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

Diff of /code/trunk/ChangeLog

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

<
revision 65 by nigel, Sat Feb 24 21:40:08 2007 UTC revision 1570 by ph10, Mon Jun 22 09:38:41 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    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65        ASCII/Unicode. This has now been added to the list of characters that are
66        recognized as white space in EBCDIC.
67    
68    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69        error (correctly) when used outside a class, but did not give an error
70        within a class.
71    
72    17. \h within a class was incorrectly compiled in EBCDIC environments.
73    
74    
75    Version 8.37 28-April-2015
76    --------------------------
77    
78    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
79        for those parentheses to be closed with whatever has been captured so far.
80        However, it was failing to mark any other groups between the hightest
81        capture so far and the currrent group as "unset". Thus, the ovector for
82        those groups contained whatever was previously there. An example is the
83        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
84    
85    2.  If an assertion condition was quantified with a minimum of zero (an odd
86        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
87    
88    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
89        unrecognized modifier, a crash could occur.
90    
91    4.  An attempt to do global matching in pcretest with a zero-length ovector
92        caused a crash.
93    
94    5.  Fixed a memory leak during matching that could occur for a subpattern
95        subroutine call (recursive or otherwise) if the number of captured groups
96        that had to be saved was greater than ten.
97    
98    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
99        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
100        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
101    
102    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
103        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
104    
105    8. When a pattern is compiled, it remembers the highest back reference so that
106       when matching, if the ovector is too small, extra memory can be obtained to
107       use instead. A conditional subpattern whose condition is a check on a
108       capture having happened, such as, for example in the pattern
109       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
110       setting the highest backreference number. This mattered only if pcre_exec()
111       was called with an ovector that was too small to hold the capture, and there
112       was no other kind of back reference (a situation which is probably quite
113       rare). The effect of the bug was that the condition was always treated as
114       FALSE when the capture could not be consulted, leading to a incorrect
115       behaviour by pcre_exec(). This bug has been fixed.
116    
117    9. A reference to a duplicated named group (either a back reference or a test
118       for being set in a conditional) that occurred in a part of the pattern where
119       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
120       to be incorrectly calculated, leading to overwriting.
121    
122    10. A mutually recursive set of back references such as (\2)(\1) caused a
123        segfault at study time (while trying to find the minimum matching length).
124        The infinite loop is now broken (with the minimum length unset, that is,
125        zero).
126    
127    11. If an assertion that was used as a condition was quantified with a minimum
128        of zero, matching went wrong. In particular, if the whole group had
129        unlimited repetition and could match an empty string, a segfault was
130        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
131        assertions to be quantified, but not if they are being used as conditions,
132        so the above pattern is faulted by Perl. PCRE has now been changed so that
133        it also rejects such patterns.
134    
135    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
136        failed to allow the zero-repeat case if pcre2_exec() was called with an
137        ovector too small to capture the group.
138    
139    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
140        Red Hat Product Security:
141    
142        (a) A crash if /K and /F were both set with the option to save the compiled
143        pattern.
144    
145        (b) Another crash if the option to print captured substrings in a callout
146        was combined with setting a null ovector, for example \O\C+ as a subject
147        string.
148    
149    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
150        forward reference repeated a large (but limited) number of times within a
151        repeated outer group that has a zero minimum quantifier, caused incorrect
152        code to be compiled, leading to the error "internal error:
153        previously-checked referenced subpattern not found" when an incorrect
154        memory address was read. This bug was reported as "heap overflow",
155        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
156        CVE-2015-2325.
157    
158    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
159        call within a group that also contained a recursive back reference caused
160        incorrect code to be compiled. This bug was reported as "heap overflow",
161        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
162        number CVE-2015-2326.
163    
164    24. Computing the size of the JIT read-only data in advance has been a source
165        of various issues, and new ones are still appear unfortunately. To fix
166        existing and future issues, size computation is eliminated from the code,
167        and replaced by on-demand memory allocation.
168    
169    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
170        adjacent to the end of the range, and the range contained characters with
171        more than one other case, caused incorrect behaviour when compiled in UTF
172        mode. In that example, the range a-j was left out of the class.
173    
174    26. Fix JIT compilation of conditional blocks, which assertion
175        is converted to (*FAIL). E.g: /(?(?!))/.
176    
177    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
178        discovered by the LLVM fuzzer.
179    
180    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
181        when this assertion was used as a condition, for example (?(?!)a|b). In
182        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
183        error about an unsupported item.
184    
185    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
186        possessification code could take exponential time to complete. A recursion
187        depth limit of 1000 has been imposed to limit the resources used by this
188        optimization.
189    
190    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
191        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
192        because \S ensures they are all in the class. The code for doing this was
193        interacting badly with the code for computing the amount of space needed to
194        compile the pattern, leading to a buffer overflow. This bug was discovered
195        by the LLVM fuzzer.
196    
197    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
198        other kinds of group caused stack overflow at compile time. This bug was
199        discovered by the LLVM fuzzer.
200    
201    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
202        between a subroutine call and its quantifier was incorrectly compiled,
203        leading to buffer overflow or other errors. This bug was discovered by the
204        LLVM fuzzer.
205    
206    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
207        assertion after (?(. The code was failing to check the character after
208        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
209        was discovered by the LLVM fuzzer.
210    
211    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
212        a fixed maximum following a group that contains a subroutine reference was
213        incorrectly compiled and could trigger buffer overflow. This bug was
214        discovered by the LLVM fuzzer.
215    
216    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
217        caused a stack overflow instead of the diagnosis of a non-fixed length
218        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
219    
220    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
221        (e.g. /(?<=\Ka)/) could make pcregrep loop.
222    
223    37. There was a similar problem to 36 in pcretest for global matches.
224    
225    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
226        and a subsequent item in the pattern caused a non-match, backtracking over
227        the repeated \X did not stop, but carried on past the start of the subject,
228        causing reference to random memory and/or a segfault. There were also some
229        other cases where backtracking after \C could crash. This set of bugs was
230        discovered by the LLVM fuzzer.
231    
232    39. The function for finding the minimum length of a matching string could take
233        a very long time if mutual recursion was present many times in a pattern,
234        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
235        method has been implemented. This infelicity was discovered by the LLVM
236        fuzzer.
237    
238    40. Static linking against the PCRE library using the pkg-config module was
239        failing on missing pthread symbols.
240    
241    
242    Version 8.36 26-September-2014
243    ------------------------------
244    
245    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
246        -Wmissing-field-initializers and -Wunused-parameter.
247    
248    2.  The tests for quantifiers being too big (greater than 65535) were being
249        applied after reading the number, and stupidly assuming that integer
250        overflow would give a negative number. The tests are now applied as the
251        numbers are read.
252    
253    3.  Tidy code in pcre_exec.c where two branches that used to be different are
254        now the same.
255    
256    4.  The JIT compiler did not generate match limit checks for certain
257        bracketed expressions with quantifiers. This may lead to exponential
258        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
259        issue should be resolved now.
260    
261    5.  Fixed an issue, which occures when nested alternatives are optimized
262        with table jumps.
263    
264    6.  Inserted two casts and changed some ints to size_t in the light of some
265        reported 64-bit compiler warnings (Bugzilla 1477).
266    
267    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
268        an empty string, which sometimes were behaving incorrectly in the
269        interpreter (though correctly in the JIT matcher). This pcretest input is
270        an example:
271    
272          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
273          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
274    
275        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
276        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
277        for an empty string was breaking the inner loop and carrying on at a lower
278        level, when possessive repeated groups should always return to a higher
279        level as they have no backtrack points in them. The empty string test now
280        occurs at the outer level.
281    
282    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
283        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
284    
285    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
286        doesn't).
287    
288    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
289        to hold a pattern to be incorrectly computed (too small) when there were
290        named back references to duplicated names. This could cause "internal
291        error: code overflow" or "double free or corruption" or other memory
292        handling errors.
293    
294    11. When named subpatterns had the same prefixes, back references could be
295        confused. For example, in this pattern:
296    
297          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
298    
299        the reference to 'Name' was incorrectly treated as a reference to a
300        duplicate name.
301    
302    12. A pattern such as /^s?c/mi8 where the optional character has more than
303        one "other case" was incorrectly compiled such that it would only try to
304        match starting at "c".
305    
306    13. When a pattern starting with \s was studied, VT was not included in the
307        list of possible starting characters; this should have been part of the
308        8.34/18 patch.
309    
310    14. If a character class started [\Qx]... where x is any character, the class
311        was incorrectly terminated at the ].
312    
313    15. If a pattern that started with a caseless match for a character with more
314        than one "other case" was studied, PCRE did not set up the starting code
315        unit bit map for the list of possible characters. Now it does. This is an
316        optimization improvement, not a bug fix.
317    
318    16. The Unicode data tables have been updated to Unicode 7.0.0.
319    
320    17. Fixed a number of memory leaks in pcregrep.
321    
322    18. Avoid a compiler warning (from some compilers) for a function call with
323        a cast that removes "const" from an lvalue by using an intermediate
324        variable (to which the compiler does not object).
325    
326    19. Incorrect code was compiled if a group that contained an internal recursive
327        back reference was optional (had quantifier with a minimum of zero). This
328        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
329        caused segmentation faults because of stack overflows at compile time.
330    
331    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
332        group that is quantified with an indefinite repeat, caused a compile-time
333        loop which used up all the system stack and provoked a segmentation fault.
334        This was not the same bug as 19 above.
335    
336    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
337        Patch by Mike Frysinger.
338    
339    
340    Version 8.35 04-April-2014
341    --------------------------
342    
343    1.  A new flag is set, when property checks are present in an XCLASS.
344        When this flag is not set, PCRE can perform certain optimizations
345        such as studying these XCLASS-es.
346    
347    2.  The auto-possessification of character sets were improved: a normal
348        and an extended character set can be compared now. Furthermore
349        the JIT compiler optimizes more character set checks.
350    
351    3.  Got rid of some compiler warnings for potentially uninitialized variables
352        that show up only when compiled with -O2.
353    
354    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
355        of a match later then the end of the match. The pcretest program was not
356        handling the case sensibly - it was outputting from the start to the next
357        binary zero. It now reports this situation in a message, and outputs the
358        text from the end to the start.
359    
360    5.  Fast forward search is improved in JIT. Instead of the first three
361        characters, any three characters with fixed position can be searched.
362        Search order: first, last, middle.
363    
364    6.  Improve character range checks in JIT. Characters are read by an inprecise
365        function now, which returns with an unknown value if the character code is
366        above a certain threshold (e.g: 256). The only limitation is that the value
367        must be bigger than the threshold as well. This function is useful when
368        the characters above the threshold are handled in the same way.
369    
370    7.  The macros whose names start with RAWUCHAR are placeholders for a future
371        mode in which only the bottom 21 bits of 32-bit data items are used. To
372        make this more memorable for those maintaining the code, the names have
373        been changed to start with UCHAR21, and an extensive comment has been added
374        to their definition.
375    
376    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
377        to the export list in Makefile.am (they were accidentally omitted from the
378        8.34 tarball).
379    
380    9.  The informational output from pcretest used the phrase "starting byte set"
381        which is inappropriate for the 16-bit and 32-bit libraries. As the output
382        for "first char" and "need char" really means "non-UTF-char", I've changed
383        "byte" to "char", and slightly reworded the output. The documentation about
384        these values has also been (I hope) clarified.
385    
386    10. Another JIT related optimization: use table jumps for selecting the correct
387        backtracking path, when more than four alternatives are present inside a
388        bracket.
389    
390    11. Empty match is not possible, when the minimum length is greater than zero,
391        and there is no \K in the pattern. JIT should avoid empty match checks in
392        such cases.
393    
394    12. In a caseless character class with UCP support, when a character with more
395        than one alternative case was not the first character of a range, not all
396        the alternative cases were added to the class. For example, s and \x{17f}
397        are both alternative cases for S: the class [RST] was handled correctly,
398        but [R-T] was not.
399    
400    13. The configure.ac file always checked for pthread support when JIT was
401        enabled. This is not used in Windows, so I have put this test inside a
402        check for the presence of windows.h (which was already tested for).
403    
404    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
405        The algorithm provides a way to skip certain starting offsets, and usually
406        faster than linear prefix searches.
407    
408    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
409        as for 'fr_FR' and 'french'. For some reason, however, it then used the
410        Windows-specific input and output files, which have 'french' screwed in.
411        So this could never have worked. One of the problems with locales is that
412        they aren't always the same. I have now updated RunTest so that it checks
413        the output of the locale test (test 3) against three different output
414        files, and it allows the test to pass if any one of them matches. With luck
415        this should make the test pass on some versions of Solaris where it was
416        failing. Because of the uncertainty, the script did not used to stop if
417        test 3 failed; it now does. If further versions of a French locale ever
418        come to light, they can now easily be added.
419    
420    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
421        there was a message during ./configure, but it did not stop. This now
422        provokes an error. The invalid example in README has been corrected.
423        If a value less than the minimum is given, the minimum value has always
424        been used, but now a warning is given.
425    
426    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
427        was a bug in the test system, which is now fixed. Also, the list of various
428        configurations that are tested for each release did not have one with both
429        16/32 bits and --enable-bar-anycrlf. It now does.
430    
431    18. pcretest was missing "-C bsr" for displaying the \R default setting.
432    
433    19. Little endian PowerPC systems are supported now by the JIT compiler.
434    
435    20. The fast forward newline mechanism could enter to an infinite loop on
436        certain invalid UTF-8 input. Although we don't support these cases
437        this issue can be fixed by a performance optimization.
438    
439    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
440        not take account if existing stack usage. There is now a new global
441        variable called pcre_stack_guard that can be set to point to an external
442        function to check stack availability. It is called at the start of
443        processing every parenthesized group.
444    
445    22. A typo in the code meant that in ungreedy mode the max/min qualifier
446        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
447        not match "ab".
448    
449    23. When UTF was disabled, the JIT program reported some incorrect compile
450        errors. These messages are silenced now.
451    
452    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
453        compiler.
454    
455    25. Change all the temporary files used in RunGrepTest to be different to those
456        used by RunTest so that the tests can be run simultaneously, for example by
457        "make -j check".
458    
459    
460    Version 8.34 15-December-2013
461    -----------------------------
462    
463    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
464        executable memory. Patch inspired by Carsten Klein.
465    
466    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
467        this macro is never tested and has no effect, because the work to support
468        coverage involves only compiling and linking options and special targets in
469        the Makefile. The comment in config.h implied that defining the macro would
470        enable coverage support, which is totally false. There was also support for
471        setting this macro in the CMake files (my fault, I just copied it from
472        configure). SUPPORT_GCOV has now been removed.
473    
474    3.  Make a small performance improvement in strlen16() and strlen32() in
475        pcretest.
476    
477    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
478        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
479        these cases). There was also one in pcretest.c.
480    
481    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
482    
483    6.  In UTF mode, the code for checking whether a group could match an empty
484        string (which is used for indefinitely repeated groups to allow for
485        breaking an infinite loop) was broken when the group contained a repeated
486        negated single-character class with a character that occupied more than one
487        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
488        UTF-8 mode). The effect was undefined: the group might or might not be
489        deemed as matching an empty string, or the program might have crashed.
490    
491    7.  The code for checking whether a group could match an empty string was not
492        recognizing that \h, \H, \v, \V, and \R must match a character.
493    
494    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
495        an empty string. If it can, pcretest shows this in its information output.
496    
497    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
498        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
499        matched by pcre_exec() without using JIT:
500    
501        (a) If the rest of the pattern did not match after a maximal run of
502            grapheme clusters, the code for backing up to try with fewer of them
503            did not always back up over a full grapheme when characters that do not
504            have the modifier quality were involved, e.g. Hangul syllables.
505    
506        (b) If the match point in a subject started with modifier character, and
507            there was no match, the code could incorrectly back up beyond the match
508            point, and potentially beyond the first character in the subject,
509            leading to a segfault or an incorrect match result.
