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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1568 by ph10, Sun Jun 14 15:53:41 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.38 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    
65    Version 8.37 28-April-2015
66    --------------------------
67    
68    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
69        for those parentheses to be closed with whatever has been captured so far.
70        However, it was failing to mark any other groups between the hightest
71        capture so far and the currrent group as "unset". Thus, the ovector for
72        those groups contained whatever was previously there. An example is the
73        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
74    
75    2.  If an assertion condition was quantified with a minimum of zero (an odd
76        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
77    
78    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
79        unrecognized modifier, a crash could occur.
80    
81    4.  An attempt to do global matching in pcretest with a zero-length ovector
82        caused a crash.
83    
84    5.  Fixed a memory leak during matching that could occur for a subpattern
85        subroutine call (recursive or otherwise) if the number of captured groups
86        that had to be saved was greater than ten.
87    
88    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
89        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
90        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
91    
92    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
93        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
94    
95    8. When a pattern is compiled, it remembers the highest back reference so that
96       when matching, if the ovector is too small, extra memory can be obtained to
97       use instead. A conditional subpattern whose condition is a check on a
98       capture having happened, such as, for example in the pattern
99       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
100       setting the highest backreference number. This mattered only if pcre_exec()
101       was called with an ovector that was too small to hold the capture, and there
102       was no other kind of back reference (a situation which is probably quite
103       rare). The effect of the bug was that the condition was always treated as
104       FALSE when the capture could not be consulted, leading to a incorrect
105       behaviour by pcre_exec(). This bug has been fixed.
106    
107    9. A reference to a duplicated named group (either a back reference or a test
108       for being set in a conditional) that occurred in a part of the pattern where
109       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
110       to be incorrectly calculated, leading to overwriting.
111    
112    10. A mutually recursive set of back references such as (\2)(\1) caused a
113        segfault at study time (while trying to find the minimum matching length).
114        The infinite loop is now broken (with the minimum length unset, that is,
115        zero).
116    
117    11. If an assertion that was used as a condition was quantified with a minimum
118        of zero, matching went wrong. In particular, if the whole group had
119        unlimited repetition and could match an empty string, a segfault was
120        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
121        assertions to be quantified, but not if they are being used as conditions,
122        so the above pattern is faulted by Perl. PCRE has now been changed so that
123        it also rejects such patterns.
124    
125    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
126        failed to allow the zero-repeat case if pcre2_exec() was called with an
127        ovector too small to capture the group.
128    
129    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
130        Red Hat Product Security:
131    
132        (a) A crash if /K and /F were both set with the option to save the compiled
133        pattern.
134    
135        (b) Another crash if the option to print captured substrings in a callout
136        was combined with setting a null ovector, for example \O\C+ as a subject
137        string.
138    
139    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
140        forward reference repeated a large (but limited) number of times within a
141        repeated outer group that has a zero minimum quantifier, caused incorrect
142        code to be compiled, leading to the error "internal error:
143        previously-checked referenced subpattern not found" when an incorrect
144        memory address was read. This bug was reported as "heap overflow",
145        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
146        CVE-2015-2325.
147    
148    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
149        call within a group that also contained a recursive back reference caused
150        incorrect code to be compiled. This bug was reported as "heap overflow",
151        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
152        number CVE-2015-2326.
153    
154    24. Computing the size of the JIT read-only data in advance has been a source
155        of various issues, and new ones are still appear unfortunately. To fix
156        existing and future issues, size computation is eliminated from the code,
157        and replaced by on-demand memory allocation.
158    
159    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
160        adjacent to the end of the range, and the range contained characters with
161        more than one other case, caused incorrect behaviour when compiled in UTF
162        mode. In that example, the range a-j was left out of the class.
163    
164    26. Fix JIT compilation of conditional blocks, which assertion
165        is converted to (*FAIL). E.g: /(?(?!))/.
166    
167    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
168        discovered by the LLVM fuzzer.
169    
170    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
171        when this assertion was used as a condition, for example (?(?!)a|b). In
172        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
173        error about an unsupported item.
174    
175    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
176        possessification code could take exponential time to complete. A recursion
177        depth limit of 1000 has been imposed to limit the resources used by this
178        optimization.
179    
180    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
181        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
182        because \S ensures they are all in the class. The code for doing this was
183        interacting badly with the code for computing the amount of space needed to
184        compile the pattern, leading to a buffer overflow. This bug was discovered
185        by the LLVM fuzzer.
186    
187    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
188        other kinds of group caused stack overflow at compile time. This bug was
189        discovered by the LLVM fuzzer.
190    
191    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
192        between a subroutine call and its quantifier was incorrectly compiled,
193        leading to buffer overflow or other errors. This bug was discovered by the
194        LLVM fuzzer.
195    
196    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
197        assertion after (?(. The code was failing to check the character after
198        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
199        was discovered by the LLVM fuzzer.
200    
201    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
202        a fixed maximum following a group that contains a subroutine reference was
203        incorrectly compiled and could trigger buffer overflow. This bug was
204        discovered by the LLVM fuzzer.
205    
206    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
207        caused a stack overflow instead of the diagnosis of a non-fixed length
208        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
209    
210    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
211        (e.g. /(?<=\Ka)/) could make pcregrep loop.
212    
213    37. There was a similar problem to 36 in pcretest for global matches.
214    
215    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
216        and a subsequent item in the pattern caused a non-match, backtracking over
217        the repeated \X did not stop, but carried on past the start of the subject,
218        causing reference to random memory and/or a segfault. There were also some
219        other cases where backtracking after \C could crash. This set of bugs was
220        discovered by the LLVM fuzzer.
221    
222    39. The function for finding the minimum length of a matching string could take
223        a very long time if mutual recursion was present many times in a pattern,
224        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
225        method has been implemented. This infelicity was discovered by the LLVM
226        fuzzer.
227    
228    40. Static linking against the PCRE library using the pkg-config module was
229        failing on missing pthread symbols.
230    
231    
232    Version 8.36 26-September-2014
233    ------------------------------
234    
235    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
236        -Wmissing-field-initializers and -Wunused-parameter.
237    
238    2.  The tests for quantifiers being too big (greater than 65535) were being
239        applied after reading the number, and stupidly assuming that integer
240        overflow would give a negative number. The tests are now applied as the
241        numbers are read.
242    
243    3.  Tidy code in pcre_exec.c where two branches that used to be different are
244        now the same.
245    
246    4.  The JIT compiler did not generate match limit checks for certain
247        bracketed expressions with quantifiers. This may lead to exponential
248        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
249        issue should be resolved now.
250    
251    5.  Fixed an issue, which occures when nested alternatives are optimized
252        with table jumps.
253    
254    6.  Inserted two casts and changed some ints to size_t in the light of some
255        reported 64-bit compiler warnings (Bugzilla 1477).
256    
257    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
258        an empty string, which sometimes were behaving incorrectly in the
259        interpreter (though correctly in the JIT matcher). This pcretest input is
260        an example:
261    
262          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
263          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
264    
265        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
266        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
267        for an empty string was breaking the inner loop and carrying on at a lower
268        level, when possessive repeated groups should always return to a higher
269        level as they have no backtrack points in them. The empty string test now
270        occurs at the outer level.
271    
272    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
273        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
274    
275    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
276        doesn't).
277    
278    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
279        to hold a pattern to be incorrectly computed (too small) when there were
280        named back references to duplicated names. This could cause "internal
281        error: code overflow" or "double free or corruption" or other memory
282        handling errors.
283    
284    11. When named subpatterns had the same prefixes, back references could be
285        confused. For example, in this pattern:
286    
287          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
288    
289        the reference to 'Name' was incorrectly treated as a reference to a
290        duplicate name.
291    
292    12. A pattern such as /^s?c/mi8 where the optional character has more than
293        one "other case" was incorrectly compiled such that it would only try to
294        match starting at "c".
295    
296    13. When a pattern starting with \s was studied, VT was not included in the
297        list of possible starting characters; this should have been part of the
298        8.34/18 patch.
299    
300    14. If a character class started [\Qx]... where x is any character, the class
301        was incorrectly terminated at the ].
302    
303    15. If a pattern that started with a caseless match for a character with more
304        than one "other case" was studied, PCRE did not set up the starting code
305        unit bit map for the list of possible characters. Now it does. This is an
306        optimization improvement, not a bug fix.
307    
308    16. The Unicode data tables have been updated to Unicode 7.0.0.
309    
310    17. Fixed a number of memory leaks in pcregrep.
311    
312    18. Avoid a compiler warning (from some compilers) for a function call with
313        a cast that removes "const" from an lvalue by using an intermediate
314        variable (to which the compiler does not object).
315    
316    19. Incorrect code was compiled if a group that contained an internal recursive
317        back reference was optional (had quantifier with a minimum of zero). This
318        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
319        caused segmentation faults because of stack overflows at compile time.
320    
321    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
322        group that is quantified with an indefinite repeat, caused a compile-time
323        loop which used up all the system stack and provoked a segmentation fault.
324        This was not the same bug as 19 above.
325    
326    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
327        Patch by Mike Frysinger.
328    
329    
330    Version 8.35 04-April-2014
331    --------------------------
332    
333    1.  A new flag is set, when property checks are present in an XCLASS.
334        When this flag is not set, PCRE can perform certain optimizations
335        such as studying these XCLASS-es.
336    
337    2.  The auto-possessification of character sets were improved: a normal
338        and an extended character set can be compared now. Furthermore
339        the JIT compiler optimizes more character set checks.
340    
341    3.  Got rid of some compiler warnings for potentially uninitialized variables
342        that show up only when compiled with -O2.
343    
344    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
345        of a match later then the end of the match. The pcretest program was not
346        handling the case sensibly - it was outputting from the start to the next
347        binary zero. It now reports this situation in a message, and outputs the
348        text from the end to the start.
349    
350    5.  Fast forward search is improved in JIT. Instead of the first three
351        characters, any three characters with fixed position can be searched.
352        Search order: first, last, middle.
353    
354    6.  Improve character range checks in JIT. Characters are read by an inprecise
355        function now, which returns with an unknown value if the character code is
356        above a certain threshold (e.g: 256). The only limitation is that the value
357        must be bigger than the threshold as well. This function is useful when
358        the characters above the threshold are handled in the same way.
359    
360    7.  The macros whose names start with RAWUCHAR are placeholders for a future
361        mode in which only the bottom 21 bits of 32-bit data items are used. To
362        make this more memorable for those maintaining the code, the names have
363        been changed to start with UCHAR21, and an extensive comment has been added
364        to their definition.
365    
366    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
367        to the export list in Makefile.am (they were accidentally omitted from the
368        8.34 tarball).
369    
370    9.  The informational output from pcretest used the phrase "starting byte set"
371        which is inappropriate for the 16-bit and 32-bit libraries. As the output
372        for "first char" and "need char" really means "non-UTF-char", I've changed
373        "byte" to "char", and slightly reworded the output. The documentation about
374        these values has also been (I hope) clarified.
375    
376    10. Another JIT related optimization: use table jumps for selecting the correct
377        backtracking path, when more than four alternatives are present inside a
378        bracket.
379    
380    11. Empty match is not possible, when the minimum length is greater than zero,
381        and there is no \K in the pattern. JIT should avoid empty match checks in
382        such cases.
383    
384    12. In a caseless character class with UCP support, when a character with more
385        than one alternative case was not the first character of a range, not all
386        the alternative cases were added to the class. For example, s and \x{17f}
387        are both alternative cases for S: the class [RST] was handled correctly,
388        but [R-T] was not.
389    
390    13. The configure.ac file always checked for pthread support when JIT was
391        enabled. This is not used in Windows, so I have put this test inside a
392        check for the presence of windows.h (which was already tested for).
393    
394    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
395        The algorithm provides a way to skip certain starting offsets, and usually
396        faster than linear prefix searches.
397    
398    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
399        as for 'fr_FR' and 'french'. For some reason, however, it then used the
400        Windows-specific input and output files, which have 'french' screwed in.
401        So this could never have worked. One of the problems with locales is that
402        they aren't always the same. I have now updated RunTest so that it checks
403        the output of the locale test (test 3) against three different output
404        files, and it allows the test to pass if any one of them matches. With luck
405        this should make the test pass on some versions of Solaris where it was
406        failing. Because of the uncertainty, the script did not used to stop if
407        test 3 failed; it now does. If further versions of a French locale ever
408        come to light, they can now easily be added.
409    
410    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
411        there was a message during ./configure, but it did not stop. This now
412        provokes an error. The invalid example in README has been corrected.
413        If a value less than the minimum is given, the minimum value has always
414        been used, but now a warning is given.
415    
416    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
417        was a bug in the test system, which is now fixed. Also, the list of various
418        configurations that are tested for each release did not have one with both
419        16/32 bits and --enable-bar-anycrlf. It now does.
420    
421    18. pcretest was missing "-C bsr" for displaying the \R default setting.
422    
423    19. Little endian PowerPC systems are supported now by the JIT compiler.
424    
425    20. The fast forward newline mechanism could enter to an infinite loop on
426        certain invalid UTF-8 input. Although we don't support these cases
427        this issue can be fixed by a performance optimization.
428    
429    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
430        not take account if existing stack usage. There is now a new global
431        variable called pcre_stack_guard that can be set to point to an external
432        function to check stack availability. It is called at the start of
433        processing every parenthesized group.
434    
435    22. A typo in the code meant that in ungreedy mode the max/min qualifier
436        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
437        not match "ab".
438    
439    23. When UTF was disabled, the JIT program reported some incorrect compile
440        errors. These messages are silenced now.
441    
442    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
443        compiler.
444    
445    25. Change all the temporary files used in RunGrepTest to be different to those
446        used by RunTest so that the tests can be run simultaneously, for example by
447        "make -j check".
448    
449    
450    Version 8.34 15-December-2013
451    -----------------------------
452    
453    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
454        executable memory. Patch inspired by Carsten Klein.
455    
456    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
457        this macro is never tested and has no effect, because the work to support
458        coverage involves only compiling and linking options and special targets in
459        the Makefile. The comment in config.h implied that defining the macro would
460        enable coverage support, which is totally false. There was also support for
461        setting this macro in the CMake files (my fault, I just copied it from
462        configure). SUPPORT_GCOV has now been removed.
463    
464    3.  Make a small performance improvement in strlen16() and strlen32() in
465        pcretest.
466    
467    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
468        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
469        these cases). There was also one in pcretest.c.
470    
471    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
472    
473    6.  In UTF mode, the code for checking whether a group could match an empty
474        string (which is used for indefinitely repeated groups to allow for
475        breaking an infinite loop) was broken when the group contained a repeated
476        negated single-character class with a character that occupied more than one
477        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
478        UTF-8 mode). The effect was undefined: the group might or might not be
479        deemed as matching an empty string, or the program might have crashed.
480    
481    7.  The code for checking whether a group could match an empty string was not
482        recognizing that \h, \H, \v, \V, and \R must match a character.
483    
484    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
485        an empty string. If it can, pcretest shows this in its information output.
486    
487    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
488        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
489        matched by pcre_exec() without using JIT:
490    
491        (a) If the rest of the pattern did not match after a maximal run of
492            grapheme clusters, the code for backing up to try with fewer of them
493            did not always back up over a full grapheme when characters that do not
494            have the modifier quality were involved, e.g. Hangul syllables.
