Backtracking control verbs are now handled in the same way in JIT and
851        interpreter.
852    
853    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
854        contained a forward subroutine reference caused a compile error.
855    
856    34. Auto-detect and optimize limited repetitions in JIT.
857    
858    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
859        blocking (*UTF) etc.
860    
861    36. In the interpreter, maximizing pattern repetitions for characters and
862        character types now use tail recursion, which reduces stack usage.
863    
864    37. The value of the max lookbehind was not correctly preserved if a compiled
865        and saved regex was reloaded on a host of different endianness.
866    
867    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
868        of the compiled pattern block, expand the flags field from 16 to 32 bits
869        because it was almost full.
870    
871    39. Try madvise first before posix_madvise.
872    
873    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
874        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
875        original change locked out all matches of empty strings. This has been
876        changed so that one match of an empty string per line is recognized.
877        Subsequent searches on the same line (for colouring or for --only-matching,
878        for example) do not recognize empty strings.
879    
880    41. Applied a user patch to fix a number of spelling mistakes in comments.
881    
882    42. Data lines longer than 65536 caused pcretest to crash.
883    
884    43. Clarified the data type for length and startoffset arguments for pcre_exec
885        and pcre_dfa_exec in the function-specific man pages, where they were
886        explicitly stated to be in bytes, never having been updated. I also added
887        some clarification to the pcreapi man page.
888    
889    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
890        a segmentation fault.
891    
892    
893    Version 8.32 30-November-2012
894    -----------------------------
895    
896    1.  Improved JIT compiler optimizations for first character search and single
897        character iterators.
898    
899    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
900        Patch by Daniel Richard G.
901    
902    3.  Single character iterator optimizations in the JIT compiler.
903    
904    4.  Improved JIT compiler optimizations for character ranges.
905    
906    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
907        Reported by Giuseppe D'Angelo.
908    
909    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
910        of a line, was being set incorrectly in cases where .* appeared inside
911        atomic brackets at the start of a pattern, or where there was a subsequent
912        *PRUNE or *SKIP.
913    
914    7.  Improved instruction cache flush for POWER/PowerPC.
915        Patch by Daniel Richard G.
916    
917    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
918        grep:
919    
920        (a) There is now no limit to the number of patterns to be matched.
921    
922        (b) An error is given if a pattern is too long.
923    
924        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
925            are now supported.
926    
927        (d) --exclude-from and --include-from (multiple use) have been added.
928    
929        (e) Exclusions and inclusions now apply to all files and directories, not
930            just to those obtained from scanning a directory recursively.
931    
932        (f) Multiple uses of -f and --file-list are now supported.
933    
934        (g) In a Windows environment, the default for -d has been changed from
935            "read" (the GNU grep default) to "skip", because otherwise the presence
936            of a directory in the file list provokes an error.
937    
938        (h) The documentation has been revised and clarified in places.
939    
940    9.  Improve the matching speed of capturing brackets.
941    
942    10. Changed the meaning of \X so that it now matches a Unicode extended
943        grapheme cluster.
944    
945    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
946        out POSIX threads when JIT support is configured.
947    
948    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
949    
950    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
951        structure could go wrong in environments where size_t is not the same size
952        as int.
953    
954    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
955        set.
956    
957    15. The EBCDIC support had decayed; later updates to the code had included
958        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
959        general tidy up of EBCDIC-related issues, and the documentation was also
960        not quite right. There is now a test that can be run on ASCII systems to
961        check some of the EBCDIC-related things (but is it not a full test).
962    
963    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
964        in a small tidy to the code.
965    
966    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
967    
968    18. If the --only-matching (-o) option in pcregrep is specified multiple
969        times, each one causes appropriate output. For example, -o1 -o2 outputs the
970        substrings matched by the 1st and 2nd capturing parentheses. A separating
971        string can be specified by --om-separator (default empty).
972    
973    19. Improving the first n character searches.
974    
975    20. Turn case lists for horizontal and vertical white space into macros so that
976        they are defined only once.
977    
978    21. This set of changes together give more compatible Unicode case-folding
979        behaviour for characters that have more than one other case when UCP
980        support is available.
981    
982        (a) The Unicode property table now has offsets into a new table of sets of
983            three or more characters that are case-equivalent. The MultiStage2.py
984            script that generates these tables (the pcre_ucd.c file) now scans
985            CaseFolding.txt instead of UnicodeData.txt for character case
986            information.
987    
988        (b) The code for adding characters or ranges of characters to a character
989            class has been abstracted into a generalized function that also handles
990            case-independence. In UTF-mode with UCP support, this uses the new data
991            to handle characters with more than one other case.
992    
993        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
994            whose other case is greater than 256 are now correctly matched
995            caselessly. Previously, the high codepoint matched the low one, but not
996            vice versa.
997    
998        (d) The processing of \h, \H, \v, and \ in character classes now makes use
999            of the new class addition function, using character lists defined as
1000            macros alongside the case definitions of 20 above.
1001    
1002        (e) Caseless back references now work with characters that have more than
1003            one other case.
1004    
1005        (f) General caseless matching of characters with more than one other case
1006            is supported.
1007    
1008    22. Unicode character properties were updated from Unicode 6.2.0
1009    
1010    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1011    
1012    24. Add support for 32-bit character strings, and UTF-32
1013    
1014    25. Major JIT compiler update (code refactoring and bugfixing).
1015        Experimental Sparc 32 support is added.
1016    
1017    26. Applied a modified version of Daniel Richard G's patch to create
1018        pcre.h.generic and config.h.generic by "make" instead of in the
1019        PrepareRelease script.
1020    
1021    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1022        pcre_compile.c when checking for a zero character.
1023    
1024    28. Introducing a native interface for JIT. Through this interface, the compiled
1025        machine code can be directly executed. The purpose of this interface is to
1026        provide fast pattern matching, so several sanity checks are not performed.
1027        However, feature tests are still performed. The new interface provides
1028        1.4x speedup compared to the old one.
1029    
1030    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1031        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1032        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1033    
1034    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1035        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1036        the "old" RFC 2279). Instead, it ended up passing a negative length to
1037        pcre_exec().
1038    
1039    31. Add support for GCC's visibility feature to hide internal functions.
1040    
1041    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1042        "unknown -C option" after outputting 0 or 1.
1043    
1044    33. There is now support for generating a code coverage report for the test
1045        suite in environments where gcc is the compiler and lcov is installed. This
1046        is mainly for the benefit of the developers.
1047    
1048    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1049        unaddressable using valgrind annotations, allowing valgrind to detect
1050        invalid memory accesses. This is mainly for the benefit of the developers.
1051    
1052    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1053    
1054    26. Give configure error if --enable-cpp but no C++ compiler found.
1055    
1056    
1057    Version 8.31 06-July-2012
1058    -------------------------
1059    
1060    1.  Fixing a wrong JIT test case and some compiler warnings.
1061    
1062    2.  Removed a bashism from the RunTest script.
1063    
1064    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1065        to unsigned type, result still unsigned" that was given by an MS compiler
1066        on encountering the code "-sizeof(xxx)".
1067    
1068    4.  Partial matching support is added to the JIT compiler.
1069    
1070    5.  Fixed several bugs concerned with partial matching of items that consist
1071        of more than one character:
1072    
1073        (a) /^(..)\1/ did not partially match "aba" because checking references was
1074            done on an "all or nothing" basis. This also applied to repeated
1075            references.
1076    
1077        (b) \R did not give a hard partial match if \r was found at the end of the
1078            subject.
1079    
1080        (c) \X did not give a hard partial match after matching one or more
1081            characters at the end of the subject.
1082    
1083        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1084            a partial match for the string "\r".
1085    
1086        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1087            a partial match for a CR character at the end of the subject string.
1088    
1089    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1090        running pcretest, the text "(JIT)" added to the output whenever JIT is
1091        actually used to run the match.
1092    
1093    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1094        or /S+[+] with a digit between 1 and 7.
1095    
1096    8.  OP_NOT now supports any UTF character not just single-byte ones.
1097    
1098    9.  (*MARK) control verb is now supported by the JIT compiler.
1099    
1100    10. The command "./RunTest list" lists the available tests without actually
1101        running any of them. (Because I keep forgetting what they all are.)
1102    
1103    11. Add PCRE_INFO_MAXLOOKBEHIND.
1104    
1105    12. Applied a (slightly modified) user-supplied patch that improves performance
1106        when the heap is used for recursion (compiled with --disable-stack-for-
1107        recursion). Instead of malloc and free for each heap frame each time a
1108        logical recursion happens, frames are retained on a chain and re-used where
1109        possible. This sometimes gives as much as 30% improvement.
1110    
1111    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1112        call.
1113    
1114    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1115    
1116    15. It is now possible to link pcretest with libedit as an alternative to
1117        libreadline.
1118    
1119    16. (*COMMIT) control verb is now supported by the JIT compiler.
1120    
1121    17. The Unicode data tables have been updated to Unicode 6.1.0.
1122    
1123    18. Added --file-list option to pcregrep.
1124    
1125    19. Added binary file support to pcregrep, including the -a, --binary-files,
1126        -I, and --text options.