510    
511    10. A conditional group with an assertion condition could lead to PCRE
512        recording an incorrect first data item for a match if no other first data
513        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
514        first data item, and therefore matched "ca" after "c" instead of at the
515        start.
516    
517    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
518        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
519    
520    12. The source of pcregrep now includes z/OS-specific code so that it can be
521        compiled for z/OS as part of the special z/OS distribution.
522    
523    13. Added the -T and -TM options to pcretest.
524    
525    14. The code in pcre_compile.c for creating the table of named capturing groups
526        has been refactored. Instead of creating the table dynamically during the
527        actual compiling pass, the information is remembered during the pre-compile
528        pass (on the stack unless there are more than 20 named groups, in which
529        case malloc() is used) and the whole table is created before the actual
530        compile happens. This has simplified the code (it is now nearly 150 lines
531        shorter) and prepared the way for better handling of references to groups
532        with duplicate names.
533    
534    15. A back reference to a named subpattern when there is more than one of the
535        same name now checks them in the order in which they appear in the pattern.
536        The first one that is set is used for the reference. Previously only the
537        first one was inspected. This change makes PCRE more compatible with Perl.
538    
539    16. Unicode character properties were updated from Unicode 6.3.0.
540    
541    17. The compile-time code for auto-possessification has been refactored, based
542        on a patch by Zoltan Herczeg. It now happens after instead of during
543        compilation. The code is cleaner, and more cases are handled. The option
544        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
545        options in pcretest are provided to set it. It can also be set by
546        (*NO_AUTO_POSSESS) at the start of a pattern.
547    
548    18. The character VT has been added to the default ("C" locale) set of
549        characters that match \s and are generally treated as white space,
550        following this same change in Perl 5.18. There is now no difference between
551        "Perl space" and "POSIX space". Whether VT is treated as white space in
552        other locales depends on the locale.
553    
554    19. The code for checking named groups as conditions, either for being set or
555        for being recursed, has been refactored (this is related to 14 and 15
556        above). Processing unduplicated named groups should now be as fast at
557        numerical groups, and processing duplicated groups should be faster than
558        before.
559    
560    20. Two patches to the CMake build system, by Alexander Barkov:
561    
562          (1) Replace the "source" command by "." in CMakeLists.txt because
563              "source" is a bash-ism.
564    
565          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
566              without these the CMake build does not work on Solaris.
567    
568    21. Perl has changed its handling of \8 and \9. If there is no previously
569        encountered capturing group of those numbers, they are treated as the
570        literal characters 8 and 9 instead of a binary zero followed by the
571        literals. PCRE now does the same.
572    
573    22. Following Perl, added \o{} to specify codepoints in octal, making it
574        possible to specify values greater than 0777 and also making them
575        unambiguous.
576    
577    23. Perl now gives an error for missing closing braces after \x{... instead of
578        treating the string as literal. PCRE now does the same.
579    
580    24. RunTest used to grumble if an inappropriate test was selected explicitly,
581        but just skip it when running all tests. This make it awkward to run ranges
582        of tests when one of them was inappropriate. Now it just skips any
583        inappropriate tests, as it always did when running all tests.
584    
585    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
586        character types such as \d or \w, too many callouts were inserted, and the
587        data that they returned was rubbish.
588    
589    26. In UCP mode, \s was not matching two of the characters that Perl matches,
590        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
591        were matched by \h. The code has now been refactored so that the lists of
592        the horizontal and vertical whitespace characters used for \h and \v (which
593        are defined only in one place) are now also used for \s.
594    
595    27. Add JIT support for the 64 bit TileGX architecture.
596        Patch by Jiong Wang (Tilera Corporation).
597    
598    28. Possessive quantifiers for classes (both explicit and automatically
599        generated) now use special opcodes instead of wrapping in ONCE brackets.
600    
601    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
602        (because it's meaningless), this was not happening when PCRE_CASELESS was
603        set. Not wrong, but inefficient.
604    
605    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
606        properties for \w, \d, etc) is present in a test regex. Otherwise if the
607        test contains no characters greater than 255, Perl doesn't realise it
608        should be using Unicode semantics.
609    
610    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
611        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
612        does in Unicode mode.
613    
614    32. Added the "forbid" facility to pcretest so that putting tests into the
615        wrong test files can sometimes be quickly detected.
616    
617    33. There is now a limit (default 250) on the depth of nesting of parentheses.
618        This limit is imposed to control the amount of system stack used at compile
619        time. It can be changed at build time by --with-parens-nest-limit=xxx or
620        the equivalent in CMake.
621    
622    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
623        errors. Perl warns for these when in warning mode, but PCRE has no facility
624        for giving warnings.
625    
626    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
627        However, this was not working for (?!) because it is optimized to (*FAIL),
628        for which PCRE does not allow quantifiers. The optimization is now disabled
629        when a quantifier follows (?!). I can't see any use for this, but it makes
630        things uniform.
631    
632    36. Perl no longer allows group names to start with digits, so I have made this
633        change also in PCRE. It simplifies the code a bit.
634    
635    37. In extended mode, Perl ignores spaces before a + that indicates a
636        possessive quantifier. PCRE allowed a space before the quantifier, but not
637        before the possessive +. It now does.
638    
639    38. The use of \K (reset reported match start) within a repeated possessive
640        group such as (a\Kb)*+ was not working.
641    
642    40. Document that the same character tables must be used at compile time and
643        run time, and that the facility to pass tables to pcre_exec() and
644        pcre_dfa_exec() is for use only with saved/restored patterns.
645    
646    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
647        features for Builds with MSVC:
648    
649        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
650           builds). The .rc files can be used to set FileDescription and many other
651           attributes.
652    
653        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
654           This allows higher-level build scripts which want .pdb files to avoid
655           hard-coding the exact files needed."
656    
657    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
658        mean "start of word" and "end of word", respectively, as a transition aid.
659    
660    43. A minimizing repeat of a class containing codepoints greater than 255 in
661        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
662        compiled to use the heap for recursion.
663    
664    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
665        is configured.
666    
667    
668    Version 8.33 28-May-2013
669    ------------------------
670    
671    1.  Added 'U' to some constants that are compared to unsigned integers, to
672        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
673        variables that are added to signed variables, to ensure the result is
674        signed and can be negated.
675    
676    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
677        CMake config files.
678    
679    3.  Revise the creation of config.h.generic so that all boolean macros are
680        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
681        overriding via -D on the command line possible.
682    
683    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
684        to unsigned int is reported to make a quite noticeable speed difference in
685        a specific Windows environment. Testing on Linux did also appear to show
686        some benefit (and it is clearly not harmful). Also fixed the definition of
687        Xop which should be unsigned.
688    
689    5.  Related to (4), changing the definition of the intermediate variable cc
690        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
691        improvements.
692    
693    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
694        unnecessary spaces.
695    
696    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
697        and later.
698    
699    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
700    
701    9.  Optimizing fast_forward_start_bits in JIT.
702    
703    10. Adding support for callouts in JIT, and fixing some issues revealed
704        during this work. Namely:
705    
706        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
707    
708        (b) Minimum length was not checked before the matching is started.
709    
710    11. The value of capture_last that is passed to callouts was incorrect in some
711        cases when there was a capture on one path that was subsequently abandoned
712        after a backtrack. Also, the capture_last value is now reset after a
713        recursion, since all captures are also reset in this case.
714    
715    12. The interpreter no longer returns the "too many substrings" error in the
716        case when an overflowing capture is in a branch that is subsequently
717        abandoned after a backtrack.
718    
719    13. In the pathological case when an offset vector of size 2 is used, pcretest
720        now prints out the matched string after a yield of 0 or 1.
721    
722    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
723        Only supported by the JIT compiler at the moment.
724    
725    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
726    
727    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
728        offset of the starting point of the matching process, provided the offsets
729        vector is large enough.
730    
731    17. The \A escape now records a lookbehind value of 1, though its execution
732        does not actually inspect the previous character. This is to ensure that,
733        in partial multi-segment matching, at least one character from the old
734        segment is retained when a new segment is processed. Otherwise, if there
735        are no lookbehinds in the pattern, \A might match incorrectly at the start
736        of a new segment.
737    
738    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
739    
740    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
741        gives some modest performance improvement in 8-bit mode.
742    
743    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
744        be expressed in certain programming languages using Universal Character
745        Names.
746    
747    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
748        which points out that "non characters" are not "characters that may not
749        appear in Unicode strings" but rather "characters that are reserved for
750        internal use and have only local meaning".
751    
752    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
753        there was a conditional group that depended on an assertion, if the
754        assertion was false, the callout that immediately followed the alternation
755        in the condition was skipped when pcre_exec() was used for matching.
756    
757    23. Allow an explicit callout to be inserted before an assertion that is the
758        condition for a conditional group, for compatibility with automatic
759        callouts, which always insert a callout at this point.
760    
761    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
762        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
763    
764    25. (*PRUNE) is now supported by the JIT compiler.
765    
766    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
767    
768    27. Fix the case where there are two or more SKIPs with arguments that may be
769        ignored.
770    
771    28. (*SKIP) is now supported by the JIT compiler.
772    
773    29. (*THEN) is now supported by the JIT compiler.
774    
775    30. Update RunTest with additional test selector options.
776    
777    31. The way PCRE handles backtracking verbs has been changed in two ways.
778    
779        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
780        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
781        backtracking. In some cases this makes it more Perl-compatible, but Perl's
782        rather obscure rules do not always do the same thing.
783    
784        (2) Previously, backtracking verbs were confined within assertions. This is
785        no longer the case for positive assertions, except for (*ACCEPT). Again,
786        this sometimes improves Perl compatibility, and sometimes does not.
787    
788    32. A number of tests that were in test 2 because Perl did things differently
789        have been moved to test 1, because either Perl or PCRE has changed, and
790        these tests are now compatible.
791    
792    32. Backtracking control verbs are now handled in the same way in JIT and
793        interpreter.
794    
795    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
796        contained a forward subroutine reference caused a compile error.
797    
798    34. Auto-detect and optimize limited repetitions in JIT.
799    
800    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
801        blocking (*UTF) etc.
802    
803    36. In the interpreter, maximizing pattern repetitions for characters and
804        character types now use tail recursion, which reduces stack usage.
805    
806    37. The value of the max lookbehind was not correctly preserved if a compiled
807        and saved regex was reloaded on a host of different endianness.
808    
809    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
810        of the compiled pattern block, expand the flags field from 16 to 32 bits
811        because it was almost full.
812    
813    39. Try madvise first before posix_madvise.
814    
815    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
816        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
817        original change locked out all matches of empty strings. This has been
818        changed so that one match of an empty string per line is recognized.
819        Subsequent searches on the same line (for colouring or for --only-matching,
820        for example) do not recognize empty strings.
821    
822    41. Applied a user patch to fix a number of spelling mistakes in comments.
823    
824    42. Data lines longer than 65536 caused pcretest to crash.
825    
826    43. Clarified the data type for length and startoffset arguments for pcre_exec
827        and pcre_dfa_exec in the function-specific man pages, where they were
828        explicitly stated to be in bytes, never having been updated. I also added
829        some clarification to the pcreapi man page.
830    
831    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
832        a segmentation fault.
833    
834    
835    Version 8.32 30-November-2012
836    -----------------------------
837    
838    1.  Improved JIT compiler optimizations for first character search and single
839        character iterators.
840    
841    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
842        Patch by Daniel Richard G.
843    
844    3.  Single character iterator optimizations in the JIT compiler.
845    
846    4.  Improved JIT compiler optimizations for character ranges.
847    
848    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
849        Reported by Giuseppe D'Angelo.
850    
851    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
852        of a line, was being set incorrectly in cases where .* appeared inside
853        atomic brackets at the start of a pattern, or where there was a subsequent
854        *PRUNE or *SKIP.
855    
856    7.  Improved instruction cache flush for POWER/PowerPC.
857        Patch by Daniel Richard G.
858    
859    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
860        grep:
861    
862        (a) There is now no limit to the number of patterns to be matched.
863    
864        (b) An error is given if a pattern is too long.
865    
866        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
867            are now supported.
868    
869        (d) --exclude-from and --include-from (multiple use) have been added.
870    
871        (e) Exclusions and inclusions now apply to all files and directories, not
872            just to those obtained from scanning a directory recursively.
873    
874        (f) Multiple uses of -f and --file-list are now supported.
875    
876        (g) In a Windows environment, the default for -d has been changed from
877            "read" (the GNU grep default) to "skip", because otherwise the presence
878            of a directory in the file list provokes an error.
879    
880        (h) The documentation has been revised and clarified in places.
881    
882    9.  Improve the matching speed of capturing brackets.
883    
884    10. Changed the meaning of \X so that it now matches a Unicode extended
885        grapheme cluster.
886    
887    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
888        out POSIX threads when JIT support is configured.
889    
890    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
891    
892    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
893        structure could go wrong in environments where size_t is not the same size
894        as int.
895    
896    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
897        set.
898    
899    15. The EBCDIC support had decayed; later updates to the code had included
900        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
901        general tidy up of EBCDIC-related issues, and the documentation was also
902        not quite right. There is now a test that can be run on ASCII systems to
903        check some of the EBCDIC-related things (but is it not a full test).
904    
905    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
906        in a small tidy to the code.
907    
908    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
909    
910    18. If the --only-matching (-o) option in pcregrep is specified multiple
911        times, each one causes appropriate output. For example, -o1 -o2 outputs the
912        substrings matched by the 1st and 2nd capturing parentheses. A separating
913        string can be specified by --om-separator (default empty).
914    
915    19. Improving the first n character searches.
916    
917    20. Turn case lists for horizontal and vertical white space into macros so that
918        they are defined only once.
919    
920    21. This set of changes together give more compatible Unicode case-folding
921        behaviour for characters that have more than one other case when UCP
922        support is available.
923    
924        (a) The Unicode property table now has offsets into a new table of sets of
925            three or more characters that are case-equivalent. The MultiStage2.py
926            script that generates these tables (the pcre_ucd.c file) now scans
927            CaseFolding.txt instead of UnicodeData.txt for character case
928            information.
929    
930        (b) The code for adding characters or ranges of characters to a character
931            class has been abstracted into a generalized function that also handles
932            case-independence. In UTF-mode with UCP support, this uses the new data
933            to handle characters with more than one other case.
934    
935        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
936            whose other case is greater than 256 are now correctly matched
937            caselessly. Previously, the high codepoint matched the low one, but not
938            vice versa.
939    
940        (d) The processing of \h, \H, \v, and \ in character classes now makes use
941            of the new class addition function, using character lists defined as
942            macros alongside the case definitions of 20 above.
943    
944        (e) Caseless back references now work with characters that have more than
945            one other case.
946    
947        (f) General caseless matching of characters with more than one other case
948            is supported.
949    
950    22. Unicode character properties were updated from Unicode 6.2.0
951    
952    23. Improved CMake support under Windows. Patch by Daniel Richard G.
953    
954    24. Add support for 32-bit character strings, and UTF-32
955    
956    25. Major JIT compiler update (code refactoring and bugfixing).
957        Experimental Sparc 32 support is added.
958    
959    26. Applied a modified version of Daniel Richard G's patch to create
960        pcre.h.generic and config.h.generic by "make" instead of in the
961        PrepareRelease script.
962    
963    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
964        pcre_compile.c when checking for a zero character.
965    
966    28. Introducing a native interface for JIT. Through this interface, the compiled
967        machine code can be directly executed. The purpose of this interface is to
968        provide fast pattern matching, so several sanity checks are not performed.
969        However, feature tests are still performed. The new interface provides
970        1.4x speedup compared to the old one.
971    
972    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
973        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
974        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
975    
976    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
977        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
978        the "old" RFC 2279). Instead, it ended up passing a negative length to
979        pcre_exec().
980    
981    31. Add support for GCC's visibility feature to hide internal functions.
982    
983    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
984        "unknown -C option" after outputting 0 or 1.