495    
496        (b) If the match point in a subject started with modifier character, and
497            there was no match, the code could incorrectly back up beyond the match
498            point, and potentially beyond the first character in the subject,
499            leading to a segfault or an incorrect match result.
500    
501    10. A conditional group with an assertion condition could lead to PCRE
502        recording an incorrect first data item for a match if no other first data
503        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
504        first data item, and therefore matched "ca" after "c" instead of at the
505        start.
506    
507    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
508        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
509    
510    12. The source of pcregrep now includes z/OS-specific code so that it can be
511        compiled for z/OS as part of the special z/OS distribution.
512    
513    13. Added the -T and -TM options to pcretest.
514    
515    14. The code in pcre_compile.c for creating the table of named capturing groups
516        has been refactored. Instead of creating the table dynamically during the
517        actual compiling pass, the information is remembered during the pre-compile
518        pass (on the stack unless there are more than 20 named groups, in which
519        case malloc() is used) and the whole table is created before the actual
520        compile happens. This has simplified the code (it is now nearly 150 lines
521        shorter) and prepared the way for better handling of references to groups
522        with duplicate names.
523    
524    15. A back reference to a named subpattern when there is more than one of the
525        same name now checks them in the order in which they appear in the pattern.
526        The first one that is set is used for the reference. Previously only the
527        first one was inspected. This change makes PCRE more compatible with Perl.
528    
529    16. Unicode character properties were updated from Unicode 6.3.0.
530    
531    17. The compile-time code for auto-possessification has been refactored, based
532        on a patch by Zoltan Herczeg. It now happens after instead of during
533        compilation. The code is cleaner, and more cases are handled. The option
534        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
535        options in pcretest are provided to set it. It can also be set by
536        (*NO_AUTO_POSSESS) at the start of a pattern.
537    
538    18. The character VT has been added to the default ("C" locale) set of
539        characters that match \s and are generally treated as white space,
540        following this same change in Perl 5.18. There is now no difference between
541        "Perl space" and "POSIX space". Whether VT is treated as white space in
542        other locales depends on the locale.
543    
544    19. The code for checking named groups as conditions, either for being set or
545        for being recursed, has been refactored (this is related to 14 and 15
546        above). Processing unduplicated named groups should now be as fast at
547        numerical groups, and processing duplicated groups should be faster than
548        before.
549    
550    20. Two patches to the CMake build system, by Alexander Barkov:
551    
552          (1) Replace the "source" command by "." in CMakeLists.txt because
553              "source" is a bash-ism.
554    
555          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
556              without these the CMake build does not work on Solaris.
557    
558    21. Perl has changed its handling of \8 and \9. If there is no previously
559        encountered capturing group of those numbers, they are treated as the
560        literal characters 8 and 9 instead of a binary zero followed by the
561        literals. PCRE now does the same.
562    
563    22. Following Perl, added \o{} to specify codepoints in octal, making it
564        possible to specify values greater than 0777 and also making them
565        unambiguous.
566    
567    23. Perl now gives an error for missing closing braces after \x{... instead of
568        treating the string as literal. PCRE now does the same.
569    
570    24. RunTest used to grumble if an inappropriate test was selected explicitly,
571        but just skip it when running all tests. This make it awkward to run ranges
572        of tests when one of them was inappropriate. Now it just skips any
573        inappropriate tests, as it always did when running all tests.
574    
575    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
576        character types such as \d or \w, too many callouts were inserted, and the
577        data that they returned was rubbish.
578    
579    26. In UCP mode, \s was not matching two of the characters that Perl matches,
580        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
581        were matched by \h. The code has now been refactored so that the lists of
582        the horizontal and vertical whitespace characters used for \h and \v (which
583        are defined only in one place) are now also used for \s.
584    
585    27. Add JIT support for the 64 bit TileGX architecture.
586        Patch by Jiong Wang (Tilera Corporation).
587    
588    28. Possessive quantifiers for classes (both explicit and automatically
589        generated) now use special opcodes instead of wrapping in ONCE brackets.
590    
591    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
592        (because it's meaningless), this was not happening when PCRE_CASELESS was
593        set. Not wrong, but inefficient.
594    
595    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
596        properties for \w, \d, etc) is present in a test regex. Otherwise if the
597        test contains no characters greater than 255, Perl doesn't realise it
598        should be using Unicode semantics.
599    
600    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
601        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
602        does in Unicode mode.
603    
604    32. Added the "forbid" facility to pcretest so that putting tests into the
605        wrong test files can sometimes be quickly detected.
606    
607    33. There is now a limit (default 250) on the depth of nesting of parentheses.
608        This limit is imposed to control the amount of system stack used at compile
609        time. It can be changed at build time by --with-parens-nest-limit=xxx or
610        the equivalent in CMake.
611    
612    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
613        errors. Perl warns for these when in warning mode, but PCRE has no facility
614        for giving warnings.
615    
616    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
617        However, this was not working for (?!) because it is optimized to (*FAIL),
618        for which PCRE does not allow quantifiers. The optimization is now disabled
619        when a quantifier follows (?!). I can't see any use for this, but it makes
620        things uniform.
621    
622    36. Perl no longer allows group names to start with digits, so I have made this
623        change also in PCRE. It simplifies the code a bit.
624    
625    37. In extended mode, Perl ignores spaces before a + that indicates a
626        possessive quantifier. PCRE allowed a space before the quantifier, but not
627        before the possessive +. It now does.
628    
629    38. The use of \K (reset reported match start) within a repeated possessive
630        group such as (a\Kb)*+ was not working.
631    
632    40. Document that the same character tables must be used at compile time and
633        run time, and that the facility to pass tables to pcre_exec() and
634        pcre_dfa_exec() is for use only with saved/restored patterns.
635    
636    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
637        features for Builds with MSVC:
638    
639        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
640           builds). The .rc files can be used to set FileDescription and many other
641           attributes.
642    
643        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
644           This allows higher-level build scripts which want .pdb files to avoid
645           hard-coding the exact files needed."
646    
647    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
648        mean "start of word" and "end of word", respectively, as a transition aid.
649    
650    43. A minimizing repeat of a class containing codepoints greater than 255 in
651        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
652        compiled to use the heap for recursion.
653    
654    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
655        is configured.
656    
657    
658    Version 8.33 28-May-2013
659    ------------------------
660    
661    1.  Added 'U' to some constants that are compared to unsigned integers, to
662        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
663        variables that are added to signed variables, to ensure the result is
664        signed and can be negated.
665    
666    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
667        CMake config files.
668    
669    3.  Revise the creation of config.h.generic so that all boolean macros are
670        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
671        overriding via -D on the command line possible.
672    
673    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
674        to unsigned int is reported to make a quite noticeable speed difference in
675        a specific Windows environment. Testing on Linux did also appear to show
676        some benefit (and it is clearly not harmful). Also fixed the definition of
677        Xop which should be unsigned.
678    
679    5.  Related to (4), changing the definition of the intermediate variable cc
680        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
681        improvements.
682    
683    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
684        unnecessary spaces.
685    
686    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
687        and later.
688    
689    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
690    
691    9.  Optimizing fast_forward_start_bits in JIT.
692    
693    10. Adding support for callouts in JIT, and fixing some issues revealed
694        during this work. Namely:
695    
696        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
697    
698        (b) Minimum length was not checked before the matching is started.
699    
700    11. The value of capture_last that is passed to callouts was incorrect in some
701        cases when there was a capture on one path that was subsequently abandoned
702        after a backtrack. Also, the capture_last value is now reset after a
703        recursion, since all captures are also reset in this case.
704    
705    12. The interpreter no longer returns the "too many substrings" error in the
706        case when an overflowing capture is in a branch that is subsequently
707        abandoned after a backtrack.
708    
709    13. In the pathological case when an offset vector of size 2 is used, pcretest
710        now prints out the matched string after a yield of 0 or 1.
711    
712    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
713        Only supported by the JIT compiler at the moment.
714    
715    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
716    
717    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
718        offset of the starting point of the matching process, provided the offsets
719        vector is large enough.
720    
721    17. The \A escape now records a lookbehind value of 1, though its execution
722        does not actually inspect the previous character. This is to ensure that,
723        in partial multi-segment matching, at least one character from the old
724        segment is retained when a new segment is processed. Otherwise, if there
725        are no lookbehinds in the pattern, \A might match incorrectly at the start
726        of a new segment.
727    
728    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
729    
730    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
731        gives some modest performance improvement in 8-bit mode.
732    
733    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
734        be expressed in certain programming languages using Universal Character
735        Names.
736    
737    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
738        which points out that "non characters" are not "characters that may not
739        appear in Unicode strings" but rather "characters that are reserved for
740        internal use and have only local meaning".
741    
742    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
743        there was a conditional group that depended on an assertion, if the
744        assertion was false, the callout that immediately followed the alternation
745        in the condition was skipped when pcre_exec() was used for matching.
746    
747    23. Allow an explicit callout to be inserted before an assertion that is the
748        condition for a conditional group, for compatibility with automatic
749        callouts, which always insert a callout at this point.
750    
751    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
752        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
753    
754    25. (*PRUNE) is now supported by the JIT compiler.
755    
756    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
757    
758    27. Fix the case where there are two or more SKIPs with arguments that may be
759        ignored.
760    
761    28. (*SKIP) is now supported by the JIT compiler.
762    
763    29. (*THEN) is now supported by the JIT compiler.
764    
765    30. Update RunTest with additional test selector options.
766    
767    31. The way PCRE handles backtracking verbs has been changed in two ways.
768    
769        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
770        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
771        backtracking. In some cases this makes it more Perl-compatible, but Perl's
772        rather obscure rules do not always do the same thing.
773    
774        (2) Previously, backtracking verbs were confined within assertions. This is
775        no longer the case for positive assertions, except for (*ACCEPT). Again,
776        this sometimes improves Perl compatibility, and sometimes does not.
777    
778    32. A number of tests that were in test 2 because Perl did things differently
779        have been moved to test 1, because either Perl or PCRE has changed, and
780        these tests are now compatible.
781    
782    32. Backtracking control verbs are now handled in the same way in JIT and
783        interpreter.
784    
785    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
786        contained a forward subroutine reference caused a compile error.
787    
788    34. Auto-detect and optimize limited repetitions in JIT.
789    
790    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
791        blocking (*UTF) etc.
792    
793    36. In the interpreter, maximizing pattern repetitions for characters and
794        character types now use tail recursion, which reduces stack usage.
795    
796    37. The value of the max lookbehind was not correctly preserved if a compiled
797        and saved regex was reloaded on a host of different endianness.
798    
799    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
800        of the compiled pattern block, expand the flags field from 16 to 32 bits
801        because it was almost full.
802    
803    39. Try madvise first before posix_madvise.
804    
805    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
806        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
807        original change locked out all matches of empty strings. This has been
808        changed so that one match of an empty string per line is recognized.
809        Subsequent searches on the same line (for colouring or for --only-matching,
810        for example) do not recognize empty strings.
811    
812    41. Applied a user patch to fix a number of spelling mistakes in comments.
813    
814    42. Data lines longer than 65536 caused pcretest to crash.
815    
816    43. Clarified the data type for length and startoffset arguments for pcre_exec
817        and pcre_dfa_exec in the function-specific man pages, where they were
818        explicitly stated to be in bytes, never having been updated. I also added
819        some clarification to the pcreapi man page.
820    
821    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
822        a segmentation fault.
823    
824    
825    Version 8.32 30-November-2012
826    -----------------------------
827    
828    1.  Improved JIT compiler optimizations for first character search and single
829        character iterators.
830    
831    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
832        Patch by Daniel Richard G.
833    
834    3.  Single character iterator optimizations in the JIT compiler.
835    
836    4.  Improved JIT compiler optimizations for character ranges.
837    
838    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
839        Reported by Giuseppe D'Angelo.
840    
841    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
842        of a line, was being set incorrectly in cases where .* appeared inside
843        atomic brackets at the start of a pattern, or where there was a subsequent
844        *PRUNE or *SKIP.
845    
846    7.  Improved instruction cache flush for POWER/PowerPC.
847        Patch by Daniel Richard G.
848    
849    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
850        grep:
851    
852        (a) There is now no limit to the number of patterns to be matched.
853    
854        (b) An error is given if a pattern is too long.
855    
856        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
857            are now supported.
858    
859        (d) --exclude-from and --include-from (multiple use) have been added.
860    
861        (e) Exclusions and inclusions now apply to all files and directories, not
862            just to those obtained from scanning a directory recursively.
863    
864        (f) Multiple uses of -f and --file-list are now supported.
865    
866        (g) In a Windows environment, the default for -d has been changed from
867            "read" (the GNU grep default) to "skip", because otherwise the presence
868            of a directory in the file list provokes an error.
869    
870        (h) The documentation has been revised and clarified in places.
871    
872    9.  Improve the matching speed of capturing brackets.
873    
874    10. Changed the meaning of \X so that it now matches a Unicode extended
875        grapheme cluster.
876    
877    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
878        out POSIX threads when JIT support is configured.
879    
880    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
881    
882    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
883        structure could go wrong in environments where size_t is not the same size
884        as int.
885    
886    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
887        set.
888    
889    15. The EBCDIC support had decayed; later updates to the code had included
890        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
891        general tidy up of EBCDIC-related issues, and the documentation was also
892        not quite right. There is now a test that can be run on ASCII systems to
893        check some of the EBCDIC-related things (but is it not a full test).
894    
895    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
896        in a small tidy to the code.
897    
898    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
899    
900    18. If the --only-matching (-o) option in pcregrep is specified multiple
901        times, each one causes appropriate output. For example, -o1 -o2 outputs the
902        substrings matched by the 1st and 2nd capturing parentheses. A separating
903        string can be specified by --om-separator (default empty).
904    
905    19. Improving the first n character searches.
906    
907    20. Turn case lists for horizontal and vertical white space into macros so that
908        they are defined only once.
909    
910    21. This set of changes together give more compatible Unicode case-folding
911        behaviour for characters that have more than one other case when UCP
912        support is available.
913    
914        (a) The Unicode property table now has offsets into a new table of sets of
915            three or more characters that are case-equivalent. The MultiStage2.py
916            script that generates these tables (the pcre_ucd.c file) now scans
917            CaseFolding.txt instead of UnicodeData.txt for character case
918            information.
919    
920        (b) The code for adding characters or ranges of characters to a character
921            class has been abstracted into a generalized function that also handles
922            case-independence. In UTF-mode with UCP support, this uses the new data
923            to handle characters with more than one other case.
924    
925        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
926            whose other case is greater than 256 are now correctly matched
927            caselessly. Previously, the high codepoint matched the low one, but not
928            vice versa.
929    
930        (d) The processing of \h, \H, \v, and \ in character classes now makes use
931            of the new class addition function, using character lists defined as
932            macros alongside the case definitions of 20 above.
933    
934        (e) Caseless back references now work with characters that have more than
935            one other case.
936    
937        (f) General caseless matching of characters with more than one other case
938            is supported.
939    
940    22. Unicode character properties were updated from Unicode 6.2.0
941    
942    23. Improved CMake support under Windows. Patch by Daniel Richard G.