1127    
1128    20. The madvise function is renamed for posix_madvise for QNX compatibility
1129        reasons. Fixed by Giuseppe D'Angelo.
1130    
1131    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1132    
1133    22. Changed the option for creating symbolic links for 16-bit man pages from
1134        -s to -sf so that re-installing does not cause issues.
1135    
1136    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1137    
1138    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1139        matches in certain environments (the workspace was not being correctly
1140        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1141        some of the workspace data at the beginning of a restart.
1142    
1143    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1144        was not doing so when it should - probably a typo introduced by SVN 528
1145        (change 8.10/14).
1146    
1147    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1148        \w+ when the character tables indicated that \x{c4} was a word character.
1149        There were several related cases, all because the tests for doing a table
1150        lookup were testing for characters less than 127 instead of 255.
1151    
1152    27. If a pattern contains capturing parentheses that are not used in a match,
1153        their slots in the ovector are set to -1. For those that are higher than
1154        any matched groups, this happens at the end of processing. In the case when
1155        there were back references that the ovector was too small to contain
1156        (causing temporary malloc'd memory to be used during matching), and the
1157        highest capturing number was not used, memory off the end of the ovector
1158        was incorrectly being set to -1. (It was using the size of the temporary
1159        memory instead of the true size.)
1160    
1161    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1162        ovector size, it uses memory at the end of the block that it has got.
1163    
1164    29. Check for an overlong MARK name and give an error at compile time. The
1165        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1166    
1167    30. JIT compiler update.
1168    
1169    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1170        Rill for the patch.
1171    
1172    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1173    
1174    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1175    
1176    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1177        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1178        was enabled.
1179    
1180    35. Improve JIT code generation for greedy plus quantifier.
1181    
1182    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1183        1 to "aa" instead of to an empty string. The bug affected repeated groups
1184        that could potentially match an empty string.
1185    
1186    37. Optimizing single character iterators in JIT.
1187    
1188    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1189        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1190        codepoints that are too big for the mode are faulted, and in a UTF mode,
1191        disallowed codepoints are also faulted.
1192    
1193    39. If PCRE was compiled with UTF support, in three places in the DFA
1194        matcher there was code that should only have been obeyed in UTF mode, but
1195        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1196        processing when bytes with values greater than 127 were present. In 16-bit
1197        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1198        both cases the values are those that cannot be the first data item in a UTF
1199        character. The three items that might have provoked this were recursions,
1200        possessively repeated groups, and atomic groups.
1201    
1202    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1203        and pcregrep, because some OS require shared objects to be explicitly
1204        passed to ld, causing the link step to fail if they are not.
1205    
1206    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1207        mode, patterns that started with \h* or \R* might be incorrectly matched.
1208    
1209    
1210    Version 8.30 04-February-2012
1211    -----------------------------
1212    
1213    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1214        name is defined in ctype.h.
1215    
1216    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1217        only in quite long subpatterns.
1218    
1219    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1220        since it was replaced by pcre_fullinfo() in February 2000.
1221    
1222    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1223        match a (*MARK), and the match failed at the start of the subject, a
1224        reference to memory before the start of the subject could occur. This bug
1225        was introduced by fix 17 of release 8.21.
1226    
1227    5.  A reference to an unset group with zero minimum repetition was giving
1228        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1229        /(another)?(\1?)test/ matched against "hello world test". This bug was
1230        introduced in release 8.13.
1231    
1232    6.  Add support for 16-bit character strings (a large amount of work involving
1233        many changes and refactorings).
1234    
1235    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1236        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1237        from a file.
1238    
1239    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1240        rounding is not applied in this particular case).
1241    
1242    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1243        if they appear, or are escaped, in patterns.
1244    
1245    10. Get rid of a number of -Wunused-but-set-variable warnings.
1246    
1247    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1248        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1249        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1250        also returns the mark "x". This bug applied to capturing parentheses,
1251        non-capturing parentheses, and atomic parentheses. It also applied to some
1252        assertions.
1253    
1254    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1255        information out of configure.ac instead of relying on pcre.h.generic, which
1256        is not stored in the repository.
1257    
1258    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1259        -lreadline.
1260    
1261    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1262    
1263    15. Applied Graycode's patch to put the top-level frame on the stack rather
1264        than the heap when not using the stack for recursion. This gives a
1265        performance improvement in many cases when recursion is not deep.
1266    
1267    16. Experimental code added to "pcretest -C" to output the stack frame size.
1268    
1269    
1270    Version 8.21 12-Dec-2011
1271    ------------------------
1272    
1273    1.  Updating the JIT compiler.
1274    
1275    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1276        are added as well.
1277    
1278    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1279        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1280        calling _pcre_jit_exec. Some extra comments are added.
1281    
1282    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1283        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1284        was introduced by change 18 for 8.20.
1285    
1286    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1287        ECMA-262 standard.
1288    
1289    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1290        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1291        This bug was probably introduced by change 9 of 8.13.
1292    
1293    7.  While fixing 6 above, I noticed that a number of other items were being
1294        incorrectly rejected as "not fixed length". This arose partly because newer
1295        opcodes had not been added to the fixed-length checking code. I have (a)
1296        corrected the bug and added tests for these items, and (b) arranged for an
1297        error to occur if an unknown opcode is encountered while checking for fixed
1298        length instead of just assuming "not fixed length". The items that were
1299        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1300        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1301        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1302    
1303    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1304        being incorrectly compiled and would have given unpredicatble results.
1305    
1306    9.  A possessively repeated subpattern with minimum repeat count greater than
1307        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1308        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1309        the first (A) could occur when it should not.
1310    
1311    10. Add a cast and remove a redundant test from the code.
1312    
1313    11. JIT should use pcre_malloc/pcre_free for allocation.
1314    
1315    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1316        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1317        is anything other than /usr, -L is still shown).
1318    
1319    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1320    
1321    14. Perl does not support \N without a following name in a [] class; PCRE now
1322        also gives an error.
1323    
1324    15. If a forward reference was repeated with an upper limit of around 2000,
1325        it caused the error "internal error: overran compiling workspace". The
1326        maximum number of forward references (including repeats) was limited by the
1327        internal workspace, and dependent on the LINK_SIZE. The code has been
1328        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1329        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1330        of around 200,000 forward references. While doing this, I also speeded up
1331        the filling in of repeated forward references.
1332    
1333    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1334        incorrectly expecting the subject to contain another "a" after the start.
1335    
1336    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1337        in the match, the SKIP should be ignored. This was not happening; instead
1338        the SKIP was being treated as NOMATCH. For patterns such as
1339        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1340        tested.
1341    
1342    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1343        now much more compatible with Perl, in particular in cases where the result
1344        is a non-match for a non-anchored pattern. For example, if
1345        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1346        "m", where previously it did not return a name. A side effect of this
1347        change is that for partial matches, the last encountered mark name is
1348        returned, as for non matches. A number of tests that were previously not
1349        Perl-compatible have been moved into the Perl-compatible test files. The
1350        refactoring has had the pleasing side effect of removing one argument from
1351        the match() function, thus reducing its stack requirements.
1352    
1353    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1354        subsequent uses of /S (without +) incorrectly behaved like /S+.
1355    
1356    21. Retrieve executable code size support for the JIT compiler and fixing
1357        some warnings.
1358    
1359    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1360        not work when the shorter character appeared right at the end of the
1361        subject string.
1362    
1363    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1364        systems.
1365    
1366    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1367        output it when the /M option is used in pcretest.
1368    
1369    25. The CheckMan script was not being included in the distribution. Also, added
1370        an explicit "perl" to run Perl scripts from the PrepareRelease script
1371        because this is reportedly needed in Windows.
1372    
1373    26. If study data was being save in a file and studying had not found a set of
1374        "starts with" bytes for the pattern, the data written to the file (though
1375        never used) was taken from uninitialized memory and so caused valgrind to
1376        complain.
1377    
1378    27. Updated RunTest.bat as provided by Sheri Pierce.
1379    
1380    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1381    
1382    29. Computation of memory usage for the table of capturing group names was
1383        giving an unnecessarily large value.
1384    
1385    
1386    Version 8.20 21-Oct-2011
1387    ------------------------
1388    
1389    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1390        a POSIX class. After further experiments with Perl, which convinced me that
1391        Perl has bugs and confusions, a closing square bracket is no longer allowed
1392        in a POSIX name. This bug also affected patterns with classes that started
1393        with full stops.
1394    
1395    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1396        captured substring, but while checking the failing first alternative,
1397        substring 1 is temporarily captured. If the output vector supplied to
1398        pcre_exec() was not big enough for this capture, the yield of the function
1399        was still zero ("insufficient space for captured substrings"). This cannot
1400        be totally fixed without adding another stack variable, which seems a lot
1401        of expense for a edge case. However, I have improved the situation in cases
1402        such as /(a)(b)x|abc/ matched against "abc", where the return code
1403        indicates that fewer than the maximum number of slots in the ovector have
1404        been set.