985    
986    33. There is now support for generating a code coverage report for the test
987        suite in environments where gcc is the compiler and lcov is installed. This
988        is mainly for the benefit of the developers.
989    
990    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
991        unaddressable using valgrind annotations, allowing valgrind to detect
992        invalid memory accesses. This is mainly for the benefit of the developers.
993    
994    25. (*UTF) can now be used to start a pattern in any of the three libraries.
995    
996    26. Give configure error if --enable-cpp but no C++ compiler found.
997    
998    
999    Version 8.31 06-July-2012
1000    -------------------------
1001    
1002    1.  Fixing a wrong JIT test case and some compiler warnings.
1003    
1004    2.  Removed a bashism from the RunTest script.
1005    
1006    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1007        to unsigned type, result still unsigned" that was given by an MS compiler
1008        on encountering the code "-sizeof(xxx)".
1009    
1010    4.  Partial matching support is added to the JIT compiler.
1011    
1012    5.  Fixed several bugs concerned with partial matching of items that consist
1013        of more than one character:
1014    
1015        (a) /^(..)\1/ did not partially match "aba" because checking references was
1016            done on an "all or nothing" basis. This also applied to repeated
1017            references.
1018    
1019        (b) \R did not give a hard partial match if \r was found at the end of the
1020            subject.
1021    
1022        (c) \X did not give a hard partial match after matching one or more
1023            characters at the end of the subject.
1024    
1025        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1026            a partial match for the string "\r".
1027    
1028        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1029            a partial match for a CR character at the end of the subject string.
1030    
1031    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1032        running pcretest, the text "(JIT)" added to the output whenever JIT is
1033        actually used to run the match.
1034    
1035    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1036        or /S+[+] with a digit between 1 and 7.
1037    
1038    8.  OP_NOT now supports any UTF character not just single-byte ones.
1039    
1040    9.  (*MARK) control verb is now supported by the JIT compiler.
1041    
1042    10. The command "./RunTest list" lists the available tests without actually
1043        running any of them. (Because I keep forgetting what they all are.)
1044    
1045    11. Add PCRE_INFO_MAXLOOKBEHIND.
1046    
1047    12. Applied a (slightly modified) user-supplied patch that improves performance
1048        when the heap is used for recursion (compiled with --disable-stack-for-
1049        recursion). Instead of malloc and free for each heap frame each time a
1050        logical recursion happens, frames are retained on a chain and re-used where
1051        possible. This sometimes gives as much as 30% improvement.
1052    
1053    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1054        call.
1055    
1056    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1057    
1058    15. It is now possible to link pcretest with libedit as an alternative to
1059        libreadline.
1060    
1061    16. (*COMMIT) control verb is now supported by the JIT compiler.
1062    
1063    17. The Unicode data tables have been updated to Unicode 6.1.0.
1064    
1065    18. Added --file-list option to pcregrep.
1066    
1067    19. Added binary file support to pcregrep, including the -a, --binary-files,
1068        -I, and --text options.
1069    
1070    20. The madvise function is renamed for posix_madvise for QNX compatibility
1071        reasons. Fixed by Giuseppe D'Angelo.
1072    
1073    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1074    
1075    22. Changed the option for creating symbolic links for 16-bit man pages from
1076        -s to -sf so that re-installing does not cause issues.
1077    
1078    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1079    
1080    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1081        matches in certain environments (the workspace was not being correctly
1082        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1083        some of the workspace data at the beginning of a restart.
1084    
1085    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1086        was not doing so when it should - probably a typo introduced by SVN 528
1087        (change 8.10/14).
1088    
1089    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1090        \w+ when the character tables indicated that \x{c4} was a word character.
1091        There were several related cases, all because the tests for doing a table
1092        lookup were testing for characters less than 127 instead of 255.
1093    
1094    27. If a pattern contains capturing parentheses that are not used in a match,
1095        their slots in the ovector are set to -1. For those that are higher than
1096        any matched groups, this happens at the end of processing. In the case when
1097        there were back references that the ovector was too small to contain
1098        (causing temporary malloc'd memory to be used during matching), and the
1099        highest capturing number was not used, memory off the end of the ovector
1100        was incorrectly being set to -1. (It was using the size of the temporary
1101        memory instead of the true size.)
1102    
1103    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1104        ovector size, it uses memory at the end of the block that it has got.
1105    
1106    29. Check for an overlong MARK name and give an error at compile time. The
1107        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1108    
1109    30. JIT compiler update.
1110    
1111    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1112        Rill for the patch.
1113    
1114    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1115    
1116    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1117    
1118    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1119        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1120        was enabled.
1121    
1122    35. Improve JIT code generation for greedy plus quantifier.
1123    
1124    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1125        1 to "aa" instead of to an empty string. The bug affected repeated groups
1126        that could potentially match an empty string.
1127    
1128    37. Optimizing single character iterators in JIT.
1129    
1130    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1131        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1132        codepoints that are too big for the mode are faulted, and in a UTF mode,
1133        disallowed codepoints are also faulted.
1134    
1135    39. If PCRE was compiled with UTF support, in three places in the DFA
1136        matcher there was code that should only have been obeyed in UTF mode, but
1137        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1138        processing when bytes with values greater than 127 were present. In 16-bit
1139        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1140        both cases the values are those that cannot be the first data item in a UTF
1141        character. The three items that might have provoked this were recursions,
1142        possessively repeated groups, and atomic groups.
1143    
1144    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1145        and pcregrep, because some OS require shared objects to be explicitly
1146        passed to ld, causing the link step to fail if they are not.
1147    
1148    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1149        mode, patterns that started with \h* or \R* might be incorrectly matched.
1150    
1151    
1152    Version 8.30 04-February-2012
1153    -----------------------------
1154    
1155    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1156        name is defined in ctype.h.
1157    
1158    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1159        only in quite long subpatterns.
1160    
1161    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1162        since it was replaced by pcre_fullinfo() in February 2000.
1163    
1164    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1165        match a (*MARK), and the match failed at the start of the subject, a
1166        reference to memory before the start of the subject could occur. This bug
1167        was introduced by fix 17 of release 8.21.
1168    
1169    5.  A reference to an unset group with zero minimum repetition was giving
1170        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1171        /(another)?(\1?)test/ matched against "hello world test". This bug was
1172        introduced in release 8.13.
1173    
1174    6.  Add support for 16-bit character strings (a large amount of work involving
1175        many changes and refactorings).
1176    
1177    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1178        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1179        from a file.
1180    
1181    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1182        rounding is not applied in this particular case).
1183    
1184    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1185        if they appear, or are escaped, in patterns.
1186    
1187    10. Get rid of a number of -Wunused-but-set-variable warnings.
1188    
1189    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1190        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1191        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1192        also returns the mark "x". This bug applied to capturing parentheses,
1193        non-capturing parentheses, and atomic parentheses. It also applied to some
1194        assertions.
1195    
1196    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1197        information out of configure.ac instead of relying on pcre.h.generic, which
1198        is not stored in the repository.
1199    
1200    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1201        -lreadline.
1202    
1203    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1204    
1205    15. Applied Graycode's patch to put the top-level frame on the stack rather
1206        than the heap when not using the stack for recursion. This gives a
1207        performance improvement in many cases when recursion is not deep.
1208    
1209    16. Experimental code added to "pcretest -C" to output the stack frame size.
1210    
1211    
1212    Version 8.21 12-Dec-2011
1213    ------------------------
1214    
1215    1.  Updating the JIT compiler.
1216    
1217    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1218        are added as well.
1219    
1220    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1221        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1222        calling _pcre_jit_exec. Some extra comments are added.
1223    
1224    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1225        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1226        was introduced by change 18 for 8.20.
1227    
1228    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1229        ECMA-262 standard.
1230    
1231    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1232        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1233        This bug was probably introduced by change 9 of 8.13.
1234    
1235    7.  While fixing 6 above, I noticed that a number of other items were being
1236        incorrectly rejected as "not fixed length". This arose partly because newer
1237        opcodes had not been added to the fixed-length checking code. I have (a)
1238        corrected the bug and added tests for these items, and (b) arranged for an
1239        error to occur if an unknown opcode is encountered while checking for fixed
1240        length instead of just assuming "not fixed length". The items that were
1241        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1242        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1243        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1244    
1245    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1246        being incorrectly compiled and would have given unpredicatble results.
1247    
1248    9.  A possessively repeated subpattern with minimum repeat count greater than
1249        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1250        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1251        the first (A) could occur when it should not.
1252    
1253    10. Add a cast and remove a redundant test from the code.
1254    
1255    11. JIT should use pcre_malloc/pcre_free for allocation.
1256    
1257    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1258        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1259        is anything other than /usr, -L is still shown).
1260    
1261    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1262    
1263    14. Perl does not support \N without a following name in a [] class; PCRE now
1264        also gives an error.
1265    
1266    15. If a forward reference was repeated with an upper limit of around 2000,
1267        it caused the error "internal error: overran compiling workspace". The
1268        maximum number of forward references (including repeats) was limited by the
1269        internal workspace, and dependent on the LINK_SIZE. The code has been
1270        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1271        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1272        of around 200,000 forward references. While doing this, I also speeded up
1273        the filling in of repeated forward references.
1274    
1275    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1276        incorrectly expecting the subject to contain another "a" after the start.
1277    
1278    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1279        in the match, the SKIP should be ignored. This was not happening; instead
1280        the SKIP was being treated as NOMATCH. For patterns such as
1281        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1282        tested.
1283    
1284    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1285        now much more compatible with Perl, in particular in cases where the result
1286        is a non-match for a non-anchored pattern. For example, if
1287        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1288        "m", where previously it did not return a name. A side effect of this
1289        change is that for partial matches, the last encountered mark name is
1290        returned, as for non matches. A number of tests that were previously not
1291        Perl-compatible have been moved into the Perl-compatible test files. The
1292        refactoring has had the pleasing side effect of removing one argument from
1293        the match() function, thus reducing its stack requirements.
1294    
1295    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1296        subsequent uses of /S (without +) incorrectly behaved like /S+.
1297    
1298    21. Retrieve executable code size support for the JIT compiler and fixing
1299        some warnings.
1300    
1301    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1302        not work when the shorter character appeared right at the end of the
1303        subject string.
1304    
1305    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1306        systems.
1307    
1308    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1309        output it when the /M option is used in pcretest.
1310    
1311    25. The CheckMan script was not being included in the distribution. Also, added
1312        an explicit "perl" to run Perl scripts from the PrepareRelease script
1313        because this is reportedly needed in Windows.
1314    
1315    26. If study data was being save in a file and studying had not found a set of
1316        "starts with" bytes for the pattern, the data written to the file (though
1317        never used) was taken from uninitialized memory and so caused valgrind to
1318        complain.
1319    
1320    27. Updated RunTest.bat as provided by Sheri Pierce.
1321    
1322    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1323    
1324    29. Computation of memory usage for the table of capturing group names was
1325        giving an unnecessarily large value.
1326    
1327    
1328    Version 8.20 21-Oct-2011
1329    ------------------------
1330    
1331    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1332        a POSIX class. After further experiments with Perl, which convinced me that
1333        Perl has bugs and confusions, a closing square bracket is no longer allowed
1334        in a POSIX name. This bug also affected patterns with classes that started
1335        with full stops.
1336    
1337    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1338        captured substring, but while checking the failing first alternative,
1339        substring 1 is temporarily captured. If the output vector supplied to
1340        pcre_exec() was not big enough for this capture, the yield of the function
1341        was still zero ("insufficient space for captured substrings"). This cannot
1342        be totally fixed without adding another stack variable, which seems a lot
1343        of expense for a edge case. However, I have improved the situation in cases
1344        such as /(a)(b)x|abc/ matched against "abc", where the return code
1345        indicates that fewer than the maximum number of slots in the ovector have
1346        been set.
1347    
1348    3.  Related to (2) above: when there are more back references in a pattern than
1349        slots in the output vector, pcre_exec() uses temporary memory during
1350        matching, and copies in the captures as far as possible afterwards. It was
1351        using the entire output vector, but this conflicts with the specification
1352        that only 2/3 is used for passing back captured substrings. Now it uses
1353        only the first 2/3, for compatibility. This is, of course, another edge
1354        case.
1355    
1356    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1357        main code base, and can be used by building with --enable-jit. When this is
1358        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1359        runtime --no-jit option is given.
1360    
1361    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1362        ovector, the return from the function was zero, implying that there were
1363        other matches that did not fit. The correct "exactly full" value is now
1364        returned.
1365    
1366    6.  If a subpattern that was called recursively or as a subroutine contained
1367        (*PRUNE) or any other control that caused it to give a non-standard return,
1368        invalid errors such as "Error -26 (nested recursion at the same subject
1369        position)" or even infinite loops could occur.
1370    
1371    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1372        computing the minimum length on reaching *ACCEPT, and so ended up with the
1373        wrong value of 1 rather than 0. Further investigation indicates that
1374        computing a minimum subject length in the presence of *ACCEPT is difficult
1375        (think back references, subroutine calls), and so I have changed the code
1376        so that no minimum is registered for a pattern that contains *ACCEPT.
1377    
1378    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1379        it was not handled as intended. [But see 16 below.]
1380    
1381    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1382        Sheri Pierce.
1383    
1384    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1385        the first byte in a match must be "a".
1386    
1387    11. Change 17 for 8.13 increased the recursion depth for patterns like
1388        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1389        pattern contains any instances of (*THEN). If it does not, the old
1390        optimizations are restored. It would be nice to do this on a per-group
1391        basis, but at the moment that is not feasible.
1392    
1393    12. In some environments, the output of pcretest -C is CRLF terminated. This
1394        broke RunTest's code that checks for the link size. A single white space
1395        character after the value is now allowed for.
1396    
1397    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1398        For "fr", it uses the Windows-specific input and output files.
1399    
1400    14. If (*THEN) appeared in a group that was called recursively or as a
1401        subroutine, it did not work as intended. [But see next item.]
1402    
1403    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1404        pattern fragments (but not containing any | characters). If A and B are
1405        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1406        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1407        D. In other words, Perl considers parentheses that do not contain any |
1408        characters to be part of a surrounding alternative, whereas PCRE was
1409        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1410        differently. PCRE now behaves in the same way as Perl, except in the case
1411        of subroutine/recursion calls such as (?1) which have in any case always
1412        been different (but PCRE had them first :-).
1413    
1414    16. Related to 15 above: Perl does not treat the | in a conditional group as
1415        creating alternatives. Such a group is treated in the same way as an
1416        ordinary group without any | characters when processing (*THEN). PCRE has
1417        been changed to match Perl's behaviour.
1418    
1419    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1420        RunGrepTest script failed.
1421    
1422    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1423        inevitable for groups that contain captures, but it can lead to a lot of
1424        stack use in large patterns. The old behaviour has been restored for atomic
1425        groups that do not contain any capturing parentheses.
1426    
1427    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1428        suppress the check for a minimum subject length at run time. (If it was
1429        given to pcre_exec() or pcre_dfa_exec() it did work.)
1430    
1431    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1432        fail to work when decoding hex characters in data strings in EBCDIC
1433        environments.
1434    
1435    21. It appears that in at least one Mac OS environment, the isxdigit() function
1436        is implemented as a macro that evaluates to its argument more than once,
1437        contravening the C 90 Standard (I haven't checked a later standard). There
1438        was an instance in pcretest which caused it to go wrong when processing
1439        \x{...} escapes in subject strings. The has been rewritten to avoid using
1440        things like p++ in the argument of isxdigit().
1441    
1442    
1443    Version 8.13 16-Aug-2011
1444    ------------------------
1445    
1446    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1447    
1448    2.  Two minor typos in pcre_internal.h have been fixed.
1449    
1450    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1451        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1452        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1453    
1454    4.  There were a number of related bugs in the code for matching backrefences
1455        caselessly in UTF-8 mode when codes for the characters concerned were
1456        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1457        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1458        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1459        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1460        2-byte code at the end of the subject (it thought there wasn't enough data
1461        left).