943    
944    24. Add support for 32-bit character strings, and UTF-32
945    
946    25. Major JIT compiler update (code refactoring and bugfixing).
947        Experimental Sparc 32 support is added.
948    
949    26. Applied a modified version of Daniel Richard G's patch to create
950        pcre.h.generic and config.h.generic by "make" instead of in the
951        PrepareRelease script.
952    
953    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
954        pcre_compile.c when checking for a zero character.
955    
956    28. Introducing a native interface for JIT. Through this interface, the compiled
957        machine code can be directly executed. The purpose of this interface is to
958        provide fast pattern matching, so several sanity checks are not performed.
959        However, feature tests are still performed. The new interface provides
960        1.4x speedup compared to the old one.
961    
962    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
963        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
964        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
965    
966    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
967        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
968        the "old" RFC 2279). Instead, it ended up passing a negative length to
969        pcre_exec().
970    
971    31. Add support for GCC's visibility feature to hide internal functions.
972    
973    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
974        "unknown -C option" after outputting 0 or 1.
975    
976    33. There is now support for generating a code coverage report for the test
977        suite in environments where gcc is the compiler and lcov is installed. This
978        is mainly for the benefit of the developers.
979    
980    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
981        unaddressable using valgrind annotations, allowing valgrind to detect
982        invalid memory accesses. This is mainly for the benefit of the developers.
983    
984    25. (*UTF) can now be used to start a pattern in any of the three libraries.
985    
986    26. Give configure error if --enable-cpp but no C++ compiler found.
987    
988    
989    Version 8.31 06-July-2012
990    -------------------------
991    
992    1.  Fixing a wrong JIT test case and some compiler warnings.
993    
994    2.  Removed a bashism from the RunTest script.
995    
996    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
997        to unsigned type, result still unsigned" that was given by an MS compiler
998        on encountering the code "-sizeof(xxx)".
999    
1000    4.  Partial matching support is added to the JIT compiler.
1001    
1002    5.  Fixed several bugs concerned with partial matching of items that consist
1003        of more than one character:
1004    
1005        (a) /^(..)\1/ did not partially match "aba" because checking references was
1006            done on an "all or nothing" basis. This also applied to repeated
1007            references.
1008    
1009        (b) \R did not give a hard partial match if \r was found at the end of the
1010            subject.
1011    
1012        (c) \X did not give a hard partial match after matching one or more
1013            characters at the end of the subject.
1014    
1015        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1016            a partial match for the string "\r".
1017    
1018        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1019            a partial match for a CR character at the end of the subject string.
1020    
1021    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1022        running pcretest, the text "(JIT)" added to the output whenever JIT is
1023        actually used to run the match.
1024    
1025    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1026        or /S+[+] with a digit between 1 and 7.
1027    
1028    8.  OP_NOT now supports any UTF character not just single-byte ones.
1029    
1030    9.  (*MARK) control verb is now supported by the JIT compiler.
1031    
1032    10. The command "./RunTest list" lists the available tests without actually
1033        running any of them. (Because I keep forgetting what they all are.)
1034    
1035    11. Add PCRE_INFO_MAXLOOKBEHIND.
1036    
1037    12. Applied a (slightly modified) user-supplied patch that improves performance
1038        when the heap is used for recursion (compiled with --disable-stack-for-
1039        recursion). Instead of malloc and free for each heap frame each time a
1040        logical recursion happens, frames are retained on a chain and re-used where
1041        possible. This sometimes gives as much as 30% improvement.
1042    
1043    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1044        call.
1045    
1046    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1047    
1048    15. It is now possible to link pcretest with libedit as an alternative to
1049        libreadline.
1050    
1051    16. (*COMMIT) control verb is now supported by the JIT compiler.
1052    
1053    17. The Unicode data tables have been updated to Unicode 6.1.0.
1054    
1055    18. Added --file-list option to pcregrep.
1056    
1057    19. Added binary file support to pcregrep, including the -a, --binary-files,
1058        -I, and --text options.
1059    
1060    20. The madvise function is renamed for posix_madvise for QNX compatibility
1061        reasons. Fixed by Giuseppe D'Angelo.
1062    
1063    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1064    
1065    22. Changed the option for creating symbolic links for 16-bit man pages from
1066        -s to -sf so that re-installing does not cause issues.
1067    
1068    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1069    
1070    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1071        matches in certain environments (the workspace was not being correctly
1072        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1073        some of the workspace data at the beginning of a restart.
1074    
1075    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1076        was not doing so when it should - probably a typo introduced by SVN 528
1077        (change 8.10/14).
1078    
1079    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1080        \w+ when the character tables indicated that \x{c4} was a word character.
1081        There were several related cases, all because the tests for doing a table
1082        lookup were testing for characters less than 127 instead of 255.
1083    
1084    27. If a pattern contains capturing parentheses that are not used in a match,
1085        their slots in the ovector are set to -1. For those that are higher than
1086        any matched groups, this happens at the end of processing. In the case when
1087        there were back references that the ovector was too small to contain
1088        (causing temporary malloc'd memory to be used during matching), and the
1089        highest capturing number was not used, memory off the end of the ovector
1090        was incorrectly being set to -1. (It was using the size of the temporary
1091        memory instead of the true size.)
1092    
1093    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1094        ovector size, it uses memory at the end of the block that it has got.
1095    
1096    29. Check for an overlong MARK name and give an error at compile time. The
1097        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1098    
1099    30. JIT compiler update.
1100    
1101    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1102        Rill for the patch.
1103    
1104    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1105    
1106    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1107    
1108    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1109        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1110        was enabled.
1111    
1112    35. Improve JIT code generation for greedy plus quantifier.
1113    
1114    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1115        1 to "aa" instead of to an empty string. The bug affected repeated groups
1116        that could potentially match an empty string.
1117    
1118    37. Optimizing single character iterators in JIT.
1119    
1120    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1121        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1122        codepoints that are too big for the mode are faulted, and in a UTF mode,
1123        disallowed codepoints are also faulted.
1124    
1125    39. If PCRE was compiled with UTF support, in three places in the DFA
1126        matcher there was code that should only have been obeyed in UTF mode, but
1127        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1128        processing when bytes with values greater than 127 were present. In 16-bit
1129        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1130        both cases the values are those that cannot be the first data item in a UTF
1131        character. The three items that might have provoked this were recursions,
1132        possessively repeated groups, and atomic groups.
1133    
1134    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1135        and pcregrep, because some OS require shared objects to be explicitly
1136        passed to ld, causing the link step to fail if they are not.
1137    
1138    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1139        mode, patterns that started with \h* or \R* might be incorrectly matched.
1140    
1141    
1142    Version 8.30 04-February-2012
1143    -----------------------------
1144    
1145    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1146        name is defined in ctype.h.
1147    
1148    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1149        only in quite long subpatterns.
1150    
1151    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1152        since it was replaced by pcre_fullinfo() in February 2000.
1153    
1154    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1155        match a (*MARK), and the match failed at the start of the subject, a
1156        reference to memory before the start of the subject could occur. This bug
1157        was introduced by fix 17 of release 8.21.
1158    
1159    5.  A reference to an unset group with zero minimum repetition was giving
1160        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1161        /(another)?(\1?)test/ matched against "hello world test". This bug was
1162        introduced in release 8.13.
1163    
1164    6.  Add support for 16-bit character strings (a large amount of work involving
1165        many changes and refactorings).
1166    
1167    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1168        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1169        from a file.
1170    
1171    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1172        rounding is not applied in this particular case).
1173    
1174    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1175        if they appear, or are escaped, in patterns.
1176    
1177    10. Get rid of a number of -Wunused-but-set-variable warnings.
1178    
1179    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1180        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1181        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1182        also returns the mark "x". This bug applied to capturing parentheses,
1183        non-capturing parentheses, and atomic parentheses. It also applied to some
1184        assertions.
1185    
1186    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1187        information out of configure.ac instead of relying on pcre.h.generic, which
1188        is not stored in the repository.
1189    
1190    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1191        -lreadline.
1192    
1193    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1194    
1195    15. Applied Graycode's patch to put the top-level frame on the stack rather
1196        than the heap when not using the stack for recursion. This gives a
1197        performance improvement in many cases when recursion is not deep.
1198    
1199    16. Experimental code added to "pcretest -C" to output the stack frame size.
1200    
1201    
1202    Version 8.21 12-Dec-2011
1203    ------------------------
1204    
1205    1.  Updating the JIT compiler.
1206    
1207    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1208        are added as well.
1209    
1210    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1211        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1212        calling _pcre_jit_exec. Some extra comments are added.
1213    
1214    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1215        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1216        was introduced by change 18 for 8.20.
1217    
1218    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1219        ECMA-262 standard.
1220    
1221    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1222        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1223        This bug was probably introduced by change 9 of 8.13.
1224    
1225    7.  While fixing 6 above, I noticed that a number of other items were being
1226        incorrectly rejected as "not fixed length". This arose partly because newer
1227        opcodes had not been added to the fixed-length checking code. I have (a)
1228        corrected the bug and added tests for these items, and (b) arranged for an
1229        error to occur if an unknown opcode is encountered while checking for fixed
1230        length instead of just assuming "not fixed length". The items that were
1231        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1232        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1233        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1234    
1235    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1236        being incorrectly compiled and would have given unpredicatble results.
1237    
1238    9.  A possessively repeated subpattern with minimum repeat count greater than
1239        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1240        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1241        the first (A) could occur when it should not.
1242    
1243    10. Add a cast and remove a redundant test from the code.
1244    
1245    11. JIT should use pcre_malloc/pcre_free for allocation.
1246    
1247    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1248        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1249        is anything other than /usr, -L is still shown).
1250    
1251    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1252    
1253    14. Perl does not support \N without a following name in a [] class; PCRE now
1254        also gives an error.
1255    
1256    15. If a forward reference was repeated with an upper limit of around 2000,
1257        it caused the error "internal error: overran compiling workspace". The
1258        maximum number of forward references (including repeats) was limited by the
1259        internal workspace, and dependent on the LINK_SIZE. The code has been
1260        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1261        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1262        of around 200,000 forward references. While doing this, I also speeded up
1263        the filling in of repeated forward references.
1264    
1265    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1266        incorrectly expecting the subject to contain another "a" after the start.
1267    
1268    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1269        in the match, the SKIP should be ignored. This was not happening; instead
1270        the SKIP was being treated as NOMATCH. For patterns such as
1271        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1272        tested.
1273    
1274    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1275        now much more compatible with Perl, in particular in cases where the result
1276        is a non-match for a non-anchored pattern. For example, if
1277        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1278        "m", where previously it did not return a name. A side effect of this
1279        change is that for partial matches, the last encountered mark name is
1280        returned, as for non matches. A number of tests that were previously not
1281        Perl-compatible have been moved into the Perl-compatible test files. The
1282        refactoring has had the pleasing side effect of removing one argument from
1283        the match() function, thus reducing its stack requirements.
1284    
1285    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1286        subsequent uses of /S (without +) incorrectly behaved like /S+.
1287    
1288    21. Retrieve executable code size support for the JIT compiler and fixing
1289        some warnings.
1290    
1291    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1292        not work when the shorter character appeared right at the end of the
1293        subject string.
1294    
1295    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1296        systems.
1297    
1298    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1299        output it when the /M option is used in pcretest.
1300    
1301    25. The CheckMan script was not being included in the distribution. Also, added
1302        an explicit "perl" to run Perl scripts from the PrepareRelease script
1303        because this is reportedly needed in Windows.
1304    
1305    26. If study data was being save in a file and studying had not found a set of
1306        "starts with" bytes for the pattern, the data written to the file (though
1307        never used) was taken from uninitialized memory and so caused valgrind to
1308        complain.
1309    
1310    27. Updated RunTest.bat as provided by Sheri Pierce.
1311    
1312    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1313    
1314    29. Computation of memory usage for the table of capturing group names was
1315        giving an unnecessarily large value.
1316    
1317    
1318    Version 8.20 21-Oct-2011
1319    ------------------------
1320    
1321    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1322        a POSIX class. After further experiments with Perl, which convinced me that
1323        Perl has bugs and confusions, a closing square bracket is no longer allowed
1324        in a POSIX name. This bug also affected patterns with classes that started
1325        with full stops.
1326    
1327    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1328        captured substring, but while checking the failing first alternative,
1329        substring 1 is temporarily captured. If the output vector supplied to
1330        pcre_exec() was not big enough for this capture, the yield of the function
1331        was still zero ("insufficient space for captured substrings"). This cannot
1332        be totally fixed without adding another stack variable, which seems a lot
1333        of expense for a edge case. However, I have improved the situation in cases
1334        such as /(a)(b)x|abc/ matched against "abc", where the return code
1335        indicates that fewer than the maximum number of slots in the ovector have
1336        been set.
1337    
1338    3.  Related to (2) above: when there are more back references in a pattern than
1339        slots in the output vector, pcre_exec() uses temporary memory during
1340        matching, and copies in the captures as far as possible afterwards. It was
1341        using the entire output vector, but this conflicts with the specification
1342        that only 2/3 is used for passing back captured substrings. Now it uses
1343        only the first 2/3, for compatibility. This is, of course, another edge
1344        case.
1345    
1346    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1347        main code base, and can be used by building with --enable-jit. When this is
1348        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1349        runtime --no-jit option is given.
1350    
1351    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1352        ovector, the return from the function was zero, implying that there were
1353        other matches that did not fit. The correct "exactly full" value is now
1354        returned.
1355    
1356    6.  If a subpattern that was called recursively or as a subroutine contained
1357        (*PRUNE) or any other control that caused it to give a non-standard return,
1358        invalid errors such as "Error -26 (nested recursion at the same subject
1359        position)" or even infinite loops could occur.
1360    
1361    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1362        computing the minimum length on reaching *ACCEPT, and so ended up with the
1363        wrong value of 1 rather than 0. Further investigation indicates that
1364        computing a minimum subject length in the presence of *ACCEPT is difficult
1365        (think back references, subroutine calls), and so I have changed the code
1366        so that no minimum is registered for a pattern that contains *ACCEPT.
1367    
1368    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1369        it was not handled as intended. [But see 16 below.]
1370    
1371    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1372        Sheri Pierce.
1373    
1374    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1375        the first byte in a match must be "a".
1376    
1377    11. Change 17 for 8.13 increased the recursion depth for patterns like
1378        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1379        pattern contains any instances of (*THEN). If it does not, the old
1380        optimizations are restored. It would be nice to do this on a per-group
1381        basis, but at the moment that is not feasible.
1382    
1383    12. In some environments, the output of pcretest -C is CRLF terminated. This
1384        broke RunTest's code that checks for the link size. A single white space
1385        character after the value is now allowed for.
1386    
1387    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1388        For "fr", it uses the Windows-specific input and output files.
1389    
1390    14. If (*THEN) appeared in a group that was called recursively or as a
1391        subroutine, it did not work as intended. [But see next item.]
1392    
1393    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1394        pattern fragments (but not containing any | characters). If A and B are
1395        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1396        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1397        D. In other words, Perl considers parentheses that do not contain any |
1398        characters to be part of a surrounding alternative, whereas PCRE was
1399        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1400        differently. PCRE now behaves in the same way as Perl, except in the case
1401        of subroutine/recursion calls such as (?1) which have in any case always
1402        been different (but PCRE had them first :-).