1405    
1406    3.  Related to (2) above: when there are more back references in a pattern than
1407        slots in the output vector, pcre_exec() uses temporary memory during
1408        matching, and copies in the captures as far as possible afterwards. It was
1409        using the entire output vector, but this conflicts with the specification
1410        that only 2/3 is used for passing back captured substrings. Now it uses
1411        only the first 2/3, for compatibility. This is, of course, another edge
1412        case.
1413    
1414    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1415        main code base, and can be used by building with --enable-jit. When this is
1416        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1417        runtime --no-jit option is given.
1418    
1419    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1420        ovector, the return from the function was zero, implying that there were
1421        other matches that did not fit. The correct "exactly full" value is now
1422        returned.
1423    
1424    6.  If a subpattern that was called recursively or as a subroutine contained
1425        (*PRUNE) or any other control that caused it to give a non-standard return,
1426        invalid errors such as "Error -26 (nested recursion at the same subject
1427        position)" or even infinite loops could occur.
1428    
1429    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1430        computing the minimum length on reaching *ACCEPT, and so ended up with the
1431        wrong value of 1 rather than 0. Further investigation indicates that
1432        computing a minimum subject length in the presence of *ACCEPT is difficult
1433        (think back references, subroutine calls), and so I have changed the code
1434        so that no minimum is registered for a pattern that contains *ACCEPT.
1435    
1436    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1437        it was not handled as intended. [But see 16 below.]
1438    
1439    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1440        Sheri Pierce.
1441    
1442    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1443        the first byte in a match must be "a".
1444    
1445    11. Change 17 for 8.13 increased the recursion depth for patterns like
1446        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1447        pattern contains any instances of (*THEN). If it does not, the old
1448        optimizations are restored. It would be nice to do this on a per-group
1449        basis, but at the moment that is not feasible.
1450    
1451    12. In some environments, the output of pcretest -C is CRLF terminated. This
1452        broke RunTest's code that checks for the link size. A single white space
1453        character after the value is now allowed for.
1454    
1455    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1456        For "fr", it uses the Windows-specific input and output files.
1457    
1458    14. If (*THEN) appeared in a group that was called recursively or as a
1459        subroutine, it did not work as intended. [But see next item.]
1460    
1461    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1462        pattern fragments (but not containing any | characters). If A and B are
1463        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1464        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1465        D. In other words, Perl considers parentheses that do not contain any |
1466        characters to be part of a surrounding alternative, whereas PCRE was
1467        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1468        differently. PCRE now behaves in the same way as Perl, except in the case
1469        of subroutine/recursion calls such as (?1) which have in any case always
1470        been different (but PCRE had them first :-).
1471    
1472    16. Related to 15 above: Perl does not treat the | in a conditional group as
1473        creating alternatives. Such a group is treated in the same way as an
1474        ordinary group without any | characters when processing (*THEN). PCRE has
1475        been changed to match Perl's behaviour.
1476    
1477    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1478        RunGrepTest script failed.
1479    
1480    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1481        inevitable for groups that contain captures, but it can lead to a lot of
1482        stack use in large patterns. The old behaviour has been restored for atomic
1483        groups that do not contain any capturing parentheses.
1484    
1485    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1486        suppress the check for a minimum subject length at run time. (If it was
1487        given to pcre_exec() or pcre_dfa_exec() it did work.)
1488    
1489    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1490        fail to work when decoding hex characters in data strings in EBCDIC
1491        environments.
1492    
1493    21. It appears that in at least one Mac OS environment, the isxdigit() function
1494        is implemented as a macro that evaluates to its argument more than once,
1495        contravening the C 90 Standard (I haven't checked a later standard). There
1496        was an instance in pcretest which caused it to go wrong when processing
1497        \x{...} escapes in subject strings. The has been rewritten to avoid using
1498        things like p++ in the argument of isxdigit().
1499    
1500    
1501    Version 8.13 16-Aug-2011
1502    ------------------------
1503    
1504    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1505    
1506    2.  Two minor typos in pcre_internal.h have been fixed.
1507    
1508    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1509        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1510        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1511    
1512    4.  There were a number of related bugs in the code for matching backrefences
1513        caselessly in UTF-8 mode when codes for the characters concerned were
1514        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1515        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1516        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1517        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1518        2-byte code at the end of the subject (it thought there wasn't enough data
1519        left).
1520    
1521    5.  Comprehensive information about what went wrong is now returned by
1522        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1523        as the output vector has at least 2 elements. The offset of the start of
1524        the failing character and a reason code are placed in the vector.
1525    
1526    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1527        now returned is for the first byte of the failing character, instead of the
1528        last byte inspected. This is an incompatible change, but I hope it is small
1529        enough not to be a problem. It makes the returned offset consistent with
1530        pcre_exec() and pcre_dfa_exec().
1531    
1532    7.  pcretest now gives a text phrase as well as the error number when
1533        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1534        failure, the offset and reason code are output.
1535    
1536    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1537        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1538        back over a single character (\n). This seems wrong (because it treated the
1539        two characters as a single entity when going forwards), conflicts with the
1540        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1541        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1542        has been changed.
1543    
1544    9.  Some internal refactoring has changed the processing so that the handling
1545        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1546        time (the PCRE_DOTALL option was changed this way some time ago: version
1547        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1548        which was always a bit of a fudge. It also means that there is one less
1549        argument for the match() function, which reduces its stack requirements
1550        slightly. This change also fixes an incompatibility with Perl: the pattern
1551        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1552    
1553    10. More internal refactoring has drastically reduced the number of recursive
1554        calls to match() for possessively repeated groups such as (abc)++ when
1555        using pcre_exec().
1556    
1557    11. While implementing 10, a number of bugs in the handling of groups were
1558        discovered and fixed:
1559    
1560        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1561        (a|)*(?1) gave a compile-time internal error.
1562        ((a|)+)+  did not notice that the outer group could match an empty string.
1563        (^a|^)+   was not marked as anchored.
1564        (.*a|.*)+ was not marked as matching at start or after a newline.
1565    
1566    12. Yet more internal refactoring has removed another argument from the match()
1567        function. Special calls to this function are now indicated by setting a
1568        value in a variable in the "match data" data block.
1569    
1570    13. Be more explicit in pcre_study() instead of relying on "default" for
1571        opcodes that mean there is no starting character; this means that when new
1572        ones are added and accidentally left out of pcre_study(), testing should
1573        pick them up.
1574    
1575    14. The -s option of pcretest has been documented for ages as being an old
1576        synonym of -m (show memory usage). I have changed it to mean "force study
1577        for every regex", that is, assume /S for every regex. This is similar to -i
1578        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1579        using it. It makes it easier to run collections of tests with and without
1580        study enabled, and thereby test pcre_study() more easily. All the standard
1581        tests are now run with and without -s (but some patterns can be marked as
1582        "never study" - see 20 below).
1583    
1584    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1585        restoration of the capturing data to the outer values was not happening
1586        correctly.
1587    
1588    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1589        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1590        pattern had matched an empty string, and so incorrectly returned a no
1591        match.
1592    
1593    17. There was optimizing code for the last branch of non-capturing parentheses,
1594        and also for the obeyed branch of a conditional subexpression, which used
1595        tail recursion to cut down on stack usage. Unfortunately, now that there is
1596        the possibility of (*THEN) occurring in these branches, tail recursion is
1597        no longer possible because the return has to be checked for (*THEN). These
1598        two optimizations have therefore been removed. [But see 8.20/11 above.]
1599    
1600    18. If a pattern containing \R was studied, it was assumed that \R always
1601        matched two bytes, thus causing the minimum subject length to be
1602        incorrectly computed because \R can also match just one byte.
1603    
1604    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1605        was incorrectly computed.
1606    
1607    20. If /S is present twice on a test pattern in pcretest input, it now
1608        *disables* studying, thereby overriding the use of -s on the command line
1609        (see 14 above). This is necessary for one or two tests to keep the output
1610        identical in both cases.
1611    
1612    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1613        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1614    
1615    22. When an atomic group that contained a capturing parenthesis was
1616        successfully matched, but the branch in which it appeared failed, the
1617        capturing was not being forgotten if a higher numbered group was later
1618        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1619        group 1 to "a", when in fact it should be unset. This applied to multi-
1620        branched capturing and non-capturing groups, repeated or not, and also to
1621        positive assertions (capturing in negative assertions does not happen
1622        in PCRE) and also to nested atomic groups.
1623    
1624    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1625        subject after a captured substring, to make it easier to tell which of a
1626        number of identical substrings has been captured.
1627    
1628    24. The way atomic groups are processed by pcre_exec() has been changed so that
1629        if they are repeated, backtracking one repetition now resets captured
1630        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1631        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1632        "aaa". Previously, it would have been "a". As part of this code
1633        refactoring, the way recursive calls are handled has also been changed.
1634    
1635    25. If an assertion condition captured any substrings, they were not passed
1636        back unless some other capturing happened later. For example, if
1637        (?(?=(a))a) was matched against "a", no capturing was returned.
1638    
1639    26. When studying a pattern that contained subroutine calls or assertions,
1640        the code for finding the minimum length of a possible match was handling
1641        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1642        group 1 called group 2 while simultaneously a separate group 2 called group
1643        1). A stack overflow occurred in this case. I have fixed this by limiting
1644        the recursion depth to 10.