1462    
1463    5.  Comprehensive information about what went wrong is now returned by
1464        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1465        as the output vector has at least 2 elements. The offset of the start of
1466        the failing character and a reason code are placed in the vector.
1467    
1468    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1469        now returned is for the first byte of the failing character, instead of the
1470        last byte inspected. This is an incompatible change, but I hope it is small
1471        enough not to be a problem. It makes the returned offset consistent with
1472        pcre_exec() and pcre_dfa_exec().
1473    
1474    7.  pcretest now gives a text phrase as well as the error number when
1475        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1476        failure, the offset and reason code are output.
1477    
1478    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1479        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1480        back over a single character (\n). This seems wrong (because it treated the
1481        two characters as a single entity when going forwards), conflicts with the
1482        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1483        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1484        has been changed.
1485    
1486    9.  Some internal refactoring has changed the processing so that the handling
1487        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1488        time (the PCRE_DOTALL option was changed this way some time ago: version
1489        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1490        which was always a bit of a fudge. It also means that there is one less
1491        argument for the match() function, which reduces its stack requirements
1492        slightly. This change also fixes an incompatibility with Perl: the pattern
1493        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1494    
1495    10. More internal refactoring has drastically reduced the number of recursive
1496        calls to match() for possessively repeated groups such as (abc)++ when
1497        using pcre_exec().
1498    
1499    11. While implementing 10, a number of bugs in the handling of groups were
1500        discovered and fixed:
1501    
1502        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1503        (a|)*(?1) gave a compile-time internal error.
1504        ((a|)+)+  did not notice that the outer group could match an empty string.
1505        (^a|^)+   was not marked as anchored.
1506        (.*a|.*)+ was not marked as matching at start or after a newline.
1507    
1508    12. Yet more internal refactoring has removed another argument from the match()
1509        function. Special calls to this function are now indicated by setting a
1510        value in a variable in the "match data" data block.
1511    
1512    13. Be more explicit in pcre_study() instead of relying on "default" for
1513        opcodes that mean there is no starting character; this means that when new
1514        ones are added and accidentally left out of pcre_study(), testing should
1515        pick them up.
1516    
1517    14. The -s option of pcretest has been documented for ages as being an old
1518        synonym of -m (show memory usage). I have changed it to mean "force study
1519        for every regex", that is, assume /S for every regex. This is similar to -i
1520        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1521        using it. It makes it easier to run collections of tests with and without
1522        study enabled, and thereby test pcre_study() more easily. All the standard
1523        tests are now run with and without -s (but some patterns can be marked as
1524        "never study" - see 20 below).
1525    
1526    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1527        restoration of the capturing data to the outer values was not happening
1528        correctly.
1529    
1530    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1531        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1532        pattern had matched an empty string, and so incorrectly returned a no
1533        match.
1534    
1535    17. There was optimizing code for the last branch of non-capturing parentheses,
1536        and also for the obeyed branch of a conditional subexpression, which used
1537        tail recursion to cut down on stack usage. Unfortunately, now that there is
1538        the possibility of (*THEN) occurring in these branches, tail recursion is
1539        no longer possible because the return has to be checked for (*THEN). These
1540        two optimizations have therefore been removed. [But see 8.20/11 above.]
1541    
1542    18. If a pattern containing \R was studied, it was assumed that \R always
1543        matched two bytes, thus causing the minimum subject length to be
1544        incorrectly computed because \R can also match just one byte.
1545    
1546    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1547        was incorrectly computed.
1548    
1549    20. If /S is present twice on a test pattern in pcretest input, it now
1550        *disables* studying, thereby overriding the use of -s on the command line
1551        (see 14 above). This is necessary for one or two tests to keep the output
1552        identical in both cases.
1553    
1554    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1555        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1556    
1557    22. When an atomic group that contained a capturing parenthesis was
1558        successfully matched, but the branch in which it appeared failed, the
1559        capturing was not being forgotten if a higher numbered group was later
1560        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1561        group 1 to "a", when in fact it should be unset. This applied to multi-
1562        branched capturing and non-capturing groups, repeated or not, and also to
1563        positive assertions (capturing in negative assertions does not happen
1564        in PCRE) and also to nested atomic groups.
1565    
1566    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1567        subject after a captured substring, to make it easier to tell which of a
1568        number of identical substrings has been captured.
1569    
1570    24. The way atomic groups are processed by pcre_exec() has been changed so that
1571        if they are repeated, backtracking one repetition now resets captured
1572        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1573        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1574        "aaa". Previously, it would have been "a". As part of this code
1575        refactoring, the way recursive calls are handled has also been changed.
1576    
1577    25. If an assertion condition captured any substrings, they were not passed
1578        back unless some other capturing happened later. For example, if
1579        (?(?=(a))a) was matched against "a", no capturing was returned.
1580    
1581    26. When studying a pattern that contained subroutine calls or assertions,
1582        the code for finding the minimum length of a possible match was handling
1583        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1584        group 1 called group 2 while simultaneously a separate group 2 called group
1585        1). A stack overflow occurred in this case. I have fixed this by limiting
1586        the recursion depth to 10.
1587    
1588    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1589        Fortmann. This supports explicit test numbers on the command line, and has
1590        argument validation and error reporting.
1591    
1592    28. An instance of \X with an unlimited repeat could fail if at any point the
1593        first character it looked at was a mark character.
1594    
1595    29. Some minor code refactoring concerning Unicode properties and scripts
1596        should reduce the stack requirement of match() slightly.
1597    
1598    30. Added the '=' option to pcretest to check the setting of unused capturing
1599        slots at the end of the pattern, which are documented as being -1, but are
1600        not included in the return count.
1601    
1602    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1603        compiled something random. Now it gives a compile-time error (as does
1604        Perl).
1605    
1606    32. A *MARK encountered during the processing of a positive assertion is now
1607        recorded and passed back (compatible with Perl).
1608    
1609    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1610        had alternative anchored branches, the search for a second match in a line
1611        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1612        matched the line "0102" twice. The same bug affected patterns that started
1613        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1614        twice.
1615    
1616    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1617        does, and because of capturing effects, quantifying parenthesized
1618        assertions may at times be useful. Quantifiers are now allowed for
1619        parenthesized assertions.
1620    
1621    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1622    
1623    36. \g was being checked for fancy things in a character class, when it should
1624        just be a literal "g".
1625    
1626    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1627        appearance of a nested POSIX class supersedes an apparent external class.
1628        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1629        unescaped square brackets may also appear as part of class names. For
1630        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1631        more like Perl. (But see 8.20/1 above.)
1632    
1633    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1634        was because it thought it was \N{name}, which is not supported).
1635    
1636    39. Add minix to OS list not supporting the -S option in pcretest.
1637    
1638    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1639        cannot analyze patterns in sufficient detail to catch mutual recursions
1640        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1641        subgroup is called recursively as a subpattern for a second time at the
1642        same position in the subject string. In previous releases this might have
1643        been caught by the recursion limit, or it might have run out of stack.
1644    
1645    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1646        happen only once. PCRE was, however incorrectly giving a compile time error
1647        "recursive call could loop indefinitely" because it cannot analyze the
1648        pattern in sufficient detail. The compile time test no longer happens when
1649        PCRE is compiling a conditional subpattern, but actual runaway loops are
1650        now caught at runtime (see 40 above).
1651    
1652    42. It seems that Perl allows any characters other than a closing parenthesis
1653        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1654        has been changed to be the same.
1655    
1656    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1657        as not to get warnings when autogen.sh is called. Also changed
1658        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1659    
1660    44. To help people who use pcregrep to scan files containing exceedingly long
1661        lines, the following changes have been made:
1662    
1663        (a) The default value of the buffer size parameter has been increased from
1664            8K to 20K. (The actual buffer used is three times this size.)
1665    
1666        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1667            PCRE is built.
1668    
1669        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1670            to be set at run time.
1671    
1672        (d) Numerical values in pcregrep options can be followed by K or M, for
1673            example --buffer-size=50K.
1674    
1675        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1676            given and the return code is set to 2.
1677    
1678    45. Add a pointer to the latest mark to the callout data block.
1679    
1680    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1681        partial match of an empty string instead of no match. This was specific to
1682        the use of ".".
1683    
1684    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1685        complete match instead of a partial match. This bug was dependent on both
1686        the PCRE_UTF8 and PCRE_DOTALL options being set.
1687    
1688    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1689        starting byte set, because \b was not being ignored.
1690    
1691    
1692    Version 8.12 15-Jan-2011
1693    ------------------------
1694    
1695    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1696        checks for such things as part of the documentation building process.
1697    
1698    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1699        --match-limit and --recursion-limit options (added for 8.11). In
1700        particular, this made one of the standard tests fail. (The integer value
1701        went into the wrong half of a long int.)
1702    
1703    3.  If the --colour option was given to pcregrep with -v (invert match), it
1704        did strange things, either producing crazy output, or crashing. It should,
1705        of course, ignore a request for colour when reporting lines that do not
1706        match.
1707    
1708    4.  Another pcregrep bug caused similar problems if --colour was specified with
1709        -M (multiline) and the pattern match finished with a line ending.
1710    
1711    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1712        matched in multiline mode, the following line was shown as part of the
1713        match. This seems wrong, so I have changed it.
1714    
1715    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1716        the check for further matches in the same line (so they could be coloured)
1717        to overrun the end of the current line. If another match was found, it was
1718        incorrectly shown (and then shown again when found in the next line).
1719    
1720    7.  If pcregrep was compiled under Windows, there was a reference to the
1721        function pcregrep_exit() before it was defined. I am assuming this was
1722        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1723        reported by a user. I've moved the definition above the reference.
1724    
1725    
1726    Version 8.11 10-Dec-2010
1727    ------------------------
1728    
1729    1.  (*THEN) was not working properly if there were untried alternatives prior
1730        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1731        backtracked to try for "b" instead of moving to the next alternative branch
1732        at the same level (in this case, to look for "c"). The Perl documentation
1733        is clear that when (*THEN) is backtracked onto, it goes to the "next
1734        alternative in the innermost enclosing group".
1735    
1736    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1737        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1738        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1739        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1740        (*THEN).
1741    
1742    3.  If \s appeared in a character class, it removed the VT character from
1743        the class, even if it had been included by some previous item, for example
1744        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1745        of \s, but is part of the POSIX "space" class.)
1746    
1747    4.  A partial match never returns an empty string (because you can always
1748        match an empty string at the end of the subject); however the checking for
1749        an empty string was starting at the "start of match" point. This has been
1750        changed to the "earliest inspected character" point, because the returned
1751        data for a partial match starts at this character. This means that, for
1752        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1753        (previously it gave "no match").
1754    
1755    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1756        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1757        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1758        has an implication that the given string is incomplete (because a partial
1759        match is preferred over a full match). For this reason, these items now
1760        give a partial match in this situation. [Aside: previously, the one case
1761        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1762        match rather than a full match, which was wrong by the old rules, but is
1763        now correct.]
1764    
1765    6.  There was a bug in the handling of #-introduced comments, recognized when
1766        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1767        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1768        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1769        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1770        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1771        places in pcre_compile().
1772    
1773    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1774        comments when looking ahead for named forward references to subpatterns,
1775        the only newline sequence it recognized was NL. It now handles newlines
1776        according to the set newline convention.
1777    
1778    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1779        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1780        cater for a lack of strerror(). These oversights have been fixed.
1781    
1782    9.  Added --match-limit and --recursion-limit to pcregrep.
1783    
1784    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1785    
1786    11. When the -o option was used, pcregrep was setting a return code of 1, even
1787        when matches were found, and --line-buffered was not being honoured.
1788    
1789    12. Added an optional parentheses number to the -o and --only-matching options
1790        of pcregrep.
1791    
1792    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1793        can match an empty string. The code to do it in pcretest and pcredemo
1794        needed fixing:
1795    
1796        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1797            only one byte after an empty string match just before CRLF (this case
1798            just got forgotten; "any" and "anycrlf" were OK).
1799    
1800        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1801            mode when an empty string match preceded an ASCII character followed by
1802            a non-ASCII character. (The code for advancing by one character rather
1803            than one byte was nonsense.)
1804    
1805        (c) The pcredemo.c sample program did not have any code at all to handle
1806            the cases when CRLF is a valid newline sequence.
1807    
1808    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1809        as a starting offset was within the subject string. There is now a new
1810        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1811        negative or greater than the length of the string. In order to test this,
1812        pcretest is extended to allow the setting of negative starting offsets.
1813    
1814    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1815        starting offset points to the beginning of a UTF-8 character was
1816        unnecessarily clumsy. I tidied it up.
1817    
1818    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1819        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1820    
1821    17. Nobody had reported that the --include_dir option, which was added in
1822        release 7.7 should have been called --include-dir (hyphen, not underscore)
1823        for compatibility with GNU grep. I have changed it to --include-dir, but
1824        left --include_dir as an undocumented synonym, and the same for
1825        --exclude-dir, though that is not available in GNU grep, at least as of
1826        release 2.5.4.
1827    
1828    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1829        characters from a string of bytes) have been redefined so as not to use
1830        loops, in order to improve performance in some environments. At the same
1831        time, I abstracted some of the common code into auxiliary macros to save
1832        repetition (this should not affect the compiled code).
1833    
1834    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1835        compile-time error is now given if \c is not followed by an ASCII
1836        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1837        different, and any byte value is allowed.)
1838    
1839    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1840        START_OPTIMIZE option, which is now allowed at compile time - but just
1841        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1842        to pcregrep and other applications that have no direct access to PCRE
1843        options. The new /Y option in pcretest sets this option when calling
1844        pcre_compile().
1845    
1846    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1847        back references. Groups containing recursive back references were forced to
1848        be atomic by that change, but in the case of named groups, the amount of
1849        memory required was incorrectly computed, leading to "Failed: internal
1850        error: code overflow". This has been fixed.
1851    
1852    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1853        pcretest.c, to avoid build problems in some Borland environments.
1854    
1855    
1856    Version 8.10 25-Jun-2010
1857    ------------------------
1858    
1859    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1860        THEN.
1861    
1862    2.  (*ACCEPT) was not working when inside an atomic group.
1863    
1864    3.  Inside a character class, \B is treated as a literal by default, but
1865        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1866        causes the error). The code is unchanged, but I tidied the documentation.
1867    
1868    4.  Inside a character class, PCRE always treated \R and \X as literals,
1869        whereas Perl faults them if its -w option is set. I have changed PCRE so
1870        that it faults them when PCRE_EXTRA is set.
1871    
1872    5.  Added support for \N, which always matches any character other than
1873        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1874    
1875    6.  When compiling pcregrep with newer versions of gcc which may have
1876        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1877        declared with attribute warn_unused_result" were given. Just casting the
1878        result to (void) does not stop the warnings; a more elaborate fudge is
1879        needed. I've used a macro to implement this.
1880    
1881    7.  Minor change to pcretest.c to avoid a compiler warning.
1882    
1883    8.  Added four artifical Unicode properties to help with an option to make
1884        \s etc use properties (see next item). The new properties are: Xan
1885        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1886    
1887    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1888        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1889        this option. Modified pcretest to add /W to test this facility. Added
1890        REG_UCP to make it available via the POSIX interface.
1891    
1892    10. Added --line-buffered to pcregrep.
1893    
1894    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1895        studied, and the match started with a letter with a code point greater than
1896        127 whose first byte was different to the first byte of the other case of
1897        the letter, the other case of this starting letter was not recognized
1898        (#976).
1899    
1900    12. If a pattern that was studied started with a repeated Unicode property
1901        test, for example, \p{Nd}+, there was the theoretical possibility of
1902        setting up an incorrect bitmap of starting bytes, but fortunately it could
1903        not have actually happened in practice until change 8 above was made (it
1904        added property types that matched character-matching opcodes).
1905    
1906    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1907        possible starting bytes for non-anchored patterns.