1403    
1404    16. Related to 15 above: Perl does not treat the | in a conditional group as
1405        creating alternatives. Such a group is treated in the same way as an
1406        ordinary group without any | characters when processing (*THEN). PCRE has
1407        been changed to match Perl's behaviour.
1408    
1409    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1410        RunGrepTest script failed.
1411    
1412    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1413        inevitable for groups that contain captures, but it can lead to a lot of
1414        stack use in large patterns. The old behaviour has been restored for atomic
1415        groups that do not contain any capturing parentheses.
1416    
1417    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1418        suppress the check for a minimum subject length at run time. (If it was
1419        given to pcre_exec() or pcre_dfa_exec() it did work.)
1420    
1421    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1422        fail to work when decoding hex characters in data strings in EBCDIC
1423        environments.
1424    
1425    21. It appears that in at least one Mac OS environment, the isxdigit() function
1426        is implemented as a macro that evaluates to its argument more than once,
1427        contravening the C 90 Standard (I haven't checked a later standard). There
1428        was an instance in pcretest which caused it to go wrong when processing
1429        \x{...} escapes in subject strings. The has been rewritten to avoid using
1430        things like p++ in the argument of isxdigit().
1431    
1432    
1433    Version 8.13 16-Aug-2011
1434    ------------------------
1435    
1436    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1437    
1438    2.  Two minor typos in pcre_internal.h have been fixed.
1439    
1440    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1441        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1442        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1443    
1444    4.  There were a number of related bugs in the code for matching backrefences
1445        caselessly in UTF-8 mode when codes for the characters concerned were
1446        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1447        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1448        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1449        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1450        2-byte code at the end of the subject (it thought there wasn't enough data
1451        left).
1452    
1453    5.  Comprehensive information about what went wrong is now returned by
1454        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1455        as the output vector has at least 2 elements. The offset of the start of
1456        the failing character and a reason code are placed in the vector.
1457    
1458    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1459        now returned is for the first byte of the failing character, instead of the
1460        last byte inspected. This is an incompatible change, but I hope it is small
1461        enough not to be a problem. It makes the returned offset consistent with
1462        pcre_exec() and pcre_dfa_exec().
1463    
1464    7.  pcretest now gives a text phrase as well as the error number when
1465        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1466        failure, the offset and reason code are output.
1467    
1468    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1469        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1470        back over a single character (\n). This seems wrong (because it treated the
1471        two characters as a single entity when going forwards), conflicts with the
1472        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1473        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1474        has been changed.
1475    
1476    9.  Some internal refactoring has changed the processing so that the handling
1477        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1478        time (the PCRE_DOTALL option was changed this way some time ago: version
1479        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1480        which was always a bit of a fudge. It also means that there is one less
1481        argument for the match() function, which reduces its stack requirements
1482        slightly. This change also fixes an incompatibility with Perl: the pattern
1483        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1484    
1485    10. More internal refactoring has drastically reduced the number of recursive
1486        calls to match() for possessively repeated groups such as (abc)++ when
1487        using pcre_exec().
1488    
1489    11. While implementing 10, a number of bugs in the handling of groups were
1490        discovered and fixed:
1491    
1492        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1493        (a|)*(?1) gave a compile-time internal error.
1494        ((a|)+)+  did not notice that the outer group could match an empty string.
1495        (^a|^)+   was not marked as anchored.
1496        (.*a|.*)+ was not marked as matching at start or after a newline.
1497    
1498    12. Yet more internal refactoring has removed another argument from the match()
1499        function. Special calls to this function are now indicated by setting a
1500        value in a variable in the "match data" data block.
1501    
1502    13. Be more explicit in pcre_study() instead of relying on "default" for
1503        opcodes that mean there is no starting character; this means that when new
1504        ones are added and accidentally left out of pcre_study(), testing should
1505        pick them up.
1506    
1507    14. The -s option of pcretest has been documented for ages as being an old
1508        synonym of -m (show memory usage). I have changed it to mean "force study
1509        for every regex", that is, assume /S for every regex. This is similar to -i
1510        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1511        using it. It makes it easier to run collections of tests with and without
1512        study enabled, and thereby test pcre_study() more easily. All the standard
1513        tests are now run with and without -s (but some patterns can be marked as
1514        "never study" - see 20 below).
1515    
1516    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1517        restoration of the capturing data to the outer values was not happening
1518        correctly.
1519    
1520    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1521        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1522        pattern had matched an empty string, and so incorrectly returned a no
1523        match.
1524    
1525    17. There was optimizing code for the last branch of non-capturing parentheses,
1526        and also for the obeyed branch of a conditional subexpression, which used
1527        tail recursion to cut down on stack usage. Unfortunately, now that there is
1528        the possibility of (*THEN) occurring in these branches, tail recursion is
1529        no longer possible because the return has to be checked for (*THEN). These
1530        two optimizations have therefore been removed. [But see 8.20/11 above.]
1531    
1532    18. If a pattern containing \R was studied, it was assumed that \R always
1533        matched two bytes, thus causing the minimum subject length to be
1534        incorrectly computed because \R can also match just one byte.
1535    
1536    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1537        was incorrectly computed.
1538    
1539    20. If /S is present twice on a test pattern in pcretest input, it now
1540        *disables* studying, thereby overriding the use of -s on the command line
1541        (see 14 above). This is necessary for one or two tests to keep the output
1542        identical in both cases.
1543    
1544    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1545        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1546    
1547    22. When an atomic group that contained a capturing parenthesis was
1548        successfully matched, but the branch in which it appeared failed, the
1549        capturing was not being forgotten if a higher numbered group was later
1550        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1551        group 1 to "a", when in fact it should be unset. This applied to multi-
1552        branched capturing and non-capturing groups, repeated or not, and also to
1553        positive assertions (capturing in negative assertions does not happen
1554        in PCRE) and also to nested atomic groups.
1555    
1556    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1557        subject after a captured substring, to make it easier to tell which of a
1558        number of identical substrings has been captured.
1559    
1560    24. The way atomic groups are processed by pcre_exec() has been changed so that
1561        if they are repeated, backtracking one repetition now resets captured
1562        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1563        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1564        "aaa". Previously, it would have been "a". As part of this code
1565        refactoring, the way recursive calls are handled has also been changed.
1566    
1567    25. If an assertion condition captured any substrings, they were not passed
1568        back unless some other capturing happened later. For example, if
1569        (?(?=(a))a) was matched against "a", no capturing was returned.
1570    
1571    26. When studying a pattern that contained subroutine calls or assertions,
1572        the code for finding the minimum length of a possible match was handling
1573        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1574        group 1 called group 2 while simultaneously a separate group 2 called group
1575        1). A stack overflow occurred in this case. I have fixed this by limiting
1576        the recursion depth to 10.
1577    
1578    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1579        Fortmann. This supports explicit test numbers on the command line, and has
1580        argument validation and error reporting.
1581    
1582    28. An instance of \X with an unlimited repeat could fail if at any point the
1583        first character it looked at was a mark character.
1584    
1585    29. Some minor code refactoring concerning Unicode properties and scripts
1586        should reduce the stack requirement of match() slightly.
1587    
1588    30. Added the '=' option to pcretest to check the setting of unused capturing
1589        slots at the end of the pattern, which are documented as being -1, but are
1590        not included in the return count.
1591    
1592    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1593        compiled something random. Now it gives a compile-time error (as does
1594        Perl).
1595    
1596    32. A *MARK encountered during the processing of a positive assertion is now
1597        recorded and passed back (compatible with Perl).
1598    
1599    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1600        had alternative anchored branches, the search for a second match in a line
1601        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1602        matched the line "0102" twice. The same bug affected patterns that started
1603        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1604        twice.
1605    
1606    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1607        does, and because of capturing effects, quantifying parenthesized
1608        assertions may at times be useful. Quantifiers are now allowed for
1609        parenthesized assertions.
1610    
1611    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1612    
1613    36. \g was being checked for fancy things in a character class, when it should
1614        just be a literal "g".
1615    
1616    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1617        appearance of a nested POSIX class supersedes an apparent external class.
1618        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1619        unescaped square brackets may also appear as part of class names. For
1620        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1621        more like Perl. (But see 8.20/1 above.)
1622    
1623    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1624        was because it thought it was \N{name}, which is not supported).
1625    
1626    39. Add minix to OS list not supporting the -S option in pcretest.
1627    
1628    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1629        cannot analyze patterns in sufficient detail to catch mutual recursions
1630        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1631        subgroup is called recursively as a subpattern for a second time at the
1632        same position in the subject string. In previous releases this might have
1633        been caught by the recursion limit, or it might have run out of stack.
1634    
1635    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1636        happen only once. PCRE was, however incorrectly giving a compile time error
1637        "recursive call could loop indefinitely" because it cannot analyze the
1638        pattern in sufficient detail. The compile time test no longer happens when
1639        PCRE is compiling a conditional subpattern, but actual runaway loops are
1640        now caught at runtime (see 40 above).
1641    
1642    42. It seems that Perl allows any characters other than a closing parenthesis
1643        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1644        has been changed to be the same.
1645    
1646    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1647        as not to get warnings when autogen.sh is called. Also changed
1648        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1649    
1650    44. To help people who use pcregrep to scan files containing exceedingly long
1651        lines, the following changes have been made:
1652    
1653        (a) The default value of the buffer size parameter has been increased from
1654            8K to 20K. (The actual buffer used is three times this size.)
1655    
1656        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1657            PCRE is built.
1658    
1659        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1660            to be set at run time.
1661    
1662        (d) Numerical values in pcregrep options can be followed by K or M, for
1663            example --buffer-size=50K.
1664    
1665        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1666            given and the return code is set to 2.
1667    
1668    45. Add a pointer to the latest mark to the callout data block.
1669    
1670    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1671        partial match of an empty string instead of no match. This was specific to
1672        the use of ".".
1673    
1674    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1675        complete match instead of a partial match. This bug was dependent on both
1676        the PCRE_UTF8 and PCRE_DOTALL options being set.
1677    
1678    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1679        starting byte set, because \b was not being ignored.
1680    
1681    
1682    Version 8.12 15-Jan-2011
1683    ------------------------
1684    
1685    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1686        checks for such things as part of the documentation building process.
1687    
1688    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1689        --match-limit and --recursion-limit options (added for 8.11). In
1690        particular, this made one of the standard tests fail. (The integer value
1691        went into the wrong half of a long int.)
1692    
1693    3.  If the --colour option was given to pcregrep with -v (invert match), it
1694        did strange things, either producing crazy output, or crashing. It should,
1695        of course, ignore a request for colour when reporting lines that do not
1696        match.
1697    
1698    4.  Another pcregrep bug caused similar problems if --colour was specified with
1699        -M (multiline) and the pattern match finished with a line ending.
1700    
1701    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1702        matched in multiline mode, the following line was shown as part of the
1703        match. This seems wrong, so I have changed it.
1704    
1705    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1706        the check for further matches in the same line (so they could be coloured)
1707        to overrun the end of the current line. If another match was found, it was
1708        incorrectly shown (and then shown again when found in the next line).
1709    
1710    7.  If pcregrep was compiled under Windows, there was a reference to the
1711        function pcregrep_exit() before it was defined. I am assuming this was
1712        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1713        reported by a user. I've moved the definition above the reference.
1714    
1715    
1716    Version 8.11 10-Dec-2010
1717    ------------------------
1718    
1719    1.  (*THEN) was not working properly if there were untried alternatives prior
1720        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1721        backtracked to try for "b" instead of moving to the next alternative branch
1722        at the same level (in this case, to look for "c"). The Perl documentation
1723        is clear that when (*THEN) is backtracked onto, it goes to the "next
1724        alternative in the innermost enclosing group".
1725    
1726    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1727        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1728        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1729        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1730        (*THEN).
1731    
1732    3.  If \s appeared in a character class, it removed the VT character from
1733        the class, even if it had been included by some previous item, for example
1734        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1735        of \s, but is part of the POSIX "space" class.)
1736    
1737    4.  A partial match never returns an empty string (because you can always
1738        match an empty string at the end of the subject); however the checking for
1739        an empty string was starting at the "start of match" point. This has been
1740        changed to the "earliest inspected character" point, because the returned
1741        data for a partial match starts at this character. This means that, for
1742        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1743        (previously it gave "no match").
1744    
1745    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1746        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1747        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1748        has an implication that the given string is incomplete (because a partial
1749        match is preferred over a full match). For this reason, these items now
1750        give a partial match in this situation. [Aside: previously, the one case
1751        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1752        match rather than a full match, which was wrong by the old rules, but is
1753        now correct.]
1754    
1755    6.  There was a bug in the handling of #-introduced comments, recognized when
1756        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1757        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1758        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1759        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1760        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1761        places in pcre_compile().
1762    
1763    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1764        comments when looking ahead for named forward references to subpatterns,
1765        the only newline sequence it recognized was NL. It now handles newlines
1766        according to the set newline convention.
1767    
1768    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1769        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1770        cater for a lack of strerror(). These oversights have been fixed.
1771    
1772    9.  Added --match-limit and --recursion-limit to pcregrep.
1773    
1774    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1775    
1776    11. When the -o option was used, pcregrep was setting a return code of 1, even
1777        when matches were found, and --line-buffered was not being honoured.
1778    
1779    12. Added an optional parentheses number to the -o and --only-matching options
1780        of pcregrep.
1781    
1782    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1783        can match an empty string. The code to do it in pcretest and pcredemo
1784        needed fixing:
1785    
1786        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1787            only one byte after an empty string match just before CRLF (this case
1788            just got forgotten; "any" and "anycrlf" were OK).
1789    
1790        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1791            mode when an empty string match preceded an ASCII character followed by
1792            a non-ASCII character. (The code for advancing by one character rather
1793            than one byte was nonsense.)
1794    
1795        (c) The pcredemo.c sample program did not have any code at all to handle
1796            the cases when CRLF is a valid newline sequence.
1797    
1798    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1799        as a starting offset was within the subject string. There is now a new
1800        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1801        negative or greater than the length of the string. In order to test this,
1802        pcretest is extended to allow the setting of negative starting offsets.
1803    
1804    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1805        starting offset points to the beginning of a UTF-8 character was
1806        unnecessarily clumsy. I tidied it up.
1807    
1808    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1809        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1810    
1811    17. Nobody had reported that the --include_dir option, which was added in
1812        release 7.7 should have been called --include-dir (hyphen, not underscore)
1813        for compatibility with GNU grep. I have changed it to --include-dir, but
1814        left --include_dir as an undocumented synonym, and the same for
1815        --exclude-dir, though that is not available in GNU grep, at least as of
1816        release 2.5.4.
1817    
1818    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1819        characters from a string of bytes) have been redefined so as not to use
1820        loops, in order to improve performance in some environments. At the same
1821        time, I abstracted some of the common code into auxiliary macros to save
1822        repetition (this should not affect the compiled code).
1823    
1824    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1825        compile-time error is now given if \c is not followed by an ASCII
1826        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1827        different, and any byte value is allowed.)
1828    
1829    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1830        START_OPTIMIZE option, which is now allowed at compile time - but just
1831        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1832        to pcregrep and other applications that have no direct access to PCRE
1833        options. The new /Y option in pcretest sets this option when calling
1834        pcre_compile().