1645    
1646    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1647        Fortmann. This supports explicit test numbers on the command line, and has
1648        argument validation and error reporting.
1649    
1650    28. An instance of \X with an unlimited repeat could fail if at any point the
1651        first character it looked at was a mark character.
1652    
1653    29. Some minor code refactoring concerning Unicode properties and scripts
1654        should reduce the stack requirement of match() slightly.
1655    
1656    30. Added the '=' option to pcretest to check the setting of unused capturing
1657        slots at the end of the pattern, which are documented as being -1, but are
1658        not included in the return count.
1659    
1660    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1661        compiled something random. Now it gives a compile-time error (as does
1662        Perl).
1663    
1664    32. A *MARK encountered during the processing of a positive assertion is now
1665        recorded and passed back (compatible with Perl).
1666    
1667    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1668        had alternative anchored branches, the search for a second match in a line
1669        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1670        matched the line "0102" twice. The same bug affected patterns that started
1671        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1672        twice.
1673    
1674    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1675        does, and because of capturing effects, quantifying parenthesized
1676        assertions may at times be useful. Quantifiers are now allowed for
1677        parenthesized assertions.
1678    
1679    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1680    
1681    36. \g was being checked for fancy things in a character class, when it should
1682        just be a literal "g".
1683    
1684    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1685        appearance of a nested POSIX class supersedes an apparent external class.
1686        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1687        unescaped square brackets may also appear as part of class names. For
1688        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1689        more like Perl. (But see 8.20/1 above.)
1690    
1691    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1692        was because it thought it was \N{name}, which is not supported).
1693    
1694    39. Add minix to OS list not supporting the -S option in pcretest.
1695    
1696    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1697        cannot analyze patterns in sufficient detail to catch mutual recursions
1698        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1699        subgroup is called recursively as a subpattern for a second time at the
1700        same position in the subject string. In previous releases this might have
1701        been caught by the recursion limit, or it might have run out of stack.
1702    
1703    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1704        happen only once. PCRE was, however incorrectly giving a compile time error
1705        "recursive call could loop indefinitely" because it cannot analyze the
1706        pattern in sufficient detail. The compile time test no longer happens when
1707        PCRE is compiling a conditional subpattern, but actual runaway loops are
1708        now caught at runtime (see 40 above).
1709    
1710    42. It seems that Perl allows any characters other than a closing parenthesis
1711        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1712        has been changed to be the same.
1713    
1714    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1715        as not to get warnings when autogen.sh is called. Also changed
1716        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1717    
1718    44. To help people who use pcregrep to scan files containing exceedingly long
1719        lines, the following changes have been made:
1720    
1721        (a) The default value of the buffer size parameter has been increased from
1722            8K to 20K. (The actual buffer used is three times this size.)
1723    
1724        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1725            PCRE is built.
1726    
1727        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1728            to be set at run time.
1729    
1730        (d) Numerical values in pcregrep options can be followed by K or M, for
1731            example --buffer-size=50K.
1732    
1733        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1734            given and the return code is set to 2.
1735    
1736    45. Add a pointer to the latest mark to the callout data block.
1737    
1738    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1739        partial match of an empty string instead of no match. This was specific to
1740        the use of ".".
1741    
1742    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1743        complete match instead of a partial match. This bug was dependent on both
1744        the PCRE_UTF8 and PCRE_DOTALL options being set.
1745    
1746    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1747        starting byte set, because \b was not being ignored.
1748    
1749    
1750    Version 8.12 15-Jan-2011
1751    ------------------------
1752    
1753    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1754        checks for such things as part of the documentation building process.
1755    
1756    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1757        --match-limit and --recursion-limit options (added for 8.11). In
1758        particular, this made one of the standard tests fail. (The integer value
1759        went into the wrong half of a long int.)
1760    
1761    3.  If the --colour option was given to pcregrep with -v (invert match), it
1762        did strange things, either producing crazy output, or crashing. It should,
1763        of course, ignore a request for colour when reporting lines that do not
1764        match.
1765    
1766    4.  Another pcregrep bug caused similar problems if --colour was specified with
1767        -M (multiline) and the pattern match finished with a line ending.
1768    
1769    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1770        matched in multiline mode, the following line was shown as part of the
1771        match. This seems wrong, so I have changed it.
1772    
1773    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1774        the check for further matches in the same line (so they could be coloured)
1775        to overrun the end of the current line. If another match was found, it was
1776        incorrectly shown (and then shown again when found in the next line).
1777    
1778    7.  If pcregrep was compiled under Windows, there was a reference to the
1779        function pcregrep_exit() before it was defined. I am assuming this was
1780        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1781        reported by a user. I've moved the definition above the reference.
1782    
1783    
1784    Version 8.11 10-Dec-2010
1785    ------------------------
1786    
1787    1.  (*THEN) was not working properly if there were untried alternatives prior
1788        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1789        backtracked to try for "b" instead of moving to the next alternative branch
1790        at the same level (in this case, to look for "c"). The Perl documentation
1791        is clear that when (*THEN) is backtracked onto, it goes to the "next
1792        alternative in the innermost enclosing group".
1793    
1794    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1795        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1796        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1797        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1798        (*THEN).
1799    
1800    3.  If \s appeared in a character class, it removed the VT character from
1801        the class, even if it had been included by some previous item, for example
1802        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1803        of \s, but is part of the POSIX "space" class.)
1804    
1805    4.  A partial match never returns an empty string (because you can always
1806        match an empty string at the end of the subject); however the checking for
1807        an empty string was starting at the "start of match" point. This has been
1808        changed to the "earliest inspected character" point, because the returned
1809        data for a partial match starts at this character. This means that, for
1810        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1811        (previously it gave "no match").
1812    
1813    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1814        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1815        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1816        has an implication that the given string is incomplete (because a partial
1817        match is preferred over a full match). For this reason, these items now
1818        give a partial match in this situation. [Aside: previously, the one case
1819        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1820        match rather than a full match, which was wrong by the old rules, but is
1821        now correct.]
1822    
1823    6.  There was a bug in the handling of #-introduced comments, recognized when
1824        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1825        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1826        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1827        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1828        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1829        places in pcre_compile().
1830    
1831    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1832        comments when looking ahead for named forward references to subpatterns,
1833        the only newline sequence it recognized was NL. It now handles newlines
1834        according to the set newline convention.
1835    
1836    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1837        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1838        cater for a lack of strerror(). These oversights have been fixed.
1839    
1840    9.  Added --match-limit and --recursion-limit to pcregrep.
1841    
1842    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1843    
1844    11. When the -o option was used, pcregrep was setting a return code of 1, even
1845        when matches were found, and --line-buffered was not being honoured.
1846    
1847    12. Added an optional parentheses number to the -o and --only-matching options
1848        of pcregrep.
1849    
1850    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1851        can match an empty string. The code to do it in pcretest and pcredemo
1852        needed fixing:
1853    
1854        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1855            only one byte after an empty string match just before CRLF (this case
1856            just got forgotten; "any" and "anycrlf" were OK).
1857    
1858        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1859            mode when an empty string match preceded an ASCII character followed by
1860            a non-ASCII character. (The code for advancing by one character rather
1861            than one byte was nonsense.)
1862    
1863        (c) The pcredemo.c sample program did not have any code at all to handle
1864            the cases when CRLF is a valid newline sequence.
1865    
1866    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1867        as a starting offset was within the subject string. There is now a new
1868        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1869        negative or greater than the length of the string. In order to test this,
1870        pcretest is extended to allow the setting of negative starting offsets.
1871    
1872    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1873        starting offset points to the beginning of a UTF-8 character was
1874        unnecessarily clumsy. I tidied it up.
1875    
1876    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1877        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1878    
1879    17. Nobody had reported that the --include_dir option, which was added in
1880        release 7.7 should have been called --include-dir (hyphen, not underscore)
1881        for compatibility with GNU grep. I have changed it to --include-dir, but
1882        left --include_dir as an undocumented synonym, and the same for
1883        --exclude-dir, though that is not available in GNU grep, at least as of
1884        release 2.5.4.
1885    
1886    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1887        characters from a string of bytes) have been redefined so as not to use
1888        loops, in order to improve performance in some environments. At the same
1889        time, I abstracted some of the common code into auxiliary macros to save
1890        repetition (this should not affect the compiled code).
1891    
1892    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1893        compile-time error is now given if \c is not followed by an ASCII
1894        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1895        different, and any byte value is allowed.)
1896    
1897    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1898        START_OPTIMIZE option, which is now allowed at compile time - but just
1899        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1900        to pcregrep and other applications that have no direct access to PCRE
1901        options. The new /Y option in pcretest sets this option when calling
1902        pcre_compile().
1903    
1904    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1905        back references. Groups containing recursive back references were forced to
1906        be atomic by that change, but in the case of named groups, the amount of
1907        memory required was incorrectly computed, leading to "Failed: internal
1908        error: code overflow". This has been fixed.
1909    
1910    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1911        pcretest.c, to avoid build problems in some Borland environments.