1908    
1909    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1910        \R, and also a number of cases that involve Unicode properties, both
1911        explicit and implicit when PCRE_UCP is set.
1912    
1913    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1914        input, it could crash or give wrong results if characters with values
1915        greater than 0xc0 were present in the subject string. (Detail: it assumed
1916        UTF-8 input when processing these items.)
1917    
1918    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1919        size_t is 64-bit (#991).
1920    
1921    17. Added a check for running out of memory when PCRE is compiled with
1922        --disable-stack-for-recursion (#990).
1923    
1924    18. If the last data line in a file for pcretest does not have a newline on
1925        the end, a newline was missing in the output.
1926    
1927    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1928        less than 128) in its various bitmaps. However, there is a facility for
1929        generating tables according to the current locale when PCRE is compiled. It
1930        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1931        characters, are recognized by isspace() and therefore were getting set in
1932        these tables, and indeed these tables seem to approximate to ISO 8859. This
1933        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1934        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1935        which of course cannot start UTF-8 characters. I have changed the code so
1936        that only real ASCII characters (less than 128) and the correct starting
1937        bytes for UTF-8 encodings are set for characters greater than 127 when in
1938        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1939        altogether.)
1940    
1941    20. Added the /T option to pcretest so as to be able to run tests with non-
1942        standard character tables, thus making it possible to include the tests
1943        used for 19 above in the standard set of tests.
1944    
1945    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1946        reference to a subpattern the other side of a comment that contains an
1947        opening parenthesis caused either an internal compiling error, or a
1948        reference to the wrong subpattern.
1949    
1950    
1951    Version 8.02 19-Mar-2010
1952    ------------------------
1953    
1954    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1955    
1956    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1957        configured.
1958    
1959    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1960        original author of that file, following a query about its status.
1961    
1962    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1963        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1964    
1965    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1966        quantifier applied to a forward-referencing subroutine call, could compile
1967        incorrect code or give the error "internal error: previously-checked
1968        referenced subpattern not found".
1969    
1970    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1971        variables to point to external functions. For these systems, therefore,
1972        pcre_malloc etc. are now initialized to local functions that call the
1973        relevant global functions.
1974    
1975    7.  There were two entries missing in the vectors called coptable and poptable
1976        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1977        I've fixed the data, and added a kludgy way of testing at compile time that
1978        the lengths are correct (equal to the number of opcodes).
1979    
1980    8.  Following on from 7, I added a similar kludge to check the length of the
1981        eint vector in pcreposix.c.
1982    
1983    9.  Error texts for pcre_compile() are held as one long string to avoid too
1984        much relocation at load time. To find a text, the string is searched,
1985        counting zeros. There was no check for running off the end of the string,
1986        which could happen if a new error number was added without updating the
1987        string.
1988    
1989    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1990    
1991    11. \K was not working if it appeared in an atomic group or in a group that
1992        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1993        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1994        the assertion is positive, but not if it is negative.
1995    
1996    12. Change 11 fortuitously reduced the size of the stack frame used in the
1997        "match()" function of pcre_exec.c by one pointer. Forthcoming
1998        implementation of support for (*MARK) will need an extra pointer on the
1999        stack; I have reserved it now, so that the stack frame size does not
2000        decrease.
2001    
2002    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2003        item in branch that calls a recursion is a subroutine call - as in the
2004        second branch in the above example - was incorrectly given the compile-
2005        time error "recursive call could loop indefinitely" because pcre_compile()
2006        was not correctly checking the subroutine for matching a non-empty string.
2007    
2008    14. The checks for overrunning compiling workspace could trigger after an
2009        overrun had occurred. This is a "should never occur" error, but it can be
2010        triggered by pathological patterns such as hundreds of nested parentheses.
2011        The checks now trigger 100 bytes before the end of the workspace.
2012    
2013    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2014    
2015    
2016    Version 8.01 19-Jan-2010
2017    ------------------------
2018    
2019    1.  If a pattern contained a conditional subpattern with only one branch (in
2020        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2021        computed the wrong minimum data length (which is of course zero for such
2022        subpatterns). This could cause incorrect "no match" results.
2023    
2024    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2025        the pattern is reset in the first branch, pcre_compile() failed with
2026        "internal error: code overflow at offset...". This happened only when
2027        the reset was to the original external option setting. (An optimization
2028        abstracts leading options settings into an external setting, which was the
2029        cause of this.)
2030    
2031    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2032        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2033        assertion pattern did not match (meaning that the assertion was true), it
2034        was incorrectly treated as false if the SKIP had been reached during the
2035        matching. This also applied to assertions used as conditions.
2036    
2037    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2038        assertion subpattern, including such a pattern used as a condition,
2039        unpredictable results occurred, instead of the error return
2040        PCRE_ERROR_DFA_UITEM.
2041    
2042    5.  The C++ GlobalReplace function was not working like Perl for the special
2043        situation when an empty string is matched. It now does the fancy magic
2044        stuff that is necessary.
2045    
2046    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2047        removed. (These were left over from very, very early versions of PCRE.)
2048    
2049    7.  Some cosmetic changes to the code to make life easier when compiling it
2050        as part of something else:
2051    
2052        (a) Change DEBUG to PCRE_DEBUG.
2053    
2054        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2055            called "current" as "current_branch", to prevent a collision with the
2056            Linux macro when compiled as a kernel module.
2057    
2058        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2059            prevent a collision with the Linux macro when compiled as a kernel
2060            module.
2061    
2062    8.  In pcre_compile() there are some checks for integer overflows that used to
2063        cast potentially large values to (double). This has been changed to that
2064        when building, a check for int64_t is made, and if it is found, it is used
2065        instead, thus avoiding the use of floating point arithmetic. (There is no
2066        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2067        double.
2068    
2069    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2070        2005 (difference between two addresses compared to an unsigned value).
2071    
2072    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2073        custom one, because of the following reported problem in Windows:
2074    
2075          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2076              under Win32.
2077          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2078              therefore missing the function definition.
2079          - The compiler thus generates a "C" signature for the test function.
2080          - The linker fails to find the "C" function.
2081          - PCRE fails to configure if asked to do so against libbz2.
2082    
2083    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2084        messages were output:
2085    
2086          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2087          rerunning libtoolize, to keep the correct libtool macros in-tree.
2088          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2089    
2090        I have done both of these things.
2091    
2092    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2093        most of the time, it *can* run out if it is given a pattern that contains a
2094        runaway infinite recursion. I updated the discussion in the pcrestack man
2095        page.
2096    
2097    13. Now that we have gone to the x.xx style of version numbers, the minor
2098        version may start with zero. Using 08 or 09 is a bad idea because users
2099        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2100        interpreted as invalid octal numbers. I've updated the previous comment in
2101        configure.ac, and also added a check that gives an error if 08 or 09 are
2102        used.
2103    
2104    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2105        causing partial matching to fail when the end of the subject matched \W
2106        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2107    
2108    15. There were some discrepancies between the declarations in pcre_internal.h
2109        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2110        their definitions. The declarations used "const uschar *" and the
2111        definitions used USPTR. Even though USPTR is normally defined as "const
2112        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2113        reported that: "This difference in casting confuses some C++ compilers, for
2114        example, SunCC recognizes above declarations as different functions and
2115        generates broken code for hbpcre." I have changed the declarations to use
2116        USPTR.
2117    
2118    16. GNU libtool is named differently on some systems. The autogen.sh script now
2119        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2120        (FreeBSD).
2121    
2122    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2123        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2124        comment: "Figure out how to create a longlong from a string: strtoll and
2125        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2126        instance, but it only takes 2 args instead of 3!"
2127    
2128    18. A subtle bug concerned with back references has been fixed by a change of
2129        specification, with a corresponding code fix. A pattern such as
2130        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2131        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2132        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2133        same bug. Such groups have to be quantified to be useful, or contained
2134        inside another quantified group. (If there's no repetition, the reference
2135        can never match.) The problem arises because, having left the group and
2136        moved on to the rest of the pattern, a later failure that backtracks into
2137        the group uses the captured value from the final iteration of the group
2138        rather than the correct earlier one. I have fixed this in PCRE by forcing
2139        any group that contains a reference to itself to be an atomic group; that
2140        is, there cannot be any backtracking into it once it has completed. This is
2141        similar to recursive and subroutine calls.
2142    
2143    
2144    Version 8.00 19-Oct-09
2145    ----------------------
2146    
2147    1.  The table for translating pcre_compile() error codes into POSIX error codes
2148        was out-of-date, and there was no check on the pcre_compile() error code
2149        being within the table. This could lead to an OK return being given in
2150        error.
2151    
2152    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2153        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2154        in a Windows environment.
2155    
2156    3.  The pcregrep --count option prints the count for each file even when it is
2157        zero, as does GNU grep. However, pcregrep was also printing all files when
2158        --files-with-matches was added. Now, when both options are given, it prints
2159        counts only for those files that have at least one match. (GNU grep just
2160        prints the file name in this circumstance, but including the count seems
2161        more useful - otherwise, why use --count?) Also ensured that the
2162        combination -clh just lists non-zero counts, with no names.
2163    
2164    4.  The long form of the pcregrep -F option was incorrectly implemented as
2165        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2166        but it seems right to fix it, and I didn't think it was worth preserving
2167        the old behaviour.
2168    
2169    5.  The command line items --regex=pattern and --regexp=pattern were not
2170        recognized by pcregrep, which required --regex pattern or --regexp pattern
2171        (with a space rather than an '='). The man page documented the '=' forms,
2172        which are compatible with GNU grep; these now work.
2173    
2174    6.  No libpcreposix.pc file was created for pkg-config; there was just
2175        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2176    
2177    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2178        when UCP support is not needed, by modifying the Python script that
2179        generates it from Unicode data files. This should not matter if the module
2180        is correctly used as a library, but I received one complaint about 50K of
2181        unwanted data. My guess is that the person linked everything into his
2182        program rather than using a library. Anyway, it does no harm.
2183    
2184    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2185        was a minimum greater than 1 for a wide character in a possessive
2186        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2187        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2188        character. Chaos in the form of incorrect output or a compiling loop could
2189        result.
2190    
2191    9.  The restrictions on what a pattern can contain when partial matching is
2192        requested for pcre_exec() have been removed. All patterns can now be
2193        partially matched by this function. In addition, if there are at least two
2194        slots in the offset vector, the offset of the earliest inspected character
2195        for the match and the offset of the end of the subject are set in them when
2196        PCRE_ERROR_PARTIAL is returned.
2197    
2198    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2199        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2200        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2201        and may be more useful for multi-segment matching.
2202    
2203    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2204        used to be given if ever the end of the subject was reached; now it is
2205        given only if matching could not proceed because another character was
2206        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2207        string "Z", which now yields "no match" instead of "partial match". In the
2208        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2209        final character ended with (*FAIL).
2210    
2211    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2212        if the pattern had a "must contain" character that was already found in the
2213        earlier partial match, unless partial matching was again requested. For
2214        example, with the pattern /dog.(body)?/, the "must contain" character is
2215        "g". If the first part-match was for the string "dog", restarting with
2216        "sbody" failed. This bug has been fixed.
2217    
2218    13. The string returned by pcre_dfa_exec() after a partial match has been
2219        changed so that it starts at the first inspected character rather than the
2220        first character of the match. This makes a difference only if the pattern
2221        starts with a lookbehind assertion or \b or \B (\K is not supported by
2222        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2223        matching functions compatible, and I think it's the right thing to do.
2224    
2225    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2226        so that the demonstration program is easily available in environments where
2227        PCRE has not been installed from source.
2228    
2229    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2230        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2231        library.
2232    
2233    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2234        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2235        is not the first non-POSIX option to be added. Clearly some people find
2236        these options useful.
2237    
2238    17. If a caller to the POSIX matching function regexec() passes a non-zero
2239        value for nmatch with a NULL value for pmatch, the value of
2240        nmatch is forced to zero.
2241    
2242    18. RunGrepTest did not have a test for the availability of the -u option of
2243        the diff command, as RunTest does. It now checks in the same way as
2244        RunTest, and also checks for the -b option.
2245    
2246    19. If an odd number of negated classes containing just a single character
2247        interposed, within parentheses, between a forward reference to a named
2248        subpattern and the definition of the subpattern, compilation crashed with
2249        an internal error, complaining that it could not find the referenced
2250        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2251        [The bug was that it was starting one character too far in when skipping
2252        over the character class, thus treating the ] as data rather than
2253        terminating the class. This meant it could skip too much.]
2254    
2255    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2256        /g option in pcretest when the pattern contains \K, which makes it possible
2257        to have an empty string match not at the start, even when the pattern is
2258        anchored. Updated pcretest and pcredemo to use this option.
2259    
2260    21. If the maximum number of capturing subpatterns in a recursion was greater
2261        than the maximum at the outer level, the higher number was returned, but
2262        with unset values at the outer level. The correct (outer level) value is
2263        now given.
2264    
2265    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2266        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2267        make it do so. The string so far is captured, making this feature
2268        compatible with Perl.
2269    
2270    23. The tests have been re-organized, adding tests 11 and 12, to make it
2271        possible to check the Perl 5.10 features against Perl 5.10.
2272    
2273    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2274        pattern matches a fixed length string. PCRE did not allow this; now it
2275        does. Neither allows recursion.
2276    
2277    25. I finally figured out how to implement a request to provide the minimum
2278        length of subject string that was needed in order to match a given pattern.
2279        (It was back references and recursion that I had previously got hung up
2280        on.) This code has now been added to pcre_study(); it finds a lower bound
2281        to the length of subject needed. It is not necessarily the greatest lower
2282        bound, but using it to avoid searching strings that are too short does give
2283        some useful speed-ups. The value is available to calling programs via
2284        pcre_fullinfo().
2285    
2286    26. While implementing 25, I discovered to my embarrassment that pcretest had
2287        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2288        study optimizations had never been tested with that matching function.
2289        Oops. What is worse, even when it was passed study data, there was a bug in
2290        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2291        were also very few tests of studied patterns with pcre_dfa_exec().
2292    
2293    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2294        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2295        on the other side of the coin, they are no longer allowed to have different
2296        names, because these cannot be distinguished in PCRE, and this has caused
2297        confusion. (This is a difference from Perl.)
2298    
2299    28. When duplicate subpattern names are present (necessarily with different
2300        numbers, as required by 27 above), and a test is made by name in a
2301        conditional pattern, either for a subpattern having been matched, or for
2302        recursion in such a pattern, all the associated numbered subpatterns are
2303        tested, and the overall condition is true if the condition is true for any
2304        one of them. This is the way Perl works, and is also more like the way
2305        testing by number works.
2306    
2307    
2308    Version 7.9 11-Apr-09
2309    ---------------------
2310    
2311    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2312        (pcretest), all targets were linked against these libraries. This included
2313        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2314        libraries. This caused unwanted dependencies to be created. This problem
2315        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2316        pcretest is linked with readline.
2317    
2318    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2319        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2320        moved outside it again, because FALSE and TRUE are already defined in AIX,
2321        but BOOL is not.
2322    
2323    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2324        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2325    
2326    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2327        hyphens) following file names and line numbers when outputting matching
2328        lines. This is not true; no spaces are inserted. I have also clarified the
2329        wording for the --colour (or --color) option.
2330    
2331    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2332        was not coloured; this is different to GNU grep, so I have changed it to be
2333        the same.
2334    
2335    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2336        each matching line was coloured. Now it goes on to look for further matches
2337        of any of the test patterns, which is the same behaviour as GNU grep.
2338    
2339    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2340        doesn't make sense to accept an empty string match in pcregrep, so I have
2341        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2342        seems to be how GNU grep behaves. [But see later change 40 for release
2343        8.33.]
2344    
2345    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2346        start or after a newline", because the conditional assertion was not being
2347        correctly handled. The rule now is that both the assertion and what follows
2348        in the first alternative must satisfy the test.
2349    
2350    9.  If auto-callout was enabled in a pattern with a conditional group whose
2351        condition was an assertion, PCRE could crash during matching, both with
2352        pcre_exec() and pcre_dfa_exec().