1835    
1836    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1837        back references. Groups containing recursive back references were forced to
1838        be atomic by that change, but in the case of named groups, the amount of
1839        memory required was incorrectly computed, leading to "Failed: internal
1840        error: code overflow". This has been fixed.
1841    
1842    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1843        pcretest.c, to avoid build problems in some Borland environments.
1844    
1845    
1846    Version 8.10 25-Jun-2010
1847    ------------------------
1848    
1849    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1850        THEN.
1851    
1852    2.  (*ACCEPT) was not working when inside an atomic group.
1853    
1854    3.  Inside a character class, \B is treated as a literal by default, but
1855        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1856        causes the error). The code is unchanged, but I tidied the documentation.
1857    
1858    4.  Inside a character class, PCRE always treated \R and \X as literals,
1859        whereas Perl faults them if its -w option is set. I have changed PCRE so
1860        that it faults them when PCRE_EXTRA is set.
1861    
1862    5.  Added support for \N, which always matches any character other than
1863        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1864    
1865    6.  When compiling pcregrep with newer versions of gcc which may have
1866        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1867        declared with attribute warn_unused_result" were given. Just casting the
1868        result to (void) does not stop the warnings; a more elaborate fudge is
1869        needed. I've used a macro to implement this.
1870    
1871    7.  Minor change to pcretest.c to avoid a compiler warning.
1872    
1873    8.  Added four artifical Unicode properties to help with an option to make
1874        \s etc use properties (see next item). The new properties are: Xan
1875        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1876    
1877    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1878        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1879        this option. Modified pcretest to add /W to test this facility. Added
1880        REG_UCP to make it available via the POSIX interface.
1881    
1882    10. Added --line-buffered to pcregrep.
1883    
1884    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1885        studied, and the match started with a letter with a code point greater than
1886        127 whose first byte was different to the first byte of the other case of
1887        the letter, the other case of this starting letter was not recognized
1888        (#976).
1889    
1890    12. If a pattern that was studied started with a repeated Unicode property
1891        test, for example, \p{Nd}+, there was the theoretical possibility of
1892        setting up an incorrect bitmap of starting bytes, but fortunately it could
1893        not have actually happened in practice until change 8 above was made (it
1894        added property types that matched character-matching opcodes).
1895    
1896    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1897        possible starting bytes for non-anchored patterns.
1898    
1899    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1900        \R, and also a number of cases that involve Unicode properties, both
1901        explicit and implicit when PCRE_UCP is set.
1902    
1903    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1904        input, it could crash or give wrong results if characters with values
1905        greater than 0xc0 were present in the subject string. (Detail: it assumed
1906        UTF-8 input when processing these items.)
1907    
1908    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1909        size_t is 64-bit (#991).
1910    
1911    17. Added a check for running out of memory when PCRE is compiled with
1912        --disable-stack-for-recursion (#990).
1913    
1914    18. If the last data line in a file for pcretest does not have a newline on
1915        the end, a newline was missing in the output.
1916    
1917    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1918        less than 128) in its various bitmaps. However, there is a facility for
1919        generating tables according to the current locale when PCRE is compiled. It
1920        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1921        characters, are recognized by isspace() and therefore were getting set in
1922        these tables, and indeed these tables seem to approximate to ISO 8859. This
1923        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1924        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1925        which of course cannot start UTF-8 characters. I have changed the code so
1926        that only real ASCII characters (less than 128) and the correct starting
1927        bytes for UTF-8 encodings are set for characters greater than 127 when in
1928        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1929        altogether.)
1930    
1931    20. Added the /T option to pcretest so as to be able to run tests with non-
1932        standard character tables, thus making it possible to include the tests
1933        used for 19 above in the standard set of tests.
1934    
1935    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1936        reference to a subpattern the other side of a comment that contains an
1937        opening parenthesis caused either an internal compiling error, or a
1938        reference to the wrong subpattern.
1939    
1940    
1941    Version 8.02 19-Mar-2010
1942    ------------------------
1943    
1944    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1945    
1946    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1947        configured.
1948    
1949    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1950        original author of that file, following a query about its status.
1951    
1952    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1953        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1954    
1955    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1956        quantifier applied to a forward-referencing subroutine call, could compile
1957        incorrect code or give the error "internal error: previously-checked
1958        referenced subpattern not found".
1959    
1960    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1961        variables to point to external functions. For these systems, therefore,
1962        pcre_malloc etc. are now initialized to local functions that call the
1963        relevant global functions.
1964    
1965    7.  There were two entries missing in the vectors called coptable and poptable
1966        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1967        I've fixed the data, and added a kludgy way of testing at compile time that
1968        the lengths are correct (equal to the number of opcodes).
1969    
1970    8.  Following on from 7, I added a similar kludge to check the length of the
1971        eint vector in pcreposix.c.
1972    
1973    9.  Error texts for pcre_compile() are held as one long string to avoid too
1974        much relocation at load time. To find a text, the string is searched,
1975        counting zeros. There was no check for running off the end of the string,
1976        which could happen if a new error number was added without updating the
1977        string.
1978    
1979    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1980    
1981    11. \K was not working if it appeared in an atomic group or in a group that
1982        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1983        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1984        the assertion is positive, but not if it is negative.
1985    
1986    12. Change 11 fortuitously reduced the size of the stack frame used in the
1987        "match()" function of pcre_exec.c by one pointer. Forthcoming
1988        implementation of support for (*MARK) will need an extra pointer on the
1989        stack; I have reserved it now, so that the stack frame size does not
1990        decrease.
1991    
1992    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1993        item in branch that calls a recursion is a subroutine call - as in the
1994        second branch in the above example - was incorrectly given the compile-
1995        time error "recursive call could loop indefinitely" because pcre_compile()
1996        was not correctly checking the subroutine for matching a non-empty string.
1997    
1998    14. The checks for overrunning compiling workspace could trigger after an
1999        overrun had occurred. This is a "should never occur" error, but it can be
2000        triggered by pathological patterns such as hundreds of nested parentheses.
2001        The checks now trigger 100 bytes before the end of the workspace.
2002    
2003    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2004    
2005    
2006    Version 8.01 19-Jan-2010
2007    ------------------------
2008    
2009    1.  If a pattern contained a conditional subpattern with only one branch (in
2010        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2011        computed the wrong minimum data length (which is of course zero for such
2012        subpatterns). This could cause incorrect "no match" results.
2013    
2014    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2015        the pattern is reset in the first branch, pcre_compile() failed with
2016        "internal error: code overflow at offset...". This happened only when
2017        the reset was to the original external option setting. (An optimization
2018        abstracts leading options settings into an external setting, which was the
2019        cause of this.)
2020    
2021    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2022        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2023        assertion pattern did not match (meaning that the assertion was true), it
2024        was incorrectly treated as false if the SKIP had been reached during the
2025        matching. This also applied to assertions used as conditions.
2026    
2027    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2028        assertion subpattern, including such a pattern used as a condition,
2029        unpredictable results occurred, instead of the error return
2030        PCRE_ERROR_DFA_UITEM.
2031    
2032    5.  The C++ GlobalReplace function was not working like Perl for the special
2033        situation when an empty string is matched. It now does the fancy magic
2034        stuff that is necessary.
2035    
2036    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2037        removed. (These were left over from very, very early versions of PCRE.)
2038    
2039    7.  Some cosmetic changes to the code to make life easier when compiling it
2040        as part of something else:
2041    
2042        (a) Change DEBUG to PCRE_DEBUG.
2043    
2044        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2045            called "current" as "current_branch", to prevent a collision with the
2046            Linux macro when compiled as a kernel module.
2047    
2048        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2049            prevent a collision with the Linux macro when compiled as a kernel
2050            module.
2051    
2052    8.  In pcre_compile() there are some checks for integer overflows that used to
2053        cast potentially large values to (double). This has been changed to that
2054        when building, a check for int64_t is made, and if it is found, it is used
2055        instead, thus avoiding the use of floating point arithmetic. (There is no
2056        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2057        double.
2058    
2059    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2060        2005 (difference between two addresses compared to an unsigned value).
2061    
2062    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2063        custom one, because of the following reported problem in Windows:
2064    
2065          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2066              under Win32.
2067          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2068              therefore missing the function definition.
2069          - The compiler thus generates a "C" signature for the test function.
2070          - The linker fails to find the "C" function.
2071          - PCRE fails to configure if asked to do so against libbz2.
2072    
2073    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2074        messages were output:
2075    
2076          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2077          rerunning libtoolize, to keep the correct libtool macros in-tree.
2078          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2079    
2080        I have done both of these things.
2081    
2082    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2083        most of the time, it *can* run out if it is given a pattern that contains a
2084        runaway infinite recursion. I updated the discussion in the pcrestack man
2085        page.
2086    
2087    13. Now that we have gone to the x.xx style of version numbers, the minor
2088        version may start with zero. Using 08 or 09 is a bad idea because users
2089        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2090        interpreted as invalid octal numbers. I've updated the previous comment in
2091        configure.ac, and also added a check that gives an error if 08 or 09 are
2092        used.
2093    
2094    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2095        causing partial matching to fail when the end of the subject matched \W
2096        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2097    
2098    15. There were some discrepancies between the declarations in pcre_internal.h
2099        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2100        their definitions. The declarations used "const uschar *" and the
2101        definitions used USPTR. Even though USPTR is normally defined as "const
2102        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2103        reported that: "This difference in casting confuses some C++ compilers, for
2104        example, SunCC recognizes above declarations as different functions and
2105        generates broken code for hbpcre." I have changed the declarations to use
2106        USPTR.
2107    
2108    16. GNU libtool is named differently on some systems. The autogen.sh script now
2109        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2110        (FreeBSD).
2111    
2112    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2113        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2114        comment: "Figure out how to create a longlong from a string: strtoll and
2115        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2116        instance, but it only takes 2 args instead of 3!"
2117    
2118    18. A subtle bug concerned with back references has been fixed by a change of
2119        specification, with a corresponding code fix. A pattern such as
2120        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2121        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2122        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2123        same bug. Such groups have to be quantified to be useful, or contained
2124        inside another quantified group. (If there's no repetition, the reference
2125        can never match.) The problem arises because, having left the group and
2126        moved on to the rest of the pattern, a later failure that backtracks into
2127        the group uses the captured value from the final iteration of the group
2128        rather than the correct earlier one. I have fixed this in PCRE by forcing
2129        any group that contains a reference to itself to be an atomic group; that
2130        is, there cannot be any backtracking into it once it has completed. This is
2131        similar to recursive and subroutine calls.
2132    
2133    
2134    Version 8.00 19-Oct-09
2135    ----------------------
2136    
2137    1.  The table for translating pcre_compile() error codes into POSIX error codes
2138        was out-of-date, and there was no check on the pcre_compile() error code
2139        being within the table. This could lead to an OK return being given in
2140        error.
2141    
2142    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2143        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2144        in a Windows environment.
2145    
2146    3.  The pcregrep --count option prints the count for each file even when it is
2147        zero, as does GNU grep. However, pcregrep was also printing all files when
2148        --files-with-matches was added. Now, when both options are given, it prints
2149        counts only for those files that have at least one match. (GNU grep just
2150        prints the file name in this circumstance, but including the count seems
2151        more useful - otherwise, why use --count?) Also ensured that the
2152        combination -clh just lists non-zero counts, with no names.
2153    
2154    4.  The long form of the pcregrep -F option was incorrectly implemented as
2155        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2156        but it seems right to fix it, and I didn't think it was worth preserving
2157        the old behaviour.
2158    
2159    5.  The command line items --regex=pattern and --regexp=pattern were not
2160        recognized by pcregrep, which required --regex pattern or --regexp pattern
2161        (with a space rather than an '='). The man page documented the '=' forms,
2162        which are compatible with GNU grep; these now work.
2163    
2164    6.  No libpcreposix.pc file was created for pkg-config; there was just
2165        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2166    
2167    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2168        when UCP support is not needed, by modifying the Python script that
2169        generates it from Unicode data files. This should not matter if the module
2170        is correctly used as a library, but I received one complaint about 50K of
2171        unwanted data. My guess is that the person linked everything into his
2172        program rather than using a library. Anyway, it does no harm.
2173    
2174    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2175        was a minimum greater than 1 for a wide character in a possessive
2176        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2177        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2178        character. Chaos in the form of incorrect output or a compiling loop could
2179        result.
2180    
2181    9.  The restrictions on what a pattern can contain when partial matching is
2182        requested for pcre_exec() have been removed. All patterns can now be
2183        partially matched by this function. In addition, if there are at least two
2184        slots in the offset vector, the offset of the earliest inspected character
2185        for the match and the offset of the end of the subject are set in them when
2186        PCRE_ERROR_PARTIAL is returned.
2187    
2188    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2189        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2190        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2191        and may be more useful for multi-segment matching.
2192    
2193    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2194        used to be given if ever the end of the subject was reached; now it is
2195        given only if matching could not proceed because another character was
2196        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2197        string "Z", which now yields "no match" instead of "partial match". In the
2198        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2199        final character ended with (*FAIL).
2200    
2201    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2202        if the pattern had a "must contain" character that was already found in the
2203        earlier partial match, unless partial matching was again requested. For
2204        example, with the pattern /dog.(body)?/, the "must contain" character is
2205        "g". If the first part-match was for the string "dog", restarting with
2206        "sbody" failed. This bug has been fixed.
2207    
2208    13. The string returned by pcre_dfa_exec() after a partial match has been
2209        changed so that it starts at the first inspected character rather than the
2210        first character of the match. This makes a difference only if the pattern
2211        starts with a lookbehind assertion or \b or \B (\K is not supported by
2212        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2213        matching functions compatible, and I think it's the right thing to do.
2214    
2215    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2216        so that the demonstration program is easily available in environments where
2217        PCRE has not been installed from source.
2218    
2219    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2220        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2221        library.
2222    
2223    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2224        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2225        is not the first non-POSIX option to be added. Clearly some people find
2226        these options useful.
2227    
2228    17. If a caller to the POSIX matching function regexec() passes a non-zero
2229        value for nmatch with a NULL value for pmatch, the value of
2230        nmatch is forced to zero.
2231    
2232    18. RunGrepTest did not have a test for the availability of the -u option of
2233        the diff command, as RunTest does. It now checks in the same way as
2234        RunTest, and also checks for the -b option.
2235    
2236    19. If an odd number of negated classes containing just a single character
2237        interposed, within parentheses, between a forward reference to a named
2238        subpattern and the definition of the subpattern, compilation crashed with
2239        an internal error, complaining that it could not find the referenced
2240        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2241        [The bug was that it was starting one character too far in when skipping
2242        over the character class, thus treating the ] as data rather than
2243        terminating the class. This meant it could skip too much.]
2244    
2245    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2246        /g option in pcretest when the pattern contains \K, which makes it possible
2247        to have an empty string match not at the start, even when the pattern is
2248        anchored. Updated pcretest and pcredemo to use this option.
2249    
2250    21. If the maximum number of capturing subpatterns in a recursion was greater
2251        than the maximum at the outer level, the higher number was returned, but
2252        with unset values at the outer level. The correct (outer level) value is
2253        now given.
2254    
2255    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2256        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2257        make it do so. The string so far is captured, making this feature
2258        compatible with Perl.