1912    
1913    
1914    Version 8.10 25-Jun-2010
1915    ------------------------
1916    
1917    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1918        THEN.
1919    
1920    2.  (*ACCEPT) was not working when inside an atomic group.
1921    
1922    3.  Inside a character class, \B is treated as a literal by default, but
1923        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1924        causes the error). The code is unchanged, but I tidied the documentation.
1925    
1926    4.  Inside a character class, PCRE always treated \R and \X as literals,
1927        whereas Perl faults them if its -w option is set. I have changed PCRE so
1928        that it faults them when PCRE_EXTRA is set.
1929    
1930    5.  Added support for \N, which always matches any character other than
1931        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1932    
1933    6.  When compiling pcregrep with newer versions of gcc which may have
1934        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1935        declared with attribute warn_unused_result" were given. Just casting the
1936        result to (void) does not stop the warnings; a more elaborate fudge is
1937        needed. I've used a macro to implement this.
1938    
1939    7.  Minor change to pcretest.c to avoid a compiler warning.
1940    
1941    8.  Added four artifical Unicode properties to help with an option to make
1942        \s etc use properties (see next item). The new properties are: Xan
1943        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1944    
1945    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1946        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1947        this option. Modified pcretest to add /W to test this facility. Added
1948        REG_UCP to make it available via the POSIX interface.
1949    
1950    10. Added --line-buffered to pcregrep.
1951    
1952    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1953        studied, and the match started with a letter with a code point greater than
1954        127 whose first byte was different to the first byte of the other case of
1955        the letter, the other case of this starting letter was not recognized
1956        (#976).
1957    
1958    12. If a pattern that was studied started with a repeated Unicode property
1959        test, for example, \p{Nd}+, there was the theoretical possibility of
1960        setting up an incorrect bitmap of starting bytes, but fortunately it could
1961        not have actually happened in practice until change 8 above was made (it
1962        added property types that matched character-matching opcodes).
1963    
1964    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1965        possible starting bytes for non-anchored patterns.
1966    
1967    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1968        \R, and also a number of cases that involve Unicode properties, both
1969        explicit and implicit when PCRE_UCP is set.
1970    
1971    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1972        input, it could crash or give wrong results if characters with values
1973        greater than 0xc0 were present in the subject string. (Detail: it assumed
1974        UTF-8 input when processing these items.)
1975    
1976    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1977        size_t is 64-bit (#991).
1978    
1979    17. Added a check for running out of memory when PCRE is compiled with
1980        --disable-stack-for-recursion (#990).
1981    
1982    18. If the last data line in a file for pcretest does not have a newline on
1983        the end, a newline was missing in the output.
1984    
1985    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1986        less than 128) in its various bitmaps. However, there is a facility for
1987        generating tables according to the current locale when PCRE is compiled. It
1988        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1989        characters, are recognized by isspace() and therefore were getting set in
1990        these tables, and indeed these tables seem to approximate to ISO 8859. This
1991        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1992        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1993        which of course cannot start UTF-8 characters. I have changed the code so
1994        that only real ASCII characters (less than 128) and the correct starting
1995        bytes for UTF-8 encodings are set for characters greater than 127 when in
1996        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1997        altogether.)
1998    
1999    20. Added the /T option to pcretest so as to be able to run tests with non-
2000        standard character tables, thus making it possible to include the tests
2001        used for 19 above in the standard set of tests.
2002    
2003    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2004        reference to a subpattern the other side of a comment that contains an
2005        opening parenthesis caused either an internal compiling error, or a
2006        reference to the wrong subpattern.
2007    
2008    
2009    Version 8.02 19-Mar-2010
2010    ------------------------
2011    
2012    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2013    
2014    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2015        configured.
2016    
2017    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2018        original author of that file, following a query about its status.
2019    
2020    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2021        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2022    
2023    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2024        quantifier applied to a forward-referencing subroutine call, could compile
2025        incorrect code or give the error "internal error: previously-checked
2026        referenced subpattern not found".
2027    
2028    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2029        variables to point to external functions. For these systems, therefore,
2030        pcre_malloc etc. are now initialized to local functions that call the
2031        relevant global functions.
2032    
2033    7.  There were two entries missing in the vectors called coptable and poptable
2034        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2035        I've fixed the data, and added a kludgy way of testing at compile time that
2036        the lengths are correct (equal to the number of opcodes).
2037    
2038    8.  Following on from 7, I added a similar kludge to check the length of the
2039        eint vector in pcreposix.c.
2040    
2041    9.  Error texts for pcre_compile() are held as one long string to avoid too
2042        much relocation at load time. To find a text, the string is searched,
2043        counting zeros. There was no check for running off the end of the string,
2044        which could happen if a new error number was added without updating the
2045        string.
2046    
2047    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2048    
2049    11. \K was not working if it appeared in an atomic group or in a group that
2050        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2051        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2052        the assertion is positive, but not if it is negative.
2053    
2054    12. Change 11 fortuitously reduced the size of the stack frame used in the
2055        "match()" function of pcre_exec.c by one pointer. Forthcoming
2056        implementation of support for (*MARK) will need an extra pointer on the
2057        stack; I have reserved it now, so that the stack frame size does not
2058        decrease.
2059    
2060    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2061        item in branch that calls a recursion is a subroutine call - as in the
2062        second branch in the above example - was incorrectly given the compile-
2063        time error "recursive call could loop indefinitely" because pcre_compile()
2064        was not correctly checking the subroutine for matching a non-empty string.
2065    
2066    14. The checks for overrunning compiling workspace could trigger after an
2067        overrun had occurred. This is a "should never occur" error, but it can be
2068        triggered by pathological patterns such as hundreds of nested parentheses.
2069        The checks now trigger 100 bytes before the end of the workspace.
2070    
2071    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2072    
2073    
2074    Version 8.01 19-Jan-2010
2075    ------------------------
2076    
2077    1.  If a pattern contained a conditional subpattern with only one branch (in
2078        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2079        computed the wrong minimum data length (which is of course zero for such
2080        subpatterns). This could cause incorrect "no match" results.
2081    
2082    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2083        the pattern is reset in the first branch, pcre_compile() failed with
2084        "internal error: code overflow at offset...". This happened only when
2085        the reset was to the original external option setting. (An optimization
2086        abstracts leading options settings into an external setting, which was the
2087        cause of this.)
2088    
2089    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2090        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2091        assertion pattern did not match (meaning that the assertion was true), it
2092        was incorrectly treated as false if the SKIP had been reached during the
2093        matching. This also applied to assertions used as conditions.
2094    
2095    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2096        assertion subpattern, including such a pattern used as a condition,
2097        unpredictable results occurred, instead of the error return
2098        PCRE_ERROR_DFA_UITEM.
2099    
2100    5.  The C++ GlobalReplace function was not working like Perl for the special
2101        situation when an empty string is matched. It now does the fancy magic
2102        stuff that is necessary.
2103    
2104    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2105        removed. (These were left over from very, very early versions of PCRE.)
2106    
2107    7.  Some cosmetic changes to the code to make life easier when compiling it
2108        as part of something else:
2109    
2110        (a) Change DEBUG to PCRE_DEBUG.
2111    
2112        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2113            called "current" as "current_branch", to prevent a collision with the
2114            Linux macro when compiled as a kernel module.
2115    
2116        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2117            prevent a collision with the Linux macro when compiled as a kernel
2118            module.
2119    
2120    8.  In pcre_compile() there are some checks for integer overflows that used to
2121        cast potentially large values to (double). This has been changed to that
2122        when building, a check for int64_t is made, and if it is found, it is used
2123        instead, thus avoiding the use of floating point arithmetic. (There is no
2124        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2125        double.
2126    
2127    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2128        2005 (difference between two addresses compared to an unsigned value).
2129    
2130    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2131        custom one, because of the following reported problem in Windows:
2132    
2133          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2134              under Win32.
2135          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2136              therefore missing the function definition.
2137          - The compiler thus generates a "C" signature for the test function.
2138          - The linker fails to find the "C" function.
2139          - PCRE fails to configure if asked to do so against libbz2.
2140    
2141    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2142        messages were output:
2143    
2144          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2145          rerunning libtoolize, to keep the correct libtool macros in-tree.
2146          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2147    
2148        I have done both of these things.
2149    
2150    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2151        most of the time, it *can* run out if it is given a pattern that contains a
2152        runaway infinite recursion. I updated the discussion in the pcrestack man
2153        page.
2154    
2155    13. Now that we have gone to the x.xx style of version numbers, the minor
2156        version may start with zero. Using 08 or 09 is a bad idea because users
2157        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2158        interpreted as invalid octal numbers. I've updated the previous comment in
2159        configure.ac, and also added a check that gives an error if 08 or 09 are
2160        used.
2161    
2162    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2163        causing partial matching to fail when the end of the subject matched \W
2164        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2165    
2166    15. There were some discrepancies between the declarations in pcre_internal.h
2167        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2168        their definitions. The declarations used "const uschar *" and the
2169        definitions used USPTR. Even though USPTR is normally defined as "const
2170        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2171        reported that: "This difference in casting confuses some C++ compilers, for
2172        example, SunCC recognizes above declarations as different functions and
2173        generates broken code for hbpcre." I have changed the declarations to use
2174        USPTR.
2175    
2176    16. GNU libtool is named differently on some systems. The autogen.sh script now
2177        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2178        (FreeBSD).