2353    
2354    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2355        used for matching.
2356    
2357    11. Unicode property support in character classes was not working for
2358        characters (bytes) greater than 127 when not in UTF-8 mode.
2359    
2360    12. Added the -M command line option to pcretest.
2361    
2362    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2363    
2364    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2365    
2366    16. Added comments and documentation about mis-use of no_arg in the C++
2367        wrapper.
2368    
2369    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2370        from Martin Jerabek that uses macro names for all relevant character and
2371        string constants.
2372    
2373    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2374        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2375        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2376        these, but not everybody uses configure.
2377    
2378    19. A conditional group that had only one branch was not being correctly
2379        recognized as an item that could match an empty string. This meant that an
2380        enclosing group might also not be so recognized, causing infinite looping
2381        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2382        with the subject "ab", where knowledge that the repeated group can match
2383        nothing is needed in order to break the loop.
2384    
2385    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2386        exec(), but without supplying a callout function, matching went wrong.
2387    
2388    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2389        leak if the size of the offset vector was greater than 30. When the vector
2390        is smaller, the saved offsets during recursion go onto a local stack
2391        vector, but for larger vectors malloc() is used. It was failing to free
2392        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2393        error, in fact).
2394    
2395    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2396        heapframe that is used only when UTF-8 support is enabled. This caused no
2397        problem, but was untidy.
2398    
2399    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2400        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2401        included within another project.
2402    
2403    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2404        slightly modified by me:
2405    
2406          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2407              not building pcregrep.
2408    
2409          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2410              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2411    
2412    25. Forward references, both numeric and by name, in patterns that made use of
2413        duplicate group numbers, could behave incorrectly or give incorrect errors,
2414        because when scanning forward to find the reference group, PCRE was not
2415        taking into account the duplicate group numbers. A pattern such as
2416        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2417    
2418    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2419        the feature of a custom pointer more persuasive (as requested by a user).
2420    
2421    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2422        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2423        pre-defined.
2424    
2425    28. Added support for (*UTF8) at the start of a pattern.
2426    
2427    29. Arrange for flags added by the "release type" setting in CMake to be shown
2428        in the configuration summary.
2429    
2430    
2431    Version 7.8 05-Sep-08
2432    ---------------------
2433    
2434    1.  Replaced UCP searching code with optimized version as implemented for Ad
2435        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2436        stage table and inline lookup instead of a function, giving speed ups of 2
2437        to 5 times on some simple patterns that I tested. Permission was given to
2438        distribute the MultiStage2.py script that generates the tables (it's not in
2439        the tarball, but is in the Subversion repository).
2440    
2441    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2442        scripts.
2443    
2444    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2445        a group with a zero qualifier. The result of the study could be incorrect,
2446        or the function might crash, depending on the pattern.
2447    
2448    4.  Caseless matching was not working for non-ASCII characters in back
2449        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2450        It now works when Unicode Property Support is available.
2451    
2452    5.  In pcretest, an escape such as \x{de} in the data was always generating
2453        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2454        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2455        truncation.
2456    
2457    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2458    
2459    7.  Added two (int) casts to pcregrep when printing the difference of two
2460        pointers, in case they are 64-bit values.
2461    
2462    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2463        test 2 if it fails.
2464    
2465    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2466        and a #define of that name to empty if it is not externally set. This is to
2467        allow users of MSVC to set it if necessary.
2468    
2469    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2470        the convenience functions in the pcre_get.c source file.
2471    
2472    11. An option change at the start of a pattern that had top-level alternatives
2473        could cause overwriting and/or a crash. This command provoked a crash in
2474        some environments:
2475    
2476          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2477    
2478        This potential security problem was recorded as CVE-2008-2371.
2479    
2480    12. For a pattern where the match had to start at the beginning or immediately
2481        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2482        pcre_dfa_exec() could read past the end of the passed subject if there was
2483        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2484        pcretest so that it places the subject at the end of its malloc-ed buffer.
2485    
2486    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2487        exec() might read past the end of the data buffer in UTF-8 mode.
2488    
2489    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2490        the data contained the byte 0x85 as part of a UTF-8 character within its
2491        first line. This applied both to normal and DFA matching.
2492    
2493    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2494        /^[^d]*?$/8 failed to match "abc".
2495    
2496    16. Added a missing copyright notice to pcrecpp_internal.h.
2497    
2498    17. Make it more clear in the documentation that values returned from
2499        pcre_exec() in ovector are byte offsets, not character counts.
2500    
2501    18. Tidied a few places to stop certain compilers from issuing warnings.
2502    
2503    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2504        supplied by Stefan Weber. I made a further small update for 7.8 because
2505        there is a change of source arrangements: the pcre_searchfuncs.c module is
2506        replaced by pcre_ucd.c.
2507    
2508    
2509    Version 7.7 07-May-08
2510    ---------------------
2511    
2512    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2513        a string to a long long, pretend we don't even have a long long." This is
2514        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2515    
2516    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2517        pre-7.6 versions, which defined a global no_arg variable instead of putting
2518        it in the RE class. (See also #8 below.)
2519    
2520    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2521        Lopes.
2522    
2523    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2524    
2525        (1) The include/exclude patterns were being applied to the whole pathnames
2526            of files, instead of just to the final components.
2527    
2528        (2) If there was more than one level of directory, the subdirectories were
2529            skipped unless they satisfied the include/exclude conditions. This is
2530            inconsistent with GNU grep (and could even be seen as contrary to the
2531            pcregrep specification - which I improved to make it absolutely clear).
2532            The action now is always to scan all levels of directory, and just
2533            apply the include/exclude patterns to regular files.
2534    
2535    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2536        --exclude_dir in the tests to avoid scanning .svn directories.
2537    
2538    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2539        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2540        doesn't support NULs in patterns.
2541    
2542    7.  Added some missing "const"s to declarations of static tables in
2543        pcre_compile.c and pcre_dfa_exec.c.
2544    
2545    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2546        caused by fix #2  above. (Subsequently also a second patch to fix the
2547        first patch. And a third patch - this was a messy problem.)
2548    
2549    9.  Applied Craig's patch to remove the use of push_back().
2550    
2551    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2552        matching function regexec().
2553    
2554    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2555        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2556        references. PCRE supports relative numbers with this syntax (I don't think
2557        Oniguruma does).
2558    
2559    12. Previously, a group with a zero repeat such as (...){0} was completely
2560        omitted from the compiled regex. However, this means that if the group
2561        was called as a subroutine from elsewhere in the pattern, things went wrong
2562        (an internal error was given). Such groups are now left in the compiled
2563        pattern, with a new opcode that causes them to be skipped at execution
2564        time.
2565    
2566    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2567        to the way PCRE behaves:
2568    
2569        (a) A lone ] character is dis-allowed (Perl treats it as data).
2570    
2571        (b) A back reference to an unmatched subpattern matches an empty string
2572            (Perl fails the current match path).
2573    
2574        (c) A data ] in a character class must be notated as \] because if the
2575            first data character in a class is ], it defines an empty class. (In
2576            Perl it is not possible to have an empty class.) The empty class []
2577            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2578            The negative empty class [^] matches any one character, independently
2579            of the DOTALL setting.
2580    
2581    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2582        non-existent subpattern following a character class starting with ']' and
2583        containing () gave an internal compiling error instead of "reference to
2584        non-existent subpattern". Fortunately, when the pattern did exist, the
2585        compiled code was correct. (When scanning forwards to check for the
2586        existence of the subpattern, it was treating the data ']' as terminating
2587        the class, so got the count wrong. When actually compiling, the reference
2588        was subsequently set up correctly.)
2589    
2590    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2591        it was being rejected as not supported by pcre_dfa_exec(), even though
2592        other assertions are supported. I have made pcre_dfa_exec() support
2593        (*FAIL).
2594    
2595    16. The implementation of 13c above involved the invention of a new opcode,
2596        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2597        cannot be changed at match time, I realized I could make a small
2598        improvement to matching performance by compiling OP_ALLANY instead of
2599        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2600        on the OP_ANY path.
2601    
2602    17. Compiling pcretest on Windows with readline support failed without the
2603        following two fixes: (1) Make the unistd.h include conditional on
2604        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2605    
2606    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2607        ncurses library to be included for pcretest when ReadLine support is
2608        requested, but also to allow for it to be overridden. This patch came from
2609        Daniel Bergström.
2610    
2611    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2612        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2613        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2614        spotting this.
2615    
2616    
2617    Version 7.6 28-Jan-08
2618    ---------------------
2619    
2620    1.  A character class containing a very large number of characters with
2621        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2622        overflow.
2623    
2624    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2625        HAVE_LONG_LONG is not defined.
2626    
2627    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2628        bring them up to date and include new features. This patch includes:
2629    
2630        - Fixed PH's badly added libz and libbz2 support.
2631        - Fixed a problem with static linking.
2632        - Added pcredemo. [But later removed - see 7 below.]
2633        - Fixed dftables problem and added an option.
2634        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2635            HAVE_LONG_LONG.
2636        - Added readline support for pcretest.
2637        - Added an listing of the option settings after cmake has run.
2638    
2639    4.  A user submitted a patch to Makefile that makes it easy to create
2640        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2641        Makefile.am that cause it to include this special target, without
2642        affecting anything else. Note that the same mingw target plus all
2643        the other distribution libraries and programs are now supported
2644        when configuring with CMake (see 6 below) instead of with
2645        Configure/Make.
2646    
2647    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2648        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2649        exported in the Windows port". It has not yet been confirmed that the patch
2650        solves the problem, but it does no harm.
2651    
2652    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2653        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2654        with CMake, and also correct the comment about stack recursion.
2655    
2656    7.  Remove the automatic building of pcredemo from the ./configure system and
2657        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2658        of a program that users should build themselves after PCRE is installed, so
2659        building it automatically is not really right. What is more, it gave
2660        trouble in some build environments.
2661    
2662    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2663    
2664    
2665    Version 7.5 10-Jan-08
2666    ---------------------
2667    
2668    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2669        values in parens when parsing an RE using the C++ wrapper."
2670    
2671    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2672        Characters greater than 255 were excluded from the class instead of being
2673        included.
2674    
2675    3.  The same bug as (2) above applied to negated POSIX classes such as
2676        [:^space:].
2677    
2678    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2679        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2680        I have changed it.
2681    
2682    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2683        first named subpattern) and a construct such as (?&a) would reference the
2684        first named subpattern whose name started with "a" (in other words, the
2685        length check was missing). Both these problems are fixed. "Subpattern name
2686        expected" is now given for (?&) (a zero-length name), and this patch also
2687        makes it give the same error for \k'' (previously it complained that that
2688        was a reference to a non-existent subpattern).
2689    
2690    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2691        this is right because (?- can be followed by option settings as well as by
2692        digits. I have, however, made the messages clearer.
2693    
2694    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2695        than the number used in the conditional) now cause a compile-time error.
2696        This is actually not compatible with Perl, which accepts such patterns, but
2697        treats the conditional as always being FALSE (as PCRE used to), but it
2698        seems to me that giving a diagnostic is better.
2699    
2700    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2701        and messages.
2702    
2703    9.  Fix two occurrences of "backslash" in comments that should have been
2704        "backspace".
2705    
2706    10. Remove two redundant lines of code that can never be obeyed (their function
2707        was moved elsewhere).
2708    
2709    11. The program that makes PCRE's Unicode character property table had a bug
2710        which caused it to generate incorrect table entries for sequences of
2711        characters that have the same character type, but are in different scripts.
2712        It amalgamated them into a single range, with the script of the first of
2713        them. In other words, some characters were in the wrong script. There were
2714        thirteen such cases, affecting characters in the following ranges:
2715    
2716          U+002b0 - U+002c1
2717          U+0060c - U+0060d
2718          U+0061e - U+00612
2719          U+0064b - U+0065e
2720          U+0074d - U+0076d
2721          U+01800 - U+01805
2722          U+01d00 - U+01d77
2723          U+01d9b - U+01dbf
2724          U+0200b - U+0200f
2725          U+030fc - U+030fe
2726          U+03260 - U+0327f
2727          U+0fb46 - U+0fbb1
2728          U+10450 - U+1049d
2729    
2730    12. The -o option (show only the matching part of a line) for pcregrep was not
2731        compatible with GNU grep in that, if there was more than one match in a
2732        line, it showed only the first of them. It now behaves in the same way as
2733        GNU grep.
2734    
2735    13. If the -o and -v options were combined for pcregrep, it printed a blank
2736        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2737        does the same. The return code can be used to tell if there were any
2738        non-matching lines.
2739    
2740    14. Added --file-offsets and --line-offsets to pcregrep.
2741    
2742    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2743        infinitely looping recursion. The bug was that positive lookaheads were not
2744        being skipped when checking for a possible empty match (negative lookaheads
2745        and both kinds of lookbehind were skipped).
2746    
2747    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2748        inclusion of <windows.h> to before rather than after the definition of
2749        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2750    
2751    17. Specifying a possessive quantifier with a specific limit for a Unicode
2752        character property caused pcre_compile() to compile bad code, which led at
2753        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2754        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2755        caused the error; without that there was no problem.
2756    
2757    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2758    
2759    19. Added --enable-pcretest-libreadline.
2760    
2761    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2762        RE::GlobalReplace(). As a result, the number of replacements returned was
2763        double what it should be. I removed one of the increments, but Craig sent a
2764        later patch that removed the other one (the right fix) and added unit tests
2765        that check the return values (which was not done before).
2766    
2767    21. Several CMake things:
2768    
2769        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2770            the names libpcre and libpcreposix, not just pcre and pcreposix.
2771    
2772        (2) The above change means that pcretest and pcregrep are now correctly
2773            linked with the newly-built libraries, not previously installed ones.
2774    
2775        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2776    
2777    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2778        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2779        UTF-8 newline character). The key issue is that the pattern starts .*;
2780        this means that the match must be either at the beginning, or after a
2781        newline. The bug was in the code for advancing after a failed match and
2782        checking that the new position followed a newline. It was not taking
2783        account of UTF-8 characters correctly.
2784    
2785    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2786        character classes. PCRE was not treating the sequence [:...:] as a
2787        character class unless the ... were all letters. Perl, however, seems to
2788        allow any characters between [: and :], though of course it rejects as
2789        unknown any "names" that contain non-letters, because all the known class
2790        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2791        for example, whereas PCRE did not - it did not recognize a POSIX character
2792        class. This seemed a bit dangerous, so the code has been changed to be
2793        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2794        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2795        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2796        Perl does, and where it didn't before.
2797    
2798    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2799        Windows environments %n is disabled by default.
2800    
2801    
2802    Version 7.4 21-Sep-07
2803    ---------------------
2804    
2805    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2806        means that a class such as [\s] counted as "explicit reference to CR or
2807        LF". That isn't really right - the whole point of the change was to try to
2808        help when there was an actual mention of one of the two characters. So now
2809        the change happens only if \r or \n (or a literal CR or LF) character is
2810        encountered.
2811    
2812    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2813        of both had grown to the point where there were only 3 bits left.
2814        Fortunately, there was spare space in the data structure, and so I have
2815        moved the internal flags into a new 16-bit field to free up more option
2816        bits.
2817    
2818    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2819        but did not set the internal JCHANGED flag - either of these is enough to
2820        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2821        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2822        start sets both bits.
2823    
2824    4.  Added options (at build time, compile time, exec time) to change \R from
2825        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2826    
2827    5.  doc/pcresyntax.html was missing from the distribution.
2828    
2829    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2830        compatibility, even though it is no longer used.
2831    
2832    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2833        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2834        windows.h file is present (where different names are used). [This was
2835        reversed later after testing - see 16 below.]
2836    
2837    8.  Changed all #include <config.h> to #include "config.h". There were also
2838        some further <pcre.h> cases that I changed to "pcre.h".
2839    
2840    9.  When pcregrep was used with the --colour option, it missed the line ending
2841        sequence off the lines that it output.