2259    
2260    23. The tests have been re-organized, adding tests 11 and 12, to make it
2261        possible to check the Perl 5.10 features against Perl 5.10.
2262    
2263    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2264        pattern matches a fixed length string. PCRE did not allow this; now it
2265        does. Neither allows recursion.
2266    
2267    25. I finally figured out how to implement a request to provide the minimum
2268        length of subject string that was needed in order to match a given pattern.
2269        (It was back references and recursion that I had previously got hung up
2270        on.) This code has now been added to pcre_study(); it finds a lower bound
2271        to the length of subject needed. It is not necessarily the greatest lower
2272        bound, but using it to avoid searching strings that are too short does give
2273        some useful speed-ups. The value is available to calling programs via
2274        pcre_fullinfo().
2275    
2276    26. While implementing 25, I discovered to my embarrassment that pcretest had
2277        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2278        study optimizations had never been tested with that matching function.
2279        Oops. What is worse, even when it was passed study data, there was a bug in
2280        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2281        were also very few tests of studied patterns with pcre_dfa_exec().
2282    
2283    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2284        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2285        on the other side of the coin, they are no longer allowed to have different
2286        names, because these cannot be distinguished in PCRE, and this has caused
2287        confusion. (This is a difference from Perl.)
2288    
2289    28. When duplicate subpattern names are present (necessarily with different
2290        numbers, as required by 27 above), and a test is made by name in a
2291        conditional pattern, either for a subpattern having been matched, or for
2292        recursion in such a pattern, all the associated numbered subpatterns are
2293        tested, and the overall condition is true if the condition is true for any
2294        one of them. This is the way Perl works, and is also more like the way
2295        testing by number works.
2296    
2297    
2298    Version 7.9 11-Apr-09
2299    ---------------------
2300    
2301    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2302        (pcretest), all targets were linked against these libraries. This included
2303        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2304        libraries. This caused unwanted dependencies to be created. This problem
2305        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2306        pcretest is linked with readline.
2307    
2308    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2309        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2310        moved outside it again, because FALSE and TRUE are already defined in AIX,
2311        but BOOL is not.
2312    
2313    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2314        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2315    
2316    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2317        hyphens) following file names and line numbers when outputting matching
2318        lines. This is not true; no spaces are inserted. I have also clarified the
2319        wording for the --colour (or --color) option.
2320    
2321    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2322        was not coloured; this is different to GNU grep, so I have changed it to be
2323        the same.
2324    
2325    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2326        each matching line was coloured. Now it goes on to look for further matches
2327        of any of the test patterns, which is the same behaviour as GNU grep.
2328    
2329    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2330        doesn't make sense to accept an empty string match in pcregrep, so I have
2331        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2332        seems to be how GNU grep behaves. [But see later change 40 for release
2333        8.33.]
2334    
2335    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2336        start or after a newline", because the conditional assertion was not being
2337        correctly handled. The rule now is that both the assertion and what follows
2338        in the first alternative must satisfy the test.
2339    
2340    9.  If auto-callout was enabled in a pattern with a conditional group whose
2341        condition was an assertion, PCRE could crash during matching, both with
2342        pcre_exec() and pcre_dfa_exec().
2343    
2344    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2345        used for matching.
2346    
2347    11. Unicode property support in character classes was not working for
2348        characters (bytes) greater than 127 when not in UTF-8 mode.
2349    
2350    12. Added the -M command line option to pcretest.
2351    
2352    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2353    
2354    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2355    
2356    16. Added comments and documentation about mis-use of no_arg in the C++
2357        wrapper.
2358    
2359    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2360        from Martin Jerabek that uses macro names for all relevant character and
2361        string constants.
2362    
2363    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2364        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2365        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2366        these, but not everybody uses configure.
2367    
2368    19. A conditional group that had only one branch was not being correctly
2369        recognized as an item that could match an empty string. This meant that an
2370        enclosing group might also not be so recognized, causing infinite looping
2371        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2372        with the subject "ab", where knowledge that the repeated group can match
2373        nothing is needed in order to break the loop.
2374    
2375    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2376        exec(), but without supplying a callout function, matching went wrong.
2377    
2378    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2379        leak if the size of the offset vector was greater than 30. When the vector
2380        is smaller, the saved offsets during recursion go onto a local stack
2381        vector, but for larger vectors malloc() is used. It was failing to free
2382        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2383        error, in fact).
2384    
2385    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2386        heapframe that is used only when UTF-8 support is enabled. This caused no
2387        problem, but was untidy.
2388    
2389    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2390        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2391        included within another project.
2392    
2393    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2394        slightly modified by me:
2395    
2396          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2397              not building pcregrep.
2398    
2399          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2400              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2401    
2402    25. Forward references, both numeric and by name, in patterns that made use of
2403        duplicate group numbers, could behave incorrectly or give incorrect errors,
2404        because when scanning forward to find the reference group, PCRE was not
2405        taking into account the duplicate group numbers. A pattern such as
2406        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2407    
2408    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2409        the feature of a custom pointer more persuasive (as requested by a user).
2410    
2411    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2412        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2413        pre-defined.
2414    
2415    28. Added support for (*UTF8) at the start of a pattern.
2416    
2417    29. Arrange for flags added by the "release type" setting in CMake to be shown
2418        in the configuration summary.
2419    
2420    
2421    Version 7.8 05-Sep-08
2422    ---------------------
2423    
2424    1.  Replaced UCP searching code with optimized version as implemented for Ad
2425        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2426        stage table and inline lookup instead of a function, giving speed ups of 2
2427        to 5 times on some simple patterns that I tested. Permission was given to
2428        distribute the MultiStage2.py script that generates the tables (it's not in
2429        the tarball, but is in the Subversion repository).
2430    
2431    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2432        scripts.
2433    
2434    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2435        a group with a zero qualifier. The result of the study could be incorrect,
2436        or the function might crash, depending on the pattern.
2437    
2438    4.  Caseless matching was not working for non-ASCII characters in back
2439        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2440        It now works when Unicode Property Support is available.
2441    
2442    5.  In pcretest, an escape such as \x{de} in the data was always generating
2443        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2444        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2445        truncation.
2446    
2447    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2448    
2449    7.  Added two (int) casts to pcregrep when printing the difference of two
2450        pointers, in case they are 64-bit values.
2451    
2452    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2453        test 2 if it fails.
2454    
2455    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2456        and a #define of that name to empty if it is not externally set. This is to
2457        allow users of MSVC to set it if necessary.
2458    
2459    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2460        the convenience functions in the pcre_get.c source file.
2461    
2462    11. An option change at the start of a pattern that had top-level alternatives
2463        could cause overwriting and/or a crash. This command provoked a crash in
2464        some environments:
2465    
2466          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2467    
2468        This potential security problem was recorded as CVE-2008-2371.
2469    
2470    12. For a pattern where the match had to start at the beginning or immediately
2471        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2472        pcre_dfa_exec() could read past the end of the passed subject if there was
2473        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2474        pcretest so that it places the subject at the end of its malloc-ed buffer.
2475    
2476    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2477        exec() might read past the end of the data buffer in UTF-8 mode.
2478    
2479    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2480        the data contained the byte 0x85 as part of a UTF-8 character within its
2481        first line. This applied both to normal and DFA matching.
2482    
2483    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2484        /^[^d]*?$/8 failed to match "abc".
2485    
2486    16. Added a missing copyright notice to pcrecpp_internal.h.
2487    
2488    17. Make it more clear in the documentation that values returned from
2489        pcre_exec() in ovector are byte offsets, not character counts.
2490    
2491    18. Tidied a few places to stop certain compilers from issuing warnings.
2492    
2493    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2494        supplied by Stefan Weber. I made a further small update for 7.8 because
2495        there is a change of source arrangements: the pcre_searchfuncs.c module is
2496        replaced by pcre_ucd.c.
2497    
2498    
2499    Version 7.7 07-May-08
2500    ---------------------
2501    
2502    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2503        a string to a long long, pretend we don't even have a long long." This is
2504        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2505    
2506    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2507        pre-7.6 versions, which defined a global no_arg variable instead of putting
2508        it in the RE class. (See also #8 below.)
2509    
2510    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2511        Lopes.
2512    
2513    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2514    
2515        (1) The include/exclude patterns were being applied to the whole pathnames
2516            of files, instead of just to the final components.
2517    
2518        (2) If there was more than one level of directory, the subdirectories were
2519            skipped unless they satisfied the include/exclude conditions. This is
2520            inconsistent with GNU grep (and could even be seen as contrary to the
2521            pcregrep specification - which I improved to make it absolutely clear).
2522            The action now is always to scan all levels of directory, and just
2523            apply the include/exclude patterns to regular files.
2524    
2525    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2526        --exclude_dir in the tests to avoid scanning .svn directories.
2527    
2528    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2529        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2530        doesn't support NULs in patterns.
2531    
2532    7.  Added some missing "const"s to declarations of static tables in
2533        pcre_compile.c and pcre_dfa_exec.c.
2534    
2535    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2536        caused by fix #2  above. (Subsequently also a second patch to fix the
2537        first patch. And a third patch - this was a messy problem.)
2538    
2539    9.  Applied Craig's patch to remove the use of push_back().
2540    
2541    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2542        matching function regexec().
2543    
2544    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2545        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2546        references. PCRE supports relative numbers with this syntax (I don't think
2547        Oniguruma does).
2548    
2549    12. Previously, a group with a zero repeat such as (...){0} was completely
2550        omitted from the compiled regex. However, this means that if the group
2551        was called as a subroutine from elsewhere in the pattern, things went wrong
2552        (an internal error was given). Such groups are now left in the compiled
2553        pattern, with a new opcode that causes them to be skipped at execution
2554        time.
2555    
2556    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2557        to the way PCRE behaves:
2558    
2559        (a) A lone ] character is dis-allowed (Perl treats it as data).
2560    
2561        (b) A back reference to an unmatched subpattern matches an empty string
2562            (Perl fails the current match path).
2563    
2564        (c) A data ] in a character class must be notated as \] because if the
2565            first data character in a class is ], it defines an empty class. (In
2566            Perl it is not possible to have an empty class.) The empty class []
2567            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2568            The negative empty class [^] matches any one character, independently
2569            of the DOTALL setting.
2570    
2571    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2572        non-existent subpattern following a character class starting with ']' and
2573        containing () gave an internal compiling error instead of "reference to
2574        non-existent subpattern". Fortunately, when the pattern did exist, the
2575        compiled code was correct. (When scanning forwards to check for the
2576        existence of the subpattern, it was treating the data ']' as terminating
2577        the class, so got the count wrong. When actually compiling, the reference
2578        was subsequently set up correctly.)
2579    
2580    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2581        it was being rejected as not supported by pcre_dfa_exec(), even though
2582        other assertions are supported. I have made pcre_dfa_exec() support
2583        (*FAIL).
2584    
2585    16. The implementation of 13c above involved the invention of a new opcode,
2586        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2587        cannot be changed at match time, I realized I could make a small
2588        improvement to matching performance by compiling OP_ALLANY instead of
2589        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2590        on the OP_ANY path.
2591    
2592    17. Compiling pcretest on Windows with readline support failed without the
2593        following two fixes: (1) Make the unistd.h include conditional on
2594        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2595    
2596    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2597        ncurses library to be included for pcretest when ReadLine support is
2598        requested, but also to allow for it to be overridden. This patch came from
2599        Daniel Bergström.
2600    
2601    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2602        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2603        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2604        spotting this.
2605    
2606    
2607    Version 7.6 28-Jan-08
2608    ---------------------
2609    
2610    1.  A character class containing a very large number of characters with
2611        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2612        overflow.
2613    
2614    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2615        HAVE_LONG_LONG is not defined.
2616    
2617    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2618        bring them up to date and include new features. This patch includes:
2619    
2620        - Fixed PH's badly added libz and libbz2 support.
2621        - Fixed a problem with static linking.
2622        - Added pcredemo. [But later removed - see 7 below.]
2623        - Fixed dftables problem and added an option.
2624        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2625            HAVE_LONG_LONG.
2626        - Added readline support for pcretest.
2627        - Added an listing of the option settings after cmake has run.
2628    
2629    4.  A user submitted a patch to Makefile that makes it easy to create
2630        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2631        Makefile.am that cause it to include this special target, without
2632        affecting anything else. Note that the same mingw target plus all
2633        the other distribution libraries and programs are now supported
2634        when configuring with CMake (see 6 below) instead of with
2635        Configure/Make.
2636    
2637    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2638        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2639        exported in the Windows port". It has not yet been confirmed that the patch
2640        solves the problem, but it does no harm.
2641    
2642    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2643        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2644        with CMake, and also correct the comment about stack recursion.
2645    
2646    7.  Remove the automatic building of pcredemo from the ./configure system and
2647        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2648        of a program that users should build themselves after PCRE is installed, so
2649        building it automatically is not really right. What is more, it gave
2650        trouble in some build environments.
2651    
2652    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2653    
2654    
2655    Version 7.5 10-Jan-08
2656    ---------------------
2657    
2658    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2659        values in parens when parsing an RE using the C++ wrapper."
2660    
2661    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2662        Characters greater than 255 were excluded from the class instead of being
2663        included.
2664    
2665    3.  The same bug as (2) above applied to negated POSIX classes such as
2666        [:^space:].
2667    
2668    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2669        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2670        I have changed it.
2671    
2672    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2673        first named subpattern) and a construct such as (?&a) would reference the
2674        first named subpattern whose name started with "a" (in other words, the
2675        length check was missing). Both these problems are fixed. "Subpattern name
2676        expected" is now given for (?&) (a zero-length name), and this patch also
2677        makes it give the same error for \k'' (previously it complained that that
2678        was a reference to a non-existent subpattern).
2679    
2680    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2681        this is right because (?- can be followed by option settings as well as by
2682        digits. I have, however, made the messages clearer.
2683    
2684    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2685        than the number used in the conditional) now cause a compile-time error.
2686        This is actually not compatible with Perl, which accepts such patterns, but
2687        treats the conditional as always being FALSE (as PCRE used to), but it
2688        seems to me that giving a diagnostic is better.
2689    
2690    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2691        and messages.
2692    
2693    9.  Fix two occurrences of "backslash" in comments that should have been
2694        "backspace".
2695    
2696    10. Remove two redundant lines of code that can never be obeyed (their function
2697        was moved elsewhere).
2698    
2699    11. The program that makes PCRE's Unicode character property table had a bug
2700        which caused it to generate incorrect table entries for sequences of
2701        characters that have the same character type, but are in different scripts.
2702        It amalgamated them into a single range, with the script of the first of
2703        them. In other words, some characters were in the wrong script. There were
2704        thirteen such cases, affecting characters in the following ranges:
2705    
2706          U+002b0 - U+002c1
2707          U+0060c - U+0060d
2708          U+0061e - U+00612
2709          U+0064b - U+0065e
2710          U+0074d - U+0076d
2711          U+01800 - U+01805
2712          U+01d00 - U+01d77
2713          U+01d9b - U+01dbf
2714          U+0200b - U+0200f
2715          U+030fc - U+030fe
2716          U+03260 - U+0327f
2717          U+0fb46 - U+0fbb1
2718          U+10450 - U+1049d
2719    
2720    12. The -o option (show only the matching part of a line) for pcregrep was not
2721        compatible with GNU grep in that, if there was more than one match in a
2722        line, it showed only the first of them. It now behaves in the same way as
2723        GNU grep.
2724    
2725    13. If the -o and -v options were combined for pcregrep, it printed a blank
2726        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2727        does the same. The return code can be used to tell if there were any
2728        non-matching lines.