2179    
2180    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2181        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2182        comment: "Figure out how to create a longlong from a string: strtoll and
2183        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2184        instance, but it only takes 2 args instead of 3!"
2185    
2186    18. A subtle bug concerned with back references has been fixed by a change of
2187        specification, with a corresponding code fix. A pattern such as
2188        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2189        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2190        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2191        same bug. Such groups have to be quantified to be useful, or contained
2192        inside another quantified group. (If there's no repetition, the reference
2193        can never match.) The problem arises because, having left the group and
2194        moved on to the rest of the pattern, a later failure that backtracks into
2195        the group uses the captured value from the final iteration of the group
2196        rather than the correct earlier one. I have fixed this in PCRE by forcing
2197        any group that contains a reference to itself to be an atomic group; that
2198        is, there cannot be any backtracking into it once it has completed. This is
2199        similar to recursive and subroutine calls.
2200    
2201    
2202    Version 8.00 19-Oct-09
2203    ----------------------
2204    
2205    1.  The table for translating pcre_compile() error codes into POSIX error codes
2206        was out-of-date, and there was no check on the pcre_compile() error code
2207        being within the table. This could lead to an OK return being given in
2208        error.
2209    
2210    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2211        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2212        in a Windows environment.
2213    
2214    3.  The pcregrep --count option prints the count for each file even when it is
2215        zero, as does GNU grep. However, pcregrep was also printing all files when
2216        --files-with-matches was added. Now, when both options are given, it prints
2217        counts only for those files that have at least one match. (GNU grep just
2218        prints the file name in this circumstance, but including the count seems
2219        more useful - otherwise, why use --count?) Also ensured that the
2220        combination -clh just lists non-zero counts, with no names.
2221    
2222    4.  The long form of the pcregrep -F option was incorrectly implemented as
2223        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2224        but it seems right to fix it, and I didn't think it was worth preserving
2225        the old behaviour.
2226    
2227    5.  The command line items --regex=pattern and --regexp=pattern were not
2228        recognized by pcregrep, which required --regex pattern or --regexp pattern
2229        (with a space rather than an '='). The man page documented the '=' forms,
2230        which are compatible with GNU grep; these now work.
2231    
2232    6.  No libpcreposix.pc file was created for pkg-config; there was just
2233        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2234    
2235    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2236        when UCP support is not needed, by modifying the Python script that
2237        generates it from Unicode data files. This should not matter if the module
2238        is correctly used as a library, but I received one complaint about 50K of
2239        unwanted data. My guess is that the person linked everything into his
2240        program rather than using a library. Anyway, it does no harm.
2241    
2242    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2243        was a minimum greater than 1 for a wide character in a possessive
2244        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2245        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2246        character. Chaos in the form of incorrect output or a compiling loop could
2247        result.
2248    
2249    9.  The restrictions on what a pattern can contain when partial matching is
2250        requested for pcre_exec() have been removed. All patterns can now be
2251        partially matched by this function. In addition, if there are at least two
2252        slots in the offset vector, the offset of the earliest inspected character
2253        for the match and the offset of the end of the subject are set in them when
2254        PCRE_ERROR_PARTIAL is returned.
2255    
2256    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2257        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2258        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2259        and may be more useful for multi-segment matching.
2260    
2261    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2262        used to be given if ever the end of the subject was reached; now it is
2263        given only if matching could not proceed because another character was
2264        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2265        string "Z", which now yields "no match" instead of "partial match". In the
2266        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2267        final character ended with (*FAIL).
2268    
2269    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2270        if the pattern had a "must contain" character that was already found in the
2271        earlier partial match, unless partial matching was again requested. For
2272        example, with the pattern /dog.(body)?/, the "must contain" character is
2273        "g". If the first part-match was for the string "dog", restarting with
2274        "sbody" failed. This bug has been fixed.
2275    
2276    13. The string returned by pcre_dfa_exec() after a partial match has been
2277        changed so that it starts at the first inspected character rather than the
2278        first character of the match. This makes a difference only if the pattern
2279        starts with a lookbehind assertion or \b or \B (\K is not supported by
2280        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2281        matching functions compatible, and I think it's the right thing to do.
2282    
2283    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2284        so that the demonstration program is easily available in environments where
2285        PCRE has not been installed from source.
2286    
2287    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2288        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2289        library.
2290    
2291    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2292        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2293        is not the first non-POSIX option to be added. Clearly some people find
2294        these options useful.
2295    
2296    17. If a caller to the POSIX matching function regexec() passes a non-zero
2297        value for nmatch with a NULL value for pmatch, the value of
2298        nmatch is forced to zero.
2299    
2300    18. RunGrepTest did not have a test for the availability of the -u option of
2301        the diff command, as RunTest does. It now checks in the same way as
2302        RunTest, and also checks for the -b option.
2303    
2304    19. If an odd number of negated classes containing just a single character
2305        interposed, within parentheses, between a forward reference to a named
2306        subpattern and the definition of the subpattern, compilation crashed with
2307        an internal error, complaining that it could not find the referenced
2308        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2309        [The bug was that it was starting one character too far in when skipping
2310        over the character class, thus treating the ] as data rather than
2311        terminating the class. This meant it could skip too much.]
2312    
2313    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2314        /g option in pcretest when the pattern contains \K, which makes it possible
2315        to have an empty string match not at the start, even when the pattern is
2316        anchored. Updated pcretest and pcredemo to use this option.
2317    
2318    21. If the maximum number of capturing subpatterns in a recursion was greater
2319        than the maximum at the outer level, the higher number was returned, but
2320        with unset values at the outer level. The correct (outer level) value is
2321        now given.
2322    
2323    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2324        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2325        make it do so. The string so far is captured, making this feature
2326        compatible with Perl.
2327    
2328    23. The tests have been re-organized, adding tests 11 and 12, to make it
2329        possible to check the Perl 5.10 features against Perl 5.10.
2330    
2331    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2332        pattern matches a fixed length string. PCRE did not allow this; now it
2333        does. Neither allows recursion.
2334    
2335    25. I finally figured out how to implement a request to provide the minimum
2336        length of subject string that was needed in order to match a given pattern.
2337        (It was back references and recursion that I had previously got hung up
2338        on.) This code has now been added to pcre_study(); it finds a lower bound
2339        to the length of subject needed. It is not necessarily the greatest lower
2340        bound, but using it to avoid searching strings that are too short does give
2341        some useful speed-ups. The value is available to calling programs via
2342        pcre_fullinfo().
2343    
2344    26. While implementing 25, I discovered to my embarrassment that pcretest had
2345        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2346        study optimizations had never been tested with that matching function.
2347        Oops. What is worse, even when it was passed study data, there was a bug in
2348        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2349        were also very few tests of studied patterns with pcre_dfa_exec().
2350    
2351    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2352        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2353        on the other side of the coin, they are no longer allowed to have different
2354        names, because these cannot be distinguished in PCRE, and this has caused
2355        confusion. (This is a difference from Perl.)
2356    
2357    28. When duplicate subpattern names are present (necessarily with different
2358        numbers, as required by 27 above), and a test is made by name in a
2359        conditional pattern, either for a subpattern having been matched, or for
2360        recursion in such a pattern, all the associated numbered subpatterns are
2361        tested, and the overall condition is true if the condition is true for any
2362        one of them. This is the way Perl works, and is also more like the way
2363        testing by number works.
2364    
2365    
2366    Version 7.9 11-Apr-09
2367    ---------------------
2368    
2369    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2370        (pcretest), all targets were linked against these libraries. This included
2371        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2372        libraries. This caused unwanted dependencies to be created. This problem
2373        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2374        pcretest is linked with readline.
2375    
2376    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2377        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2378        moved outside it again, because FALSE and TRUE are already defined in AIX,
2379        but BOOL is not.
2380    
2381    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2382        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2383    
2384    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2385        hyphens) following file names and line numbers when outputting matching
2386        lines. This is not true; no spaces are inserted. I have also clarified the
2387        wording for the --colour (or --color) option.
2388    
2389    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2390        was not coloured; this is different to GNU grep, so I have changed it to be
2391        the same.
2392    
2393    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2394        each matching line was coloured. Now it goes on to look for further matches
2395        of any of the test patterns, which is the same behaviour as GNU grep.
2396    
2397    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2398        doesn't make sense to accept an empty string match in pcregrep, so I have
2399        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2400        seems to be how GNU grep behaves. [But see later change 40 for release
2401        8.33.]
2402    
2403    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2404        start or after a newline", because the conditional assertion was not being
2405        correctly handled. The rule now is that both the assertion and what follows
2406        in the first alternative must satisfy the test.
2407    
2408    9.  If auto-callout was enabled in a pattern with a conditional group whose
2409        condition was an assertion, PCRE could crash during matching, both with
2410        pcre_exec() and pcre_dfa_exec().