2842    
2843    10. It was pointed out to me that arrays of string pointers cause lots of
2844        relocations when a shared library is dynamically loaded. A technique of
2845        using a single long string with a table of offsets can drastically reduce
2846        these. I have refactored PCRE in four places to do this. The result is
2847        dramatic:
2848    
2849          Originally:                          290
2850          After changing UCP table:            187
2851          After changing error message table:   43
2852          After changing table of "verbs"       36
2853          After changing table of Posix names   22
2854    
2855        Thanks to the folks working on Gregex for glib for this insight.
2856    
2857    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2858        unicode-properties was also set.
2859    
2860    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2861    
2862    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2863        checked only for CRLF.
2864    
2865    14. Added casts to pcretest.c to avoid compiler warnings.
2866    
2867    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2868    
2869    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2870        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2871        entirely. This removes changes made in 7 above.
2872    
2873    17. The CMake files have been updated, and there is now more information about
2874        building with CMake in the NON-UNIX-USE document.
2875    
2876    
2877    Version 7.3 28-Aug-07
2878    ---------------------
2879    
2880     1. In the rejigging of the build system that eventually resulted in 7.1, the
2881        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2882        brackets there is not right, since it causes compilers to look for an
2883        installed pcre.h, not the version that is in the source that is being
2884        compiled (which of course may be different). I have changed it back to:
2885    
2886          #include "pcre.h"
2887    
2888        I have a vague recollection that the change was concerned with compiling in
2889        different directories, but in the new build system, that is taken care of
2890        by the VPATH setting the Makefile.
2891    
2892     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2893        when the subject happened to end in the byte 0x85 (e.g. if the last
2894        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2895        characters but of course it shouldn't be taken as a newline when it is part
2896        of another character. The bug was that, for an unlimited repeat of . in
2897        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2898        characters when looking for a newline.
2899    
2900     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2901    
2902     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2903        in debug output.
2904    
2905     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2906        long printing in the pcrecpp unittest when running under MinGW.
2907    
2908     6. ESC_K was left out of the EBCDIC table.
2909    
2910     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2911        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2912        limit also applies to "virtual nesting" when a pattern is recursive, and in
2913        this case 1000 isn't so big. I have been able to remove this limit at the
2914        expense of backing off one optimization in certain circumstances. Normally,
2915        when pcre_exec() would call its internal match() function recursively and
2916        immediately return the result unconditionally, it uses a "tail recursion"
2917        feature to save stack. However, when a subpattern that can match an empty
2918        string has an unlimited repetition quantifier, it no longer makes this
2919        optimization. That gives it a stack frame in which to save the data for
2920        checking that an empty string has been matched. Previously this was taken
2921        from the 1000-entry workspace that had been reserved. So now there is no
2922        explicit limit, but more stack is used.
2923    
2924     8. Applied Daniel's patches to solve problems with the import/export magic
2925        syntax that is required for Windows, and which was going wrong for the
2926        pcreposix and pcrecpp parts of the library. These were overlooked when this
2927        problem was solved for the main library.
2928    
2929     9. There were some crude static tests to avoid integer overflow when computing
2930        the size of patterns that contain repeated groups with explicit upper
2931        limits. As the maximum quantifier is 65535, the maximum group length was
2932        set at 30,000 so that the product of these two numbers did not overflow a
2933        32-bit integer. However, it turns out that people want to use groups that
2934        are longer than 30,000 bytes (though not repeat them that many times).
2935        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2936        made it possible to implement the integer overflow checks in a much more
2937        dynamic way, which I have now done. The artificial limitation on group
2938        length has been removed - we now have only the limit on the total length of
2939        the compiled pattern, which depends on the LINK_SIZE setting.
2940    
2941    10. Fixed a bug in the documentation for get/copy named substring when
2942        duplicate names are permitted. If none of the named substrings are set, the
2943        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2944        empty string.
2945    
2946    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2947        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2948        because the ] is interpreted as the first data character and the
2949        terminating ] is not found. PCRE has been made compatible with Perl in this
2950        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2951        cause memory overwriting.
2952    
2953    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2954        string has been matched (to stop an infinite loop). It was not recognizing
2955        a conditional subpattern that could match an empty string if that
2956        subpattern was within another subpattern. For example, it looped when
2957        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2958        condition was not nested. This bug has been fixed.
2959    
2960    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2961        past the start of the subject in the presence of bytes with the top bit
2962        set, for example "\x8aBCD".
2963    
2964    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2965        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2966    
2967    14. Optimized (?!) to (*FAIL).
2968    
2969    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2970        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2971        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2972        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2973        does: it's just the validity check that is more restrictive.
2974    
2975    16. Inserted checks for integer overflows during escape sequence (backslash)
2976        processing, and also fixed erroneous offset values for syntax errors during
2977        backslash processing.
2978    
2979    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2980        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2981    
2982    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2983        caused an overrun.
2984    
2985    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2986        something other than just ASCII characters) inside a group that had an
2987        unlimited repeat caused a loop at compile time (while checking to see
2988        whether the group could match an empty string).
2989    
2990    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2991        [\P{Any}] did so. (Error in the code for printing property names.)
2992    
2993    21. An orphan \E inside a character class could cause a crash.
2994    
2995    22. A repeated capturing bracket such as (A)? could cause a wild memory
2996        reference during compilation.
2997    
2998    23. There are several functions in pcre_compile() that scan along a compiled
2999        expression for various reasons (e.g. to see if it's fixed length for look
3000        behind). There were bugs in these functions when a repeated \p or \P was
3001        present in the pattern. These operators have additional parameters compared
3002        with \d, etc, and these were not being taken into account when moving along
3003        the compiled data. Specifically:
3004    
3005        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3006            length.
3007    
3008        (b) An item such as \pL+ within a repeated group could cause crashes or
3009            loops.
3010    
3011        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3012            "reference to non-existent subpattern" error.
3013    
3014        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3015    
3016    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3017        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3018    
3019    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3020    
3021    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3022        character were causing crashes (broken optimization).
3023    
3024    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3025        \p or \P) caused a compile-time loop.
3026    
3027    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3028        break. For example, the unstudied pattern [\r\n]A does not match the string
3029        "\r\nA" because change 7.0/46 below moves the current point on by two
3030        characters after failing to match at the start. However, the pattern \nA
3031        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3032        the same is true. There doesn't seem any very clean way out of this, but
3033        what I have chosen to do makes the common cases work: PCRE now takes note
3034        of whether there can be an explicit match for \r or \n anywhere in the
3035        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3036        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3037        pattern has explicit CR or LF references.
3038    
3039    29. Added (*CR) etc for changing newline setting at start of pattern.
3040    
3041    
3042    Version 7.2 19-Jun-07
3043    ---------------------
3044    
3045     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3046        which is apparently normally available under Windows.
3047    
3048     2. Re-jig the pcregrep tests with different newline settings in an attempt
3049        to make them independent of the local environment's newline setting.
3050    
3051     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3052    
3053     4. Some of the "internals" tests were previously cut out when the link size
3054        was not 2, because the output contained actual offsets. The recent new
3055        "Z" feature of pcretest means that these can be cut out, making the tests
3056        usable with all link sizes.
3057    
3058     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3059        stack recursion. This gives a massive performance boost under BSD, but just
3060        a small improvement under Linux. However, it saves one field in the frame
3061        in all cases.
3062    
3063     6. Added more features from the forthcoming Perl 5.10:
3064    
3065        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3066            recursion call. It refers to the nth most recently opened parentheses.
3067    
3068        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3069            to be opened parentheses.
3070    
3071        (c) Conditions that refer to capturing parentheses can be specified
3072            relatively, for example, (?(-2)... or (?(+3)...
3073    
3074        (d) \K resets the start of the current match so that everything before
3075            is not part of it.
3076    
3077        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3078    
3079        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3080            reference syntax.
3081    
3082        (g) (?| introduces a group in which the numbering of parentheses in each
3083            alternative starts with the same number.
3084    
3085        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3086    
3087     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3088        PCRE_INFO_JCHANGED.
3089    
3090     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3091        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3092        for detecting groups that can match an empty string.
3093    
3094     9. A pattern with a very large number of alternatives (more than several
3095        hundred) was running out of internal workspace during the pre-compile
3096        phase, where pcre_compile() figures out how much memory will be needed. A
3097        bit of new cunning has reduced the workspace needed for groups with
3098        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3099        workspace instead of running out of the 4096 that are available.
3100    
3101    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3102    
3103    11. Applied patch from Google to remove an optimization that didn't quite work.
3104        The report of the bug said:
3105    
3106          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3107          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3108          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3109    
3110    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3111        it matched the wrong number of bytes.
3112    
3113    
3114    Version 7.1 24-Apr-07
3115    ---------------------
3116    
3117     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3118        that is more "standard", making use of automake and other Autotools. There
3119        is some re-arrangement of the files and adjustment of comments consequent
3120        on this.
3121    
3122     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3123        for recursive directory scanning broke on some systems because the files
3124        are not scanned in any specific order and on different systems the order
3125        was different. A call to "sort" has been inserted into RunGrepTest for the
3126        approprate test as a short-term fix. In the longer term there may be an
3127        alternative.
3128    
3129     3. I had an email from Eric Raymond about problems translating some of PCRE's
3130        man pages to HTML (despite the fact that I distribute HTML pages, some
3131        people do their own conversions for various reasons). The problems
3132        concerned the use of low-level troff macros .br and .in. I have therefore
3133        removed all such uses from the man pages (some were redundant, some could
3134        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3135        HTML has been updated to handle .nf/.fi and to complain if it encounters
3136        .br or .in.
3137    
3138     4. Updated comments in configure.ac that get placed in config.h.in and also
3139        arranged for config.h to be included in the distribution, with the name
3140        config.h.generic, for the benefit of those who have to compile without
3141        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3142    
3143     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3144        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3145        makevp.bat for the current PCRE, using the additional files
3146        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3147    
3148     6. A Windows user reported a minor discrepancy with test 2, which turned out
3149        to be caused by a trailing space on an input line that had got lost in his
3150        copy. The trailing space was an accident, so I've just removed it.
3151    
3152     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3153        that is needed.
3154    
3155     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3156        as "const" (a) because they are and (b) because it helps the PHP
3157        maintainers who have recently made a script to detect big data structures
3158        in the php code that should be moved to the .rodata section. I remembered
3159        to update Builducptable as well, so it won't revert if ucptable.h is ever
3160        re-created.
3161    
3162     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3163        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3164        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3165        support is not required. This saves 1.5-2K of code, which is important in
3166        some applications.
3167    
3168        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3169        so as not to refer to the tables, even though these functions will never be
3170        called when UTF-8 support is disabled. Otherwise there are problems with a
3171        shared library.
3172    
3173    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3174    
3175        (a) It was defining its arguments as char * instead of void *.
3176    
3177        (b) It was assuming that all moves were upwards in memory; this was true
3178            a long time ago when I wrote it, but is no longer the case.
3179    
3180        The emulated memove() is provided for those environments that have neither
3181        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3182        is clearly not the case, as these two bugs were recently reported.
3183    
3184    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3185        and Detrail to create the HTML documentation, the .txt form of the man
3186        pages, and it removes trailing spaces from listed files. It also creates
3187        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3188        case, it wraps all the #defines with #ifndefs. This script should be run
3189        before "make dist".
3190    
3191    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3192        with Unicode property support.
3193    
3194        (a) For a maximizing quantifier, if the two different cases of the
3195            character were of different lengths in their UTF-8 codings (there are
3196            some cases like this - I found 11), and the matching function had to
3197            back up over a mixture of the two cases, it incorrectly assumed they
3198            were both the same length.
3199    
3200        (b) When PCRE was configured to use the heap rather than the stack for
3201            recursion during matching, it was not correctly preserving the data for
3202            the other case of a UTF-8 character when checking ahead for a match
3203            while processing a minimizing repeat. If the check also involved
3204            matching a wide character, but failed, corruption could cause an
3205            erroneous result when trying to check for a repeat of the original
3206            character.
3207    
3208    13. Some tidying changes to the testing mechanism:
3209    
3210        (a) The RunTest script now detects the internal link size and whether there
3211            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3212            values substituted by "configure". (The RunGrepTest script already did
3213            this for UTF-8.) The configure.ac script no longer substitutes the
3214            relevant variables.
3215    
3216        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3217            with length and offset values. This means that the output is different
3218            for different internal link sizes. Test 2 is skipped for link sizes
3219            other than 2 because of this, bypassing the problem. Unfortunately,
3220            there was also a test in test 3 (the locale tests) that used /B and
3221            failed for link sizes other than 2. Rather than cut the whole test out,
3222            I have added a new /Z option to pcretest that replaces the length and
3223            offset values with spaces. This is now used to make test 3 independent
3224            of link size. (Test 2 will be tidied up later.)
3225    
3226    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3227        segmentation fault instead of returning the appropriate error message.
3228    
3229    15. In multiline mode when the newline sequence was set to "any", the pattern
3230        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3231        This doesn't seem right; it now treats the CRLF combination as the line
3232        ending, and so does not match in that case. It's only a pattern such as ^$
3233        that would hit this one: something like ^ABC$ would have failed after \r
3234        and then tried again after \r\n.
3235    
3236    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3237        in an attempt to make files that differ only in their line terminators
3238        compare equal. This works on Linux.
3239    
3240    17. Under certain error circumstances pcregrep might try to free random memory
3241        as it exited. This is now fixed, thanks to valgrind.
3242    
3243    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3244        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3245        was because its rules for how to advance for /g after matching an empty
3246        string at the end of a line did not allow for this case. They now check for
3247        it specially.
3248    
3249    20. pcretest is supposed to handle patterns and data of any length, by
3250        extending its buffers when necessary. It was getting this wrong when the
3251        buffer for a data line had to be extended.
3252    
3253    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3254        CRLF as a newline sequence.
3255    
3256    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3257        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3258        I have nevertheless tidied it up.
3259    
3260    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3261    
3262    24. Added a man page for pcre-config.
3263    
3264    
3265    Version 7.0 19-Dec-06
3266    ---------------------
3267    
3268     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3269        moving to gcc 4.1.1.
3270    
3271     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3272        sys/time.h, which is documented as needed for this function. It doesn't
3273        seem to matter on Linux, but it showed up on some releases of OS X.
3274    
3275     3. It seems that there are systems where bytes whose values are greater than
3276        127 match isprint() in the "C" locale. The "C" locale should be the
3277        default when a C program starts up. In most systems, only ASCII printing
3278        characters match isprint(). This difference caused the output from pcretest
3279        to vary, making some of the tests fail. I have changed pcretest so that:
3280    
3281        (a) When it is outputting text in the compiled version of a pattern, bytes
3282            other than 32-126 are always shown as hex escapes.
3283    
3284        (b) When it is outputting text that is a matched part of a subject string,
3285            it does the same, unless a different locale has been set for the match
3286            (using the /L modifier). In this case, it uses isprint() to decide.
3287    
3288     4. Fixed a major bug that caused incorrect computation of the amount of memory
3289        required for a compiled pattern when options that changed within the
3290        pattern affected the logic of the preliminary scan that determines the
3291        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3292        that the computed length was too small. The symptoms of this bug were
3293        either the PCRE error "internal error: code overflow" from pcre_compile(),
3294        or a glibc crash with a message such as "pcretest: free(): invalid next
3295        size (fast)". Examples of patterns that provoked this bug (shown in
3296        pcretest format) are:
3297    
3298          /(?-x: )/x
3299          /(?x)(?-x: \s*#\s*)/
3300          /((?i)[\x{c0}])/8
3301          /(?i:[\x{c0}])/8
3302    
3303        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3304        is now done differently.
3305    
3306     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3307        wrapper classes; (b) implement a new function in the C++ scanner that is
3308        more efficient than the old way of doing things because it avoids levels of
3309        recursion in the regex matching; (c) add a paragraph to the documentation
3310        for the FullMatch() function.