2729    
2730    14. Added --file-offsets and --line-offsets to pcregrep.
2731    
2732    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2733        infinitely looping recursion. The bug was that positive lookaheads were not
2734        being skipped when checking for a possible empty match (negative lookaheads
2735        and both kinds of lookbehind were skipped).
2736    
2737    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2738        inclusion of <windows.h> to before rather than after the definition of
2739        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2740    
2741    17. Specifying a possessive quantifier with a specific limit for a Unicode
2742        character property caused pcre_compile() to compile bad code, which led at
2743        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2744        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2745        caused the error; without that there was no problem.
2746    
2747    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2748    
2749    19. Added --enable-pcretest-libreadline.
2750    
2751    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2752        RE::GlobalReplace(). As a result, the number of replacements returned was
2753        double what it should be. I removed one of the increments, but Craig sent a
2754        later patch that removed the other one (the right fix) and added unit tests
2755        that check the return values (which was not done before).
2756    
2757    21. Several CMake things:
2758    
2759        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2760            the names libpcre and libpcreposix, not just pcre and pcreposix.
2761    
2762        (2) The above change means that pcretest and pcregrep are now correctly
2763            linked with the newly-built libraries, not previously installed ones.
2764    
2765        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2766    
2767    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2768        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2769        UTF-8 newline character). The key issue is that the pattern starts .*;
2770        this means that the match must be either at the beginning, or after a
2771        newline. The bug was in the code for advancing after a failed match and
2772        checking that the new position followed a newline. It was not taking
2773        account of UTF-8 characters correctly.
2774    
2775    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2776        character classes. PCRE was not treating the sequence [:...:] as a
2777        character class unless the ... were all letters. Perl, however, seems to
2778        allow any characters between [: and :], though of course it rejects as
2779        unknown any "names" that contain non-letters, because all the known class
2780        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2781        for example, whereas PCRE did not - it did not recognize a POSIX character
2782        class. This seemed a bit dangerous, so the code has been changed to be
2783        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2784        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2785        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2786        Perl does, and where it didn't before.
2787    
2788    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2789        Windows environments %n is disabled by default.
2790    
2791    
2792    Version 7.4 21-Sep-07
2793    ---------------------
2794    
2795    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2796        means that a class such as [\s] counted as "explicit reference to CR or
2797        LF". That isn't really right - the whole point of the change was to try to
2798        help when there was an actual mention of one of the two characters. So now
2799        the change happens only if \r or \n (or a literal CR or LF) character is
2800        encountered.
2801    
2802    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2803        of both had grown to the point where there were only 3 bits left.
2804        Fortunately, there was spare space in the data structure, and so I have
2805        moved the internal flags into a new 16-bit field to free up more option
2806        bits.
2807    
2808    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2809        but did not set the internal JCHANGED flag - either of these is enough to
2810        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2811        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2812        start sets both bits.
2813    
2814    4.  Added options (at build time, compile time, exec time) to change \R from
2815        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2816    
2817    5.  doc/pcresyntax.html was missing from the distribution.
2818    
2819    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2820        compatibility, even though it is no longer used.
2821    
2822    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2823        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2824        windows.h file is present (where different names are used). [This was
2825        reversed later after testing - see 16 below.]
2826    
2827    8.  Changed all #include <config.h> to #include "config.h". There were also
2828        some further <pcre.h> cases that I changed to "pcre.h".
2829    
2830    9.  When pcregrep was used with the --colour option, it missed the line ending
2831        sequence off the lines that it output.
2832    
2833    10. It was pointed out to me that arrays of string pointers cause lots of
2834        relocations when a shared library is dynamically loaded. A technique of
2835        using a single long string with a table of offsets can drastically reduce
2836        these. I have refactored PCRE in four places to do this. The result is
2837        dramatic:
2838    
2839          Originally:                          290
2840          After changing UCP table:            187
2841          After changing error message table:   43
2842          After changing table of "verbs"       36
2843          After changing table of Posix names   22
2844    
2845        Thanks to the folks working on Gregex for glib for this insight.
2846    
2847    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2848        unicode-properties was also set.
2849    
2850    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2851    
2852    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2853        checked only for CRLF.
2854    
2855    14. Added casts to pcretest.c to avoid compiler warnings.
2856    
2857    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2858    
2859    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2860        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2861        entirely. This removes changes made in 7 above.
2862    
2863    17. The CMake files have been updated, and there is now more information about
2864        building with CMake in the NON-UNIX-USE document.
2865    
2866    
2867    Version 7.3 28-Aug-07
2868    ---------------------
2869    
2870     1. In the rejigging of the build system that eventually resulted in 7.1, the
2871        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2872        brackets there is not right, since it causes compilers to look for an
2873        installed pcre.h, not the version that is in the source that is being
2874        compiled (which of course may be different). I have changed it back to:
2875    
2876          #include "pcre.h"
2877    
2878        I have a vague recollection that the change was concerned with compiling in
2879        different directories, but in the new build system, that is taken care of
2880        by the VPATH setting the Makefile.
2881    
2882     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2883        when the subject happened to end in the byte 0x85 (e.g. if the last
2884        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2885        characters but of course it shouldn't be taken as a newline when it is part
2886        of another character. The bug was that, for an unlimited repeat of . in
2887        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2888        characters when looking for a newline.
2889    
2890     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2891    
2892     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2893        in debug output.
2894    
2895     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2896        long printing in the pcrecpp unittest when running under MinGW.
2897    
2898     6. ESC_K was left out of the EBCDIC table.
2899    
2900     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2901        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2902        limit also applies to "virtual nesting" when a pattern is recursive, and in
2903        this case 1000 isn't so big. I have been able to remove this limit at the
2904        expense of backing off one optimization in certain circumstances. Normally,
2905        when pcre_exec() would call its internal match() function recursively and
2906        immediately return the result unconditionally, it uses a "tail recursion"
2907        feature to save stack. However, when a subpattern that can match an empty
2908        string has an unlimited repetition quantifier, it no longer makes this
2909        optimization. That gives it a stack frame in which to save the data for
2910        checking that an empty string has been matched. Previously this was taken
2911        from the 1000-entry workspace that had been reserved. So now there is no
2912        explicit limit, but more stack is used.
2913    
2914     8. Applied Daniel's patches to solve problems with the import/export magic
2915        syntax that is required for Windows, and which was going wrong for the
2916        pcreposix and pcrecpp parts of the library. These were overlooked when this
2917        problem was solved for the main library.
2918    
2919     9. There were some crude static tests to avoid integer overflow when computing
2920        the size of patterns that contain repeated groups with explicit upper
2921        limits. As the maximum quantifier is 65535, the maximum group length was
2922        set at 30,000 so that the product of these two numbers did not overflow a
2923        32-bit integer. However, it turns out that people want to use groups that
2924        are longer than 30,000 bytes (though not repeat them that many times).
2925        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2926        made it possible to implement the integer overflow checks in a much more
2927        dynamic way, which I have now done. The artificial limitation on group
2928        length has been removed - we now have only the limit on the total length of
2929        the compiled pattern, which depends on the LINK_SIZE setting.
2930    
2931    10. Fixed a bug in the documentation for get/copy named substring when
2932        duplicate names are permitted. If none of the named substrings are set, the
2933        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2934        empty string.
2935    
2936    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2937        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2938        because the ] is interpreted as the first data character and the
2939        terminating ] is not found. PCRE has been made compatible with Perl in this
2940        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2941        cause memory overwriting.
2942    
2943    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2944        string has been matched (to stop an infinite loop). It was not recognizing
2945        a conditional subpattern that could match an empty string if that
2946        subpattern was within another subpattern. For example, it looped when
2947        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2948        condition was not nested. This bug has been fixed.
2949    
2950    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2951        past the start of the subject in the presence of bytes with the top bit
2952        set, for example "\x8aBCD".
2953    
2954    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2955        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2956    
2957    14. Optimized (?!) to (*FAIL).
2958    
2959    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2960        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2961        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2962        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2963        does: it's just the validity check that is more restrictive.
2964    
2965    16. Inserted checks for integer overflows during escape sequence (backslash)
2966        processing, and also fixed erroneous offset values for syntax errors during
2967        backslash processing.
2968    
2969    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2970        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2971    
2972    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2973        caused an overrun.
2974    
2975    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2976        something other than just ASCII characters) inside a group that had an
2977        unlimited repeat caused a loop at compile time (while checking to see
2978        whether the group could match an empty string).
2979    
2980    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2981        [\P{Any}] did so. (Error in the code for printing property names.)
2982    
2983    21. An orphan \E inside a character class could cause a crash.
2984    
2985    22. A repeated capturing bracket such as (A)? could cause a wild memory
2986        reference during compilation.
2987    
2988    23. There are several functions in pcre_compile() that scan along a compiled
2989        expression for various reasons (e.g. to see if it's fixed length for look
2990        behind). There were bugs in these functions when a repeated \p or \P was
2991        present in the pattern. These operators have additional parameters compared
2992        with \d, etc, and these were not being taken into account when moving along
2993        the compiled data. Specifically:
2994    
2995        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2996            length.
2997    
2998        (b) An item such as \pL+ within a repeated group could cause crashes or
2999            loops.
3000    
3001        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3002            "reference to non-existent subpattern" error.
3003    
3004        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3005    
3006    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3007        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3008    
3009    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3010    
3011    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3012        character were causing crashes (broken optimization).
3013    
3014    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3015        \p or \P) caused a compile-time loop.
3016    
3017    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3018        break. For example, the unstudied pattern [\r\n]A does not match the string
3019        "\r\nA" because change 7.0/46 below moves the current point on by two
3020        characters after failing to match at the start. However, the pattern \nA
3021        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3022        the same is true. There doesn't seem any very clean way out of this, but
3023        what I have chosen to do makes the common cases work: PCRE now takes note
3024        of whether there can be an explicit match for \r or \n anywhere in the
3025        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3026        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3027        pattern has explicit CR or LF references.
3028    
3029    29. Added (*CR) etc for changing newline setting at start of pattern.
3030    
3031    
3032    Version 7.2 19-Jun-07
3033    ---------------------
3034    
3035     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3036        which is apparently normally available under Windows.
3037    
3038     2. Re-jig the pcregrep tests with different newline settings in an attempt
3039        to make them independent of the local environment's newline setting.
3040    
3041     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3042    
3043     4. Some of the "internals" tests were previously cut out when the link size
3044        was not 2, because the output contained actual offsets. The recent new
3045        "Z" feature of pcretest means that these can be cut out, making the tests
3046        usable with all link sizes.
3047    
3048     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3049        stack recursion. This gives a massive performance boost under BSD, but just
3050        a small improvement under Linux. However, it saves one field in the frame
3051        in all cases.
3052    
3053     6. Added more features from the forthcoming Perl 5.10:
3054    
3055        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3056            recursion call. It refers to the nth most recently opened parentheses.
3057    
3058        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3059            to be opened parentheses.
3060    
3061        (c) Conditions that refer to capturing parentheses can be specified
3062            relatively, for example, (?(-2)... or (?(+3)...
3063    
3064        (d) \K resets the start of the current match so that everything before
3065            is not part of it.
3066    
3067        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3068    
3069        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3070            reference syntax.
3071    
3072        (g) (?| introduces a group in which the numbering of parentheses in each
3073            alternative starts with the same number.
3074    
3075        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3076    
3077     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3078        PCRE_INFO_JCHANGED.
3079    
3080     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3081        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3082        for detecting groups that can match an empty string.
3083    
3084     9. A pattern with a very large number of alternatives (more than several
3085        hundred) was running out of internal workspace during the pre-compile
3086        phase, where pcre_compile() figures out how much memory will be needed. A
3087        bit of new cunning has reduced the workspace needed for groups with
3088        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3089        workspace instead of running out of the 4096 that are available.
3090    
3091    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3092    
3093    11. Applied patch from Google to remove an optimization that didn't quite work.
3094        The report of the bug said:
3095    
3096          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3097          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3098          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3099    
3100    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3101        it matched the wrong number of bytes.
3102    
3103    
3104    Version 7.1 24-Apr-07
3105    ---------------------
3106    
3107     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3108        that is more "standard", making use of automake and other Autotools. There
3109        is some re-arrangement of the files and adjustment of comments consequent
3110        on this.
3111    
3112     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3113        for recursive directory scanning broke on some systems because the files
3114        are not scanned in any specific order and on different systems the order
3115        was different. A call to "sort" has been inserted into RunGrepTest for the
3116        approprate test as a short-term fix. In the longer term there may be an
3117        alternative.
3118    
3119     3. I had an email from Eric Raymond about problems translating some of PCRE's
3120        man pages to HTML (despite the fact that I distribute HTML pages, some
3121        people do their own conversions for various reasons). The problems
3122        concerned the use of low-level troff macros .br and .in. I have therefore
3123        removed all such uses from the man pages (some were redundant, some could
3124        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3125        HTML has been updated to handle .nf/.fi and to complain if it encounters
3126        .br or .in.
3127    
3128     4. Updated comments in configure.ac that get placed in config.h.in and also
3129        arranged for config.h to be included in the distribution, with the name
3130        config.h.generic, for the benefit of those who have to compile without
3131        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3132    
3133     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3134        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3135        makevp.bat for the current PCRE, using the additional files
3136        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3137    
3138     6. A Windows user reported a minor discrepancy with test 2, which turned out
3139        to be caused by a trailing space on an input line that had got lost in his
3140        copy. The trailing space was an accident, so I've just removed it.
3141    
3142     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3143        that is needed.
3144    
3145     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3146        as "const" (a) because they are and (b) because it helps the PHP
3147        maintainers who have recently made a script to detect big data structures
3148        in the php code that should be moved to the .rodata section. I remembered
3149        to update Builducptable as well, so it won't revert if ucptable.h is ever
3150        re-created.
3151    
3152     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3153        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3154        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3155        support is not required. This saves 1.5-2K of code, which is important in
3156        some applications.
3157    
3158        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3159        so as not to refer to the tables, even though these functions will never be
3160        called when UTF-8 support is disabled. Otherwise there are problems with a
3161        shared library.
3162    
3163    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3164    
3165        (a) It was defining its arguments as char * instead of void *.
3166    
3167        (b) It was assuming that all moves were upwards in memory; this was true
3168            a long time ago when I wrote it, but is no longer the case.
3169    
3170        The emulated memove() is provided for those environments that have neither
3171        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3172        is clearly not the case, as these two bugs were recently reported.
3173    
3174    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3175        and Detrail to create the HTML documentation, the .txt form of the man
3176        pages, and it removes trailing spaces from listed files. It also creates
3177        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3178        case, it wraps all the #defines with #ifndefs. This script should be run
3179        before "make dist".
3180    
3181    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3182        with Unicode property support.