2411    
2412    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2413        used for matching.
2414    
2415    11. Unicode property support in character classes was not working for
2416        characters (bytes) greater than 127 when not in UTF-8 mode.
2417    
2418    12. Added the -M command line option to pcretest.
2419    
2420    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2421    
2422    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2423    
2424    16. Added comments and documentation about mis-use of no_arg in the C++
2425        wrapper.
2426    
2427    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2428        from Martin Jerabek that uses macro names for all relevant character and
2429        string constants.
2430    
2431    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2432        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2433        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2434        these, but not everybody uses configure.
2435    
2436    19. A conditional group that had only one branch was not being correctly
2437        recognized as an item that could match an empty string. This meant that an
2438        enclosing group might also not be so recognized, causing infinite looping
2439        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2440        with the subject "ab", where knowledge that the repeated group can match
2441        nothing is needed in order to break the loop.
2442    
2443    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2444        exec(), but without supplying a callout function, matching went wrong.
2445    
2446    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2447        leak if the size of the offset vector was greater than 30. When the vector
2448        is smaller, the saved offsets during recursion go onto a local stack
2449        vector, but for larger vectors malloc() is used. It was failing to free
2450        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2451        error, in fact).
2452    
2453    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2454        heapframe that is used only when UTF-8 support is enabled. This caused no
2455        problem, but was untidy.
2456    
2457    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2458        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2459        included within another project.
2460    
2461    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2462        slightly modified by me:
2463    
2464          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2465              not building pcregrep.
2466    
2467          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2468              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2469    
2470    25. Forward references, both numeric and by name, in patterns that made use of
2471        duplicate group numbers, could behave incorrectly or give incorrect errors,
2472        because when scanning forward to find the reference group, PCRE was not
2473        taking into account the duplicate group numbers. A pattern such as
2474        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2475    
2476    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2477        the feature of a custom pointer more persuasive (as requested by a user).
2478    
2479    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2480        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2481        pre-defined.
2482    
2483    28. Added support for (*UTF8) at the start of a pattern.
2484    
2485    29. Arrange for flags added by the "release type" setting in CMake to be shown
2486        in the configuration summary.
2487    
2488    
2489    Version 7.8 05-Sep-08
2490    ---------------------
2491    
2492    1.  Replaced UCP searching code with optimized version as implemented for Ad
2493        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2494        stage table and inline lookup instead of a function, giving speed ups of 2
2495        to 5 times on some simple patterns that I tested. Permission was given to
2496        distribute the MultiStage2.py script that generates the tables (it's not in
2497        the tarball, but is in the Subversion repository).
2498    
2499    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2500        scripts.
2501    
2502    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2503        a group with a zero qualifier. The result of the study could be incorrect,
2504        or the function might crash, depending on the pattern.
2505    
2506    4.  Caseless matching was not working for non-ASCII characters in back
2507        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2508        It now works when Unicode Property Support is available.
2509    
2510    5.  In pcretest, an escape such as \x{de} in the data was always generating
2511        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2512        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2513        truncation.
2514    
2515    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2516    
2517    7.  Added two (int) casts to pcregrep when printing the difference of two
2518        pointers, in case they are 64-bit values.
2519    
2520    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2521        test 2 if it fails.
2522    
2523    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2524        and a #define of that name to empty if it is not externally set. This is to
2525        allow users of MSVC to set it if necessary.
2526    
2527    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2528        the convenience functions in the pcre_get.c source file.
2529    
2530    11. An option change at the start of a pattern that had top-level alternatives
2531        could cause overwriting and/or a crash. This command provoked a crash in
2532        some environments:
2533    
2534          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2535    
2536        This potential security problem was recorded as CVE-2008-2371.
2537    
2538    12. For a pattern where the match had to start at the beginning or immediately
2539        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2540        pcre_dfa_exec() could read past the end of the passed subject if there was
2541        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2542        pcretest so that it places the subject at the end of its malloc-ed buffer.
2543    
2544    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2545        exec() might read past the end of the data buffer in UTF-8 mode.
2546    
2547    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2548        the data contained the byte 0x85 as part of a UTF-8 character within its
2549        first line. This applied both to normal and DFA matching.
2550    
2551    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2552        /^[^d]*?$/8 failed to match "abc".
2553    
2554    16. Added a missing copyright notice to pcrecpp_internal.h.
2555    
2556    17. Make it more clear in the documentation that values returned from
2557        pcre_exec() in ovector are byte offsets, not character counts.
2558    
2559    18. Tidied a few places to stop certain compilers from issuing warnings.
2560    
2561    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2562        supplied by Stefan Weber. I made a further small update for 7.8 because
2563        there is a change of source arrangements: the pcre_searchfuncs.c module is
2564        replaced by pcre_ucd.c.
2565    
2566    
2567    Version 7.7 07-May-08
2568    ---------------------
2569    
2570    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2571        a string to a long long, pretend we don't even have a long long." This is
2572        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2573    
2574    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2575        pre-7.6 versions, which defined a global no_arg variable instead of putting
2576        it in the RE class. (See also #8 below.)
2577    
2578    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2579        Lopes.
2580    
2581    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2582    
2583        (1) The include/exclude patterns were being applied to the whole pathnames
2584            of files, instead of just to the final components.
2585    
2586        (2) If there was more than one level of directory, the subdirectories were
2587            skipped unless they satisfied the include/exclude conditions. This is
2588            inconsistent with GNU grep (and could even be seen as contrary to the
2589            pcregrep specification - which I improved to make it absolutely clear).
2590            The action now is always to scan all levels of directory, and just
2591            apply the include/exclude patterns to regular files.
2592    
2593    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2594        --exclude_dir in the tests to avoid scanning .svn directories.
2595    
2596    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2597        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2598        doesn't support NULs in patterns.
2599    
2600    7.  Added some missing "const"s to declarations of static tables in
2601        pcre_compile.c and pcre_dfa_exec.c.
2602    
2603    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2604        caused by fix #2  above. (Subsequently also a second patch to fix the
2605        first patch. And a third patch - this was a messy problem.)
2606    
2607    9.  Applied Craig's patch to remove the use of push_back().
2608    
2609    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2610        matching function regexec().
2611    
2612    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2613        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2614        references. PCRE supports relative numbers with this syntax (I don't think
2615        Oniguruma does).
2616    
2617    12. Previously, a group with a zero repeat such as (...){0} was completely
2618        omitted from the compiled regex. However, this means that if the group
2619        was called as a subroutine from elsewhere in the pattern, things went wrong
2620        (an internal error was given). Such groups are now left in the compiled
2621        pattern, with a new opcode that causes them to be skipped at execution
2622        time.
2623    
2624    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2625        to the way PCRE behaves:
2626    
2627        (a) A lone ] character is dis-allowed (Perl treats it as data).
2628    
2629        (b) A back reference to an unmatched subpattern matches an empty string
2630            (Perl fails the current match path).
2631    
2632        (c) A data ] in a character class must be notated as \] because if the
2633            first data character in a class is ], it defines an empty class. (In
2634            Perl it is not possible to have an empty class.) The empty class []
2635            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2636            The negative empty class [^] matches any one character, independently
2637            of the DOTALL setting.
2638    
2639    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2640        non-existent subpattern following a character class starting with ']' and
2641        containing () gave an internal compiling error instead of "reference to
2642        non-existent subpattern". Fortunately, when the pattern did exist, the
2643        compiled code was correct. (When scanning forwards to check for the
2644        existence of the subpattern, it was treating the data ']' as terminating
2645        the class, so got the count wrong. When actually compiling, the reference
2646        was subsequently set up correctly.)
2647    
2648    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2649        it was being rejected as not supported by pcre_dfa_exec(), even though
2650        other assertions are supported. I have made pcre_dfa_exec() support
2651        (*FAIL).
2652    
2653    16. The implementation of 13c above involved the invention of a new opcode,
2654        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2655        cannot be changed at match time, I realized I could make a small
2656        improvement to matching performance by compiling OP_ALLANY instead of
2657        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2658        on the OP_ANY path.
2659    
2660    17. Compiling pcretest on Windows with readline support failed without the
2661        following two fixes: (1) Make the unistd.h include conditional on
2662        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2663    
2664    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2665        ncurses library to be included for pcretest when ReadLine support is
2666        requested, but also to allow for it to be overridden. This patch came from
2667        Daniel Bergström.
2668    
2669    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2670        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2671        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2672        spotting this.
2673    
2674    
2675    Version 7.6 28-Jan-08
2676    ---------------------
2677    
2678    1.  A character class containing a very large number of characters with
2679        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2680        overflow.
2681    
2682    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2683        HAVE_LONG_LONG is not defined.
2684    
2685    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2686        bring them up to date and include new features. This patch includes:
2687    
2688        - Fixed PH's badly added libz and libbz2 support.
2689        - Fixed a problem with static linking.
2690        - Added pcredemo. [But later removed - see 7 below.]
2691        - Fixed dftables problem and added an option.
2692        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2693            HAVE_LONG_LONG.
2694        - Added readline support for pcretest.