3183    
3184        (a) For a maximizing quantifier, if the two different cases of the
3185            character were of different lengths in their UTF-8 codings (there are
3186            some cases like this - I found 11), and the matching function had to
3187            back up over a mixture of the two cases, it incorrectly assumed they
3188            were both the same length.
3189    
3190        (b) When PCRE was configured to use the heap rather than the stack for
3191            recursion during matching, it was not correctly preserving the data for
3192            the other case of a UTF-8 character when checking ahead for a match
3193            while processing a minimizing repeat. If the check also involved
3194            matching a wide character, but failed, corruption could cause an
3195            erroneous result when trying to check for a repeat of the original
3196            character.
3197    
3198    13. Some tidying changes to the testing mechanism:
3199    
3200        (a) The RunTest script now detects the internal link size and whether there
3201            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3202            values substituted by "configure". (The RunGrepTest script already did
3203            this for UTF-8.) The configure.ac script no longer substitutes the
3204            relevant variables.
3205    
3206        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3207            with length and offset values. This means that the output is different
3208            for different internal link sizes. Test 2 is skipped for link sizes
3209            other than 2 because of this, bypassing the problem. Unfortunately,
3210            there was also a test in test 3 (the locale tests) that used /B and
3211            failed for link sizes other than 2. Rather than cut the whole test out,
3212            I have added a new /Z option to pcretest that replaces the length and
3213            offset values with spaces. This is now used to make test 3 independent
3214            of link size. (Test 2 will be tidied up later.)
3215    
3216    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3217        segmentation fault instead of returning the appropriate error message.
3218    
3219    15. In multiline mode when the newline sequence was set to "any", the pattern
3220        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3221        This doesn't seem right; it now treats the CRLF combination as the line
3222        ending, and so does not match in that case. It's only a pattern such as ^$
3223        that would hit this one: something like ^ABC$ would have failed after \r
3224        and then tried again after \r\n.
3225    
3226    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3227        in an attempt to make files that differ only in their line terminators
3228        compare equal. This works on Linux.
3229    
3230    17. Under certain error circumstances pcregrep might try to free random memory
3231        as it exited. This is now fixed, thanks to valgrind.
3232    
3233    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3234        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3235        was because its rules for how to advance for /g after matching an empty
3236        string at the end of a line did not allow for this case. They now check for
3237        it specially.
3238    
3239    20. pcretest is supposed to handle patterns and data of any length, by
3240        extending its buffers when necessary. It was getting this wrong when the
3241        buffer for a data line had to be extended.
3242    
3243    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3244        CRLF as a newline sequence.
3245    
3246    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3247        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3248        I have nevertheless tidied it up.
3249    
3250    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3251    
3252    24. Added a man page for pcre-config.
3253    
3254    
3255    Version 7.0 19-Dec-06
3256    ---------------------
3257    
3258     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3259        moving to gcc 4.1.1.
3260    
3261     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3262        sys/time.h, which is documented as needed for this function. It doesn't
3263        seem to matter on Linux, but it showed up on some releases of OS X.
3264    
3265     3. It seems that there are systems where bytes whose values are greater than
3266        127 match isprint() in the "C" locale. The "C" locale should be the
3267        default when a C program starts up. In most systems, only ASCII printing
3268        characters match isprint(). This difference caused the output from pcretest
3269        to vary, making some of the tests fail. I have changed pcretest so that:
3270    
3271        (a) When it is outputting text in the compiled version of a pattern, bytes
3272            other than 32-126 are always shown as hex escapes.
3273    
3274        (b) When it is outputting text that is a matched part of a subject string,
3275            it does the same, unless a different locale has been set for the match
3276            (using the /L modifier). In this case, it uses isprint() to decide.
3277    
3278     4. Fixed a major bug that caused incorrect computation of the amount of memory
3279        required for a compiled pattern when options that changed within the
3280        pattern affected the logic of the preliminary scan that determines the
3281        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3282        that the computed length was too small. The symptoms of this bug were
3283        either the PCRE error "internal error: code overflow" from pcre_compile(),
3284        or a glibc crash with a message such as "pcretest: free(): invalid next
3285        size (fast)". Examples of patterns that provoked this bug (shown in
3286        pcretest format) are:
3287    
3288          /(?-x: )/x
3289          /(?x)(?-x: \s*#\s*)/
3290          /((?i)[\x{c0}])/8
3291          /(?i:[\x{c0}])/8
3292    
3293        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3294        is now done differently.
3295    
3296     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3297        wrapper classes; (b) implement a new function in the C++ scanner that is
3298        more efficient than the old way of doing things because it avoids levels of
3299        recursion in the regex matching; (c) add a paragraph to the documentation
3300        for the FullMatch() function.
3301    
3302     6. The escape sequence \n was being treated as whatever was defined as
3303        "newline". Not only was this contrary to the documentation, which states
3304        that \n is character 10 (hex 0A), but it also went horribly wrong when
3305        "newline" was defined as CRLF. This has been fixed.
3306    
3307     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3308        was being set to -1 for the "end of line" case (supposedly a value that no
3309        character can have). Though this value is never used (the check for end of
3310        line is "zero bytes in current character"), it caused compiler complaints.
3311        I've changed it to 0xffffffff.
3312    
3313     8. In pcre_version.c, the version string was being built by a sequence of
3314        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3315        string (as it is for production releases) called a macro with an empty
3316        argument. The C standard says the result of this is undefined. The gcc
3317        compiler treats it as an empty string (which was what was wanted) but it is
3318        reported that Visual C gives an error. The source has been hacked around to
3319        avoid this problem.
3320    
3321     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3322        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3323        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3324        of them did).
3325    
3326    10. Originally, pcretest opened its input and output without "b"; then I was
3327        told that "b" was needed in some environments, so it was added for release
3328        5.0 to both the input and output. (It makes no difference on Unix-like
3329        systems.) Later I was told that it is wrong for the input on Windows. I've
3330        now abstracted the modes into two macros, to make it easier to fiddle with
3331        them, and removed "b" from the input mode under Windows.
3332    
3333    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3334    
3335    12. Added -help and --help to pcretest as an official way of being reminded
3336        of the options.
3337    
3338    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3339        and pcrecpp.cc because they annoy compilers at high warning levels.
3340    
3341    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3342    
3343    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3344        scripts are not C programs :-) and which was not noticed because it works
3345        on Linux.
3346    
3347    16. pcretest is supposed to handle any length of pattern and data line (as one
3348        line or as a continued sequence of lines) by extending its input buffer if
3349        necessary. This feature was broken for very long pattern lines, leading to
3350        a string of junk being passed to pcre_compile() if the pattern was longer
3351        than about 50K.
3352    
3353    17. I have done a major re-factoring of the way pcre_compile() computes the
3354        amount of memory needed for a compiled pattern. Previously, there was code
3355        that made a preliminary scan of the pattern in order to do this. That was
3356        OK when PCRE was new, but as the facilities have expanded, it has become
3357        harder and harder to keep it in step with the real compile phase, and there
3358        have been a number of bugs (see for example, 4 above). I have now found a
3359        cunning way of running the real compile function in a "fake" mode that
3360        enables it to compute how much memory it would need, while actually only
3361        ever using a few hundred bytes of working memory and without too many
3362        tests of the mode. This should make future maintenance and development
3363        easier. A side effect of this work is that the limit of 200 on the nesting
3364        depth of parentheses has been removed (though this was never a serious
3365        limitation, I suspect). However, there is a downside: pcre_compile() now
3366        runs more slowly than before (30% or more, depending on the pattern). I
3367        hope this isn't a big issue. There is no effect on runtime performance.
3368    
3369    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3370        newline (only possible for the last line of a file) and it was a
3371        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3372    
3373    19. Added additional timing features to pcretest. (1) The -tm option now times
3374        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3375        separate command line item, by a number that specifies the number of
3376        repeats to use when timing. The default is 50000; this gives better
3377        precision, but takes uncomfortably long for very large patterns.
3378    
3379    20. Extended pcre_study() to be more clever in cases where a branch of a
3380        subpattern has no definite first character. For example, (a*|b*)[cd] would
3381        previously give no result from pcre_study(). Now it recognizes that the
3382        first character must be a, b, c, or d.
3383    
3384    21. There was an incorrect error "recursive call could loop indefinitely" if
3385        a subpattern (or the entire pattern) that was being tested for matching an
3386        empty string contained only one non-empty item after a nested subpattern.
3387        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3388        incorrectly, because the \d was being skipped in the check.
3389    
3390    22. The pcretest program now has a new pattern option /B and a command line
3391        option -b, which is equivalent to adding /B to every pattern. This causes
3392        it to show the compiled bytecode, without the additional information that
3393        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3394        is the same as /B/I).
3395    
3396    23. A new optimization is now able automatically to treat some sequences such
3397        as a*b as a*+b. More specifically, if something simple (such as a character
3398        or a simple class like \d) has an unlimited quantifier, and is followed by
3399        something that cannot possibly match the quantified thing, the quantifier
3400        is automatically "possessified".
3401    
3402    24. A recursive reference to a subpattern whose number was greater than 39
3403        went wrong under certain circumstances in UTF-8 mode. This bug could also
3404        have affected the operation of pcre_study().
3405    
3406    25. Realized that a little bit of performance could be had by replacing
3407        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3408    
3409    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3410    
3411    27. Possessive quantifiers such as a++ were previously implemented by turning
3412        them into atomic groups such as ($>a+). Now they have their own opcodes,
3413        which improves performance. This includes the automatically created ones
3414        from 23 above.
3415    
3416    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3417        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3418        the first matched character to be a colon. This applied both to named and
3419        numbered groups.
3420    
3421    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3422    
3423    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3424        building PCRE on Windows easier, so I have included it in the distribution.
3425    
3426    31. There is now a check in pcretest against a ridiculously large number being
3427        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3428        loop, the loop is abandoned.
3429    
3430    32. Forward references to subpatterns in conditions such as (?(2)...) where
3431        subpattern 2 is defined later cause pcre_compile() to search forwards in
3432        the pattern for the relevant set of parentheses. This search went wrong
3433        when there were unescaped parentheses in a character class, parentheses
3434        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3435    
3436    33. "Subroutine" calls and backreferences were previously restricted to
3437        referencing subpatterns earlier in the regex. This restriction has now
3438        been removed.
3439    
3440    34. Added a number of extra features that are going to be in Perl 5.10. On the
3441        whole, these are just syntactic alternatives for features that PCRE had
3442        previously implemented using the Python syntax or my own invention. The
3443        other formats are all retained for compatibility.
3444    
3445        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3446            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3447            also .NET compatible.
3448    
3449        (b) A recursion or subroutine call to a named group can now be defined as
3450            (?&name) as well as (?P>name).
3451    
3452        (c) A backreference to a named group can now be defined as \k<name> or
3453            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3454            5.10, are also .NET compatible.
3455    
3456        (d) A conditional reference to a named group can now use the syntax
3457            (?(<name>) or (?('name') as well as (?(name).
3458    
3459        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3460            groups (named and numbered) that are never evaluated inline, but can be
3461            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3462            is always false. There may be only one alternative in such a group.
3463    
3464        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3465            as the simple (?(R). The condition is true only if the most recent
3466            recursion is that of the given number or name. It does not search out
3467            through the entire recursion stack.
3468    
3469        (g) The escape \gN or \g{N} has been added, where N is a positive or
3470            negative number, specifying an absolute or relative reference.
3471    
3472    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3473        some "unreachable code" warnings.
3474    
3475    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3476        things, this adds five new scripts.
3477    
3478    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3479        There were also incompatibilities regarding the handling of \Q..\E inside
3480        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3481        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3482    
3483    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3484        matches an empty string, and forcibly breaks the loop. There were bugs in
3485        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3486        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3487        separate and independent bugs (that affected different cases) have been
3488        fixed.
3489    
3490    39. Refactored the code to abolish the use of different opcodes for small
3491        capturing bracket numbers. This is a tidy that I avoided doing when I
3492        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3493        The new approach is not only tidier, it makes it possible to reduce the
3494        memory needed to fix the previous bug (38).
3495    
3496    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3497        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3498        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3499        mode.
3500    
3501    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3502        report.
3503    
3504    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3505        copy construction and assignment in the C++ wrapper.
3506    
3507    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3508        couple of bugs that could have given wrong results in the "--newline=crlf"
3509        case.
3510    
3511    44. Added a number of casts and did some reorganization of signed/unsigned int
3512        variables following suggestions from Dair Grant. Also renamed the variable
3513        "this" as "item" because it is a C++ keyword.
3514    
3515    45. Arranged for dftables to add
3516    
3517          #include "pcre_internal.h"
3518    
3519        to pcre_chartables.c because without it, gcc 4.x may remove the array
3520        definition from the final binary if PCRE is built into a static library and
3521        dead code stripping is activated.
3522    
3523    46. For an unanchored pattern, if a match attempt fails at the start of a
3524        newline sequence, and the newline setting is CRLF or ANY, and the next two
3525        characters are CRLF, advance by two characters instead of one.
3526    
3527    
3528    Version 6.7 04-Jul-06
3529    ---------------------
3530    
3531     1. In order to handle tests when input lines are enormously long, pcretest has
3532        been re-factored so that it automatically extends its buffers when
3533        necessary. The code is crude, but this _is_ just a test program. The
3534        default size has been increased from 32K to 50K.
3535    
3536     2. The code in pcre_study() was using the value of the re argument before
3537        testing it for NULL. (Of course, in any sensible call of the function, it
3538        won't be NULL.)
3539    
3540     3. The memmove() emulation function in pcre_internal.h, which is used on
3541        systems that lack both memmove() and bcopy() - that is, hardly ever -
3542        was missing a "static" storage class specifier.
3543    
3544     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3545        containing an extended class (one that cannot be represented by a bitmap
3546        because it contains high-valued characters or Unicode property items, e.g.
3547        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3548        pattern, but PCRE should not loop if you do not (it no longer does).
3549        [Detail: two cases were found: (a) a repeated subpattern containing an
3550        extended class; (b) a recursive reference to a subpattern that followed a
3551        previous extended class. It wasn't skipping over the extended class
3552        correctly when UTF-8 mode was not set.]
3553    
3554     5. A negated single-character class was not being recognized as fixed-length
3555        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3556        compile error "lookbehind assertion is not fixed length".
3557    
3558     6. The RunPerlTest auxiliary script was showing an unexpected difference
3559        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3560        write a Perl script that can interpret lines of an input file either as
3561        byte characters or as UTF-8, which is what "perltest" was being required to
3562        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3563        can't do is switch easily at run time between having the "use utf8;" pragma
3564        or not. In the end, I fudged it by using the RunPerlTest script to insert
3565        "use utf8;" explicitly for the UTF-8 tests.
3566    
3567     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3568        the end of the subject string, contrary to the documentation and to what
3569        Perl does. This was true of both matching functions. Now it matches only at
3570        the start of the subject and immediately after *internal* newlines.
3571    
3572     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3573        a pointer to an int instead of a pointer to an unsigned long int. This
3574        caused problems on 64-bit systems.
3575    
3576     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3577        instance of the 'standard' template library not being so standard".
3578    
3579    10. There was no check on the number of named subpatterns nor the maximum
3580        length of a subpattern name. The product of these values is used to compute
3581        the size of the memory block for a compiled pattern. By supplying a very
3582        long subpattern name and a large number of named subpatterns, the size
3583