2695        - Added an listing of the option settings after cmake has run.
2696    
2697    4.  A user submitted a patch to Makefile that makes it easy to create
2698        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2699        Makefile.am that cause it to include this special target, without
2700        affecting anything else. Note that the same mingw target plus all
2701        the other distribution libraries and programs are now supported
2702        when configuring with CMake (see 6 below) instead of with
2703        Configure/Make.
2704    
2705    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2706        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2707        exported in the Windows port". It has not yet been confirmed that the patch
2708        solves the problem, but it does no harm.
2709    
2710    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2711        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2712        with CMake, and also correct the comment about stack recursion.
2713    
2714    7.  Remove the automatic building of pcredemo from the ./configure system and
2715        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2716        of a program that users should build themselves after PCRE is installed, so
2717        building it automatically is not really right. What is more, it gave
2718        trouble in some build environments.
2719    
2720    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2721    
2722    
2723    Version 7.5 10-Jan-08
2724    ---------------------
2725    
2726    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2727        values in parens when parsing an RE using the C++ wrapper."
2728    
2729    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2730        Characters greater than 255 were excluded from the class instead of being
2731        included.
2732    
2733    3.  The same bug as (2) above applied to negated POSIX classes such as
2734        [:^space:].
2735    
2736    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2737        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2738        I have changed it.
2739    
2740    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2741        first named subpattern) and a construct such as (?&a) would reference the
2742        first named subpattern whose name started with "a" (in other words, the
2743        length check was missing). Both these problems are fixed. "Subpattern name
2744        expected" is now given for (?&) (a zero-length name), and this patch also
2745        makes it give the same error for \k'' (previously it complained that that
2746        was a reference to a non-existent subpattern).
2747    
2748    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2749        this is right because (?- can be followed by option settings as well as by
2750        digits. I have, however, made the messages clearer.
2751    
2752    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2753        than the number used in the conditional) now cause a compile-time error.
2754        This is actually not compatible with Perl, which accepts such patterns, but
2755        treats the conditional as always being FALSE (as PCRE used to), but it
2756        seems to me that giving a diagnostic is better.
2757    
2758    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2759        and messages.
2760    
2761    9.  Fix two occurrences of "backslash" in comments that should have been
2762        "backspace".
2763    
2764    10. Remove two redundant lines of code that can never be obeyed (their function
2765        was moved elsewhere).
2766    
2767    11. The program that makes PCRE's Unicode character property table had a bug
2768        which caused it to generate incorrect table entries for sequences of
2769        characters that have the same character type, but are in different scripts.
2770        It amalgamated them into a single range, with the script of the first of
2771        them. In other words, some characters were in the wrong script. There were
2772        thirteen such cases, affecting characters in the following ranges:
2773    
2774          U+002b0 - U+002c1
2775          U+0060c - U+0060d
2776          U+0061e - U+00612
2777          U+0064b - U+0065e
2778          U+0074d - U+0076d
2779          U+01800 - U+01805
2780          U+01d00 - U+01d77
2781          U+01d9b - U+01dbf
2782          U+0200b - U+0200f
2783          U+030fc - U+030fe
2784          U+03260 - U+0327f
2785          U+0fb46 - U+0fbb1
2786          U+10450 - U+1049d
2787    
2788    12. The -o option (show only the matching part of a line) for pcregrep was not
2789        compatible with GNU grep in that, if there was more than one match in a
2790        line, it showed only the first of them. It now behaves in the same way as
2791        GNU grep.
2792    
2793    13. If the -o and -v options were combined for pcregrep, it printed a blank
2794        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2795        does the same. The return code can be used to tell if there were any
2796        non-matching lines.
2797    
2798    14. Added --file-offsets and --line-offsets to pcregrep.
2799    
2800    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2801        infinitely looping recursion. The bug was that positive lookaheads were not
2802        being skipped when checking for a possible empty match (negative lookaheads
2803        and both kinds of lookbehind were skipped).
2804    
2805    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2806        inclusion of <windows.h> to before rather than after the definition of
2807        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2808    
2809    17. Specifying a possessive quantifier with a specific limit for a Unicode
2810        character property caused pcre_compile() to compile bad code, which led at
2811        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2812        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2813        caused the error; without that there was no problem.
2814    
2815    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2816    
2817    19. Added --enable-pcretest-libreadline.
2818    
2819    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2820        RE::GlobalReplace(). As a result, the number of replacements returned was
2821        double what it should be. I removed one of the increments, but Craig sent a
2822        later patch that removed the other one (the right fix) and added unit tests
2823        that check the return values (which was not done before).
2824    
2825    21. Several CMake things:
2826    
2827        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2828            the names libpcre and libpcreposix, not just pcre and pcreposix.
2829    
2830        (2) The above change means that pcretest and pcregrep are now correctly
2831            linked with the newly-built libraries, not previously installed ones.
2832    
2833        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2834    
2835    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2836        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2837        UTF-8 newline character). The key issue is that the pattern starts .*;
2838        this means that the match must be either at the beginning, or after a
2839        newline. The bug was in the code for advancing after a failed match and
2840        checking that the new position followed a newline. It was not taking
2841        account of UTF-8 characters correctly.
2842    
2843    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2844        character classes. PCRE was not treating the sequence [:...:] as a
2845        character class unless the ... were all letters. Perl, however, seems to
2846        allow any characters between [: and :], though of course it rejects as
2847        unknown any "names" that contain non-letters, because all the known class
2848        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2849        for example, whereas PCRE did not - it did not recognize a POSIX character
2850        class. This seemed a bit dangerous, so the code has been changed to be
2851        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2852        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2853        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2854        Perl does, and where it didn't before.
2855    
2856    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2857        Windows environments %n is disabled by default.
2858    
2859    
2860    Version 7.4 21-Sep-07
2861    ---------------------
2862    
2863    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2864        means that a class such as [\s] counted as "explicit reference to CR or
2865        LF". That isn't really right - the whole point of the change was to try to
2866        help when there was an actual mention of one of the two characters. So now
2867        the change happens only if \r or \n (or a literal CR or LF) character is
2868        encountered.
2869    
2870    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2871        of both had grown to the point where there were only 3 bits left.
2872        Fortunately, there was spare space in the data structure, and so I have
2873        moved the internal flags into a new 16-bit field to free up more option
2874        bits.
2875    
2876    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2877        but did not set the internal JCHANGED flag - either of these is enough to
2878        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2879        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2880        start sets both bits.
2881    
2882    4.  Added options (at build time, compile time, exec time) to change \R from
2883        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2884    
2885    5.  doc/pcresyntax.html was missing from the distribution.
2886    
2887    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2888        compatibility, even though it is no longer used.
2889    
2890    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2891        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2892        windows.h file is present (where different names are used). [This was
2893        reversed later after testing - see 16 below.]
2894    
2895    8.  Changed all #include <config.h> to #include "config.h". There were also
2896        some further <pcre.h> cases that I changed to "pcre.h".
2897    
2898    9.  When pcregrep was used with the --colour option, it missed the line ending
2899        sequence off the lines that it output.
2900    
2901    10. It was pointed out to me that arrays of string pointers cause lots of
2902        relocations when a shared library is dynamically loaded. A technique of
2903        using a single long string with a table of offsets can drastically reduce
2904        these. I have refactored PCRE in four places to do this. The result is
2905        dramatic:
2906    
2907          Originally:                          290
2908          After changing UCP table:            187
2909          After changing error message table:   43
2910          After changing table of "verbs"       36
2911          After changing table of Posix names   22
2912    
2913        Thanks to the folks working on Gregex for glib for this insight.
2914    
2915    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2916        unicode-properties was also set.
2917    
2918    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2919    
2920    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2921        checked only for CRLF.
2922    
2923    14. Added casts to pcretest.c to avoid compiler warnings.
2924    
2925    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2926    
2927    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2928        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2929        entirely. This removes changes made in 7 above.
2930    
2931    17. The CMake files have been updated, and there is now more information about
2932        building with CMake in the NON-UNIX-USE document.
2933    
2934    
2935    Version 7.3 28-Aug-07
2936  ---------------------  ---------------------
2937    
2938   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 3074  Version 7.3 17-Aug-07
3074  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
3075      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").
3076    
3077    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3078    
3079    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3080        character were causing crashes (broken optimization).
3081    
3082    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3083        \p or \P) caused a compile-time loop.
3084    
3085    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3086        break. For example, the unstudied pattern [\r\n]A does not match the string
3087        "\r\nA" because change 7.0/46 below moves the current point on by two
3088        characters after failing to match at the start. However, the pattern \nA
3089        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3090        the same is true. There doesn't seem any very clean way out of this, but
3091        what I have chosen to do makes the common cases work: PCRE now takes note
3092        of whether there can be an explicit match for \r or \n anywhere in the
3093        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3094        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3095        pattern has explicit CR or LF references.
3096    
3097    29. Added (*CR) etc for changing newline setting at start of pattern.
3098    
3099    
3100  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3101  ---------------------  ---------------------

Legend:
Removed from v.221  
changed lines
  Added in v.1591

  ViewVC Help
Powered by ViewVC 1.1.5