3311    
3312     6. The escape sequence \n was being treated as whatever was defined as
3313        "newline". Not only was this contrary to the documentation, which states
3314        that \n is character 10 (hex 0A), but it also went horribly wrong when
3315        "newline" was defined as CRLF. This has been fixed.
3316    
3317     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3318        was being set to -1 for the "end of line" case (supposedly a value that no
3319        character can have). Though this value is never used (the check for end of
3320        line is "zero bytes in current character"), it caused compiler complaints.
3321        I've changed it to 0xffffffff.
3322    
3323     8. In pcre_version.c, the version string was being built by a sequence of
3324        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3325        string (as it is for production releases) called a macro with an empty
3326        argument. The C standard says the result of this is undefined. The gcc
3327        compiler treats it as an empty string (which was what was wanted) but it is
3328        reported that Visual C gives an error. The source has been hacked around to
3329        avoid this problem.
3330    
3331     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3332        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3333        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3334        of them did).
3335    
3336    10. Originally, pcretest opened its input and output without "b"; then I was
3337        told that "b" was needed in some environments, so it was added for release
3338        5.0 to both the input and output. (It makes no difference on Unix-like
3339        systems.) Later I was told that it is wrong for the input on Windows. I've
3340        now abstracted the modes into two macros, to make it easier to fiddle with
3341        them, and removed "b" from the input mode under Windows.
3342    
3343    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3344    
3345    12. Added -help and --help to pcretest as an official way of being reminded
3346        of the options.
3347    
3348    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3349        and pcrecpp.cc because they annoy compilers at high warning levels.
3350    
3351    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3352    
3353    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3354        scripts are not C programs :-) and which was not noticed because it works
3355        on Linux.
3356    
3357    16. pcretest is supposed to handle any length of pattern and data line (as one
3358        line or as a continued sequence of lines) by extending its input buffer if
3359        necessary. This feature was broken for very long pattern lines, leading to
3360        a string of junk being passed to pcre_compile() if the pattern was longer
3361        than about 50K.
3362    
3363    17. I have done a major re-factoring of the way pcre_compile() computes the
3364        amount of memory needed for a compiled pattern. Previously, there was code
3365        that made a preliminary scan of the pattern in order to do this. That was
3366        OK when PCRE was new, but as the facilities have expanded, it has become
3367        harder and harder to keep it in step with the real compile phase, and there
3368        have been a number of bugs (see for example, 4 above). I have now found a
3369        cunning way of running the real compile function in a "fake" mode that
3370        enables it to compute how much memory it would need, while actually only
3371        ever using a few hundred bytes of working memory and without too many
3372        tests of the mode. This should make future maintenance and development
3373        easier. A side effect of this work is that the limit of 200 on the nesting
3374        depth of parentheses has been removed (though this was never a serious
3375        limitation, I suspect). However, there is a downside: pcre_compile() now
3376        runs more slowly than before (30% or more, depending on the pattern). I
3377        hope this isn't a big issue. There is no effect on runtime performance.
3378    
3379    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3380        newline (only possible for the last line of a file) and it was a
3381        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3382    
3383    19. Added additional timing features to pcretest. (1) The -tm option now times
3384        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3385        separate command line item, by a number that specifies the number of
3386        repeats to use when timing. The default is 50000; this gives better
3387        precision, but takes uncomfortably long for very large patterns.
3388    
3389    20. Extended pcre_study() to be more clever in cases where a branch of a
3390        subpattern has no definite first character. For example, (a*|b*)[cd] would
3391        previously give no result from pcre_study(). Now it recognizes that the
3392        first character must be a, b, c, or d.
3393    
3394    21. There was an incorrect error "recursive call could loop indefinitely" if
3395        a subpattern (or the entire pattern) that was being tested for matching an
3396        empty string contained only one non-empty item after a nested subpattern.
3397        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3398        incorrectly, because the \d was being skipped in the check.
3399    
3400    22. The pcretest program now has a new pattern option /B and a command line
3401        option -b, which is equivalent to adding /B to every pattern. This causes
3402        it to show the compiled bytecode, without the additional information that
3403        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3404        is the same as /B/I).
3405    
3406    23. A new optimization is now able automatically to treat some sequences such
3407        as a*b as a*+b. More specifically, if something simple (such as a character
3408        or a simple class like \d) has an unlimited quantifier, and is followed by
3409        something that cannot possibly match the quantified thing, the quantifier
3410        is automatically "possessified".
3411    
3412    24. A recursive reference to a subpattern whose number was greater than 39
3413        went wrong under certain circumstances in UTF-8 mode. This bug could also
3414        have affected the operation of pcre_study().
3415    
3416    25. Realized that a little bit of performance could be had by replacing
3417        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3418    
3419    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3420    
3421    27. Possessive quantifiers such as a++ were previously implemented by turning
3422        them into atomic groups such as ($>a+). Now they have their own opcodes,
3423        which improves performance. This includes the automatically created ones
3424        from 23 above.
3425    
3426    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3427        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3428        the first matched character to be a colon. This applied both to named and
3429        numbered groups.
3430    
3431    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3432    
3433    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3434        building PCRE on Windows easier, so I have included it in the distribution.
3435    
3436    31. There is now a check in pcretest against a ridiculously large number being
3437        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3438        loop, the loop is abandoned.
3439    
3440    32. Forward references to subpatterns in conditions such as (?(2)...) where
3441        subpattern 2 is defined later cause pcre_compile() to search forwards in
3442        the pattern for the relevant set of parentheses. This search went wrong
3443        when there were unescaped parentheses in a character class, parentheses
3444        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3445    
3446    33. "Subroutine" calls and backreferences were previously restricted to
3447        referencing subpatterns earlier in the regex. This restriction has now
3448        been removed.
3449    
3450    34. Added a number of extra features that are going to be in Perl 5.10. On the
3451        whole, these are just syntactic alternatives for features that PCRE had
3452        previously implemented using the Python syntax or my own invention. The
3453        other formats are all retained for compatibility.
3454    
3455        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3456            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3457            also .NET compatible.
3458    
3459        (b) A recursion or subroutine call to a named group can now be defined as
3460            (?&name) as well as (?P>name).
3461    
3462        (c) A backreference to a named group can now be defined as \k<name> or
3463            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3464            5.10, are also .NET compatible.
3465    
3466        (d) A conditional reference to a named group can now use the syntax
3467            (?(<name>) or (?('name') as well as (?(name).
3468    
3469        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3470            groups (named and numbered) that are never evaluated inline, but can be
3471            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3472            is always false. There may be only one alternative in such a group.
3473    
3474        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3475            as the simple (?(R). The condition is true only if the most recent
3476            recursion is that of the given number or name. It does not search out
3477            through the entire recursion stack.
3478    
3479        (g) The escape \gN or \g{N} has been added, where N is a positive or
3480            negative number, specifying an absolute or relative reference.
3481    
3482    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3483        some "unreachable code" warnings.
3484    
3485    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3486        things, this adds five new scripts.
3487    
3488    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3489        There were also incompatibilities regarding the handling of \Q..\E inside
3490        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3491        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3492    
3493    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3494        matches an empty string, and forcibly breaks the loop. There were bugs in
3495        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3496        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3497        separate and independent bugs (that affected different cases) have been
3498        fixed.
3499    
3500    39. Refactored the code to abolish the use of different opcodes for small
3501        capturing bracket numbers. This is a tidy that I avoided doing when I
3502        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3503        The new approach is not only tidier, it makes it possible to reduce the
3504        memory needed to fix the previous bug (38).
3505    
3506    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3507        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3508        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3509        mode.
3510    
3511    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3512        report.
3513    
3514    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3515        copy construction and assignment in the C++ wrapper.
3516    
3517    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3518        couple of bugs that could have given wrong results in the "--newline=crlf"
3519        case.
3520    
3521    44. Added a number of casts and did some reorganization of signed/unsigned int
3522        variables following suggestions from Dair Grant. Also renamed the variable
3523        "this" as "item" because it is a C++ keyword.
3524    
3525    45. Arranged for dftables to add
3526    
3527          #include "pcre_internal.h"
3528    
3529        to pcre_chartables.c because without it, gcc 4.x may remove the array
3530        definition from the final binary if PCRE is built into a static library and
3531        dead code stripping is activated.
3532    
3533    46. For an unanchored pattern, if a match attempt fails at the start of a
3534        newline sequence, and the newline setting is CRLF or ANY, and the next two
3535        characters are CRLF, advance by two characters instead of one.
3536    
3537    
3538    Version 6.7 04-Jul-06
3539    ---------------------
3540    
3541     1. In order to handle tests when input lines are enormously long, pcretest has
3542        been re-factored so that it automatically extends its buffers when
3543        necessary. The code is crude, but this _is_ just a test program. The
3544        default size has been increased from 32K to 50K.
3545    
3546     2. The code in pcre_study() was using the value of the re argument before
3547        testing it for NULL. (Of course, in any sensible call of the function, it
3548        won't be NULL.)
3549    
3550     3. The memmove() emulation function in pcre_internal.h, which is used on
3551        systems that lack both memmove() and bcopy() - that is, hardly ever -
3552        was missing a "static" storage class specifier.
3553    
3554     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3555        containing an extended class (one that cannot be represented by a bitmap
3556        because it contains high-valued characters or Unicode property items, e.g.
3557        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3558        pattern, but PCRE should not loop if you do not (it no longer does).
3559        [Detail: two cases were found: (a) a repeated subpattern containing an
3560        extended class; (b) a recursive reference to a subpattern that followed a
3561        previous extended class. It wasn't skipping over the extended class
3562        correctly when UTF-8 mode was not set.]
3563    
3564     5. A negated single-character class was not being recognized as fixed-length
3565        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3566        compile error "lookbehind assertion is not fixed length".
3567    
3568     6. The RunPerlTest auxiliary script was showing an unexpected difference
3569        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3570        write a Perl script that can interpret lines of an input file either as
3571        byte characters or as UTF-8, which is what "perltest" was being required to
3572        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3573        can't do is switch easily at run time between having the "use utf8;" pragma
3574        or not. In the end, I fudged it by using the RunPerlTest script to insert
3575        "use utf8;" explicitly for the UTF-8 tests.
3576    
3577     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3578        the end of the subject string, contrary to the documentation and to what
3579        Perl does. This was true of both matching functions. Now it matches only at
3580        the start of the subject and immediately after *internal* newlines.
3581    
3582     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3583        a pointer to an int instead of a pointer to an unsigned long int. This
3584        caused problems on 64-bit systems.
3585    
3586     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3587        instance of the 'standard' template library not being so standard".
3588    
3589    10. There was no check on the number of named subpatterns nor the maximum
3590        length of a subpattern name. The product of these values is used to compute
3591        the size of the memory block for a compiled pattern. By supplying a very
3592        long subpattern name and a large number of named subpatterns, the size
3593        computation could be caused to overflow. This is now prevented by limiting
3594        the length of names to 32 characters, and the number of named subpatterns
3595        to 10,000.
3596    
3597    11. Subpatterns that are repeated with specific counts have to be replicated in
3598        the compiled pattern. The size of memory for this was computed from the
3599        length of the subpattern and the repeat count. The latter is limited to
3600        65535, but there was no limit on the former, meaning that integer overflow
3601        could in principle occur. The compiled length of a repeated subpattern is
3602        now limited to 30,000 bytes in order to prevent this.
3603    
3604    12. Added the optional facility to have named substrings with the same name.
3605    
3606    13. Added the ability to use a named substring as a condition, using the
3607        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3608        are numbers (not recommended). Forward references are permitted.
3609    
3610    14. Added forward references in named backreferences (if you see what I mean).
3611    
3612    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3613        pattern could run off the end of the subject. For example, the pattern
3614        "(?s)(.{1,5})"8 did this with the subject "ab".
3615    
3616    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3617        PCRE_CASELESS was set when matching characters that were quantified with ?
3618        or *.
3619    
3620    17. A character class other than a single negated character that had a minimum
3621        but no maximum quantifier - for example [ab]{6,} - was not handled
3622        correctly by pce_dfa_exec(). It would match only one character.
3623    
3624    18. A valid (though odd) pattern that looked like a POSIX character
3625        class but used an invalid character after [ (for example [[,abc,]]) caused
3626        pcre_compile() to give the error "Failed: internal error: code overflow" or
3627        in some cases to crash with a glibc free() error. This could even happen if
3628        the pattern terminated after [[ but there just happened to be a sequence of
3629        letters, a binary zero, and a closing ] in the memory that followed.
3630    
3631    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3632        over the years. Originally (before any Unicode support), just the bottom 8
3633        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3634        output from "man perlunicode" includes this:
3635    
3636          The regular expression compiler produces polymorphic opcodes.  That
3637          is, the pattern adapts to the data and automatically switches to
3638          the Unicode character scheme when presented with Unicode data--or
3639          instead uses a traditional byte scheme when presented with byte
3640          data.
3641    
3642        Sadly, a wide octal escape does not cause a switch, and in a string with
3643        no other multibyte characters, these octal escapes are treated as before.
3644        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3645        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3646        Unicode string.
3647    
3648        I have not perpetrated such confusion in PCRE. Up till now, it took just
3649        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3650        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3651        translate to the appropriate multibyte character.
3652    
3653    29. Applied some refactoring to reduce the number of warnings from Microsoft
3654        and Borland compilers. This has included removing the fudge introduced
3655        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3656        a warning about an unused variable.
3657    
3658    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3659        characters since release 4.0, because Perl (from release 5.004) does not.
3660        [Or at least, is documented not to: some releases seem to be in conflict
3661        with the documentation.] However, when a pattern was studied with
3662        pcre_study() and all its branches started with \s, PCRE still included VT
3663        as a possible starting character. Of course, this did no harm; it just
3664        caused an unnecessary match attempt.
3665    
3666    22. Removed a now-redundant internal flag bit that recorded the fact that case
3667        dependency changed within the pattern. This was once needed for "required
3668        byte" processing, but is no longer used. This recovers a now-scarce options
3669        bit. Also moved the least significant internal flag bit to the most-
3670        significant bit of the word, which was not previously used (hangover from
3671        the days when it was an int rather than a uint) to free up another bit for
3672        the future.
3673    
3674    23. Added support for CRLF line endings as well as CR and LF. As well as the
3675        default being selectable at build time, it can now be changed at runtime
3676        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3677        specify that it is scanning data with non-default line endings.
3678    
3679    24. Changed the definition of CXXLINK to make it agree with the definition of
3680        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3681    
3682    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3683        recursions. This makes a big different to stack usage for some patterns.
3684    
3685    26. If a subpattern containing a named recursion or subroutine reference such
3686        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3687        the space required for the compiled pattern went wrong and gave too small a
3688        value. Depending on the environment, this could lead to "Failed: internal
3689        error: code overflow at offset 49" or "glibc detected double free or
3690        corruption" errors.
3691    
3692    27. Applied patches from Google (a) to support the new newline modes and (b) to
3693        advance over multibyte UTF-8 characters in GlobalReplace.
3694    
3695    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3696        difference for some implementation of PCRE in some Windows version.
3697    
3698    29. Added some extra testing facilities to pcretest:
3699    
3700        \q<number>   in a data line sets the "match limit" value
3701        \Q<number>   in a data line sets the "match recursion limt" value
3702        -S <number>  sets the stack size, where <number> is in megabytes
3703    
3704        The -S option isn't available for Windows.
3705    
3706    
3707    Version 6.6 06-Feb-06
3708    ---------------------
3709    
3710     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3711        in pcreposix.h. I have copied the definition from pcre.h.
3712    
3713     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3714        because pcre.h is no longer a built file.
3715    
3716     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3717        not normally included in the compiled code.
3718    
3719    
3720    Version 6.5 01-Feb-06
3721    ---------------------
3722    
3723     1. When using the partial match feature with pcre_dfa_exec(), it was not
3724        anchoring the second and subsequent partial matches at the new starting
3725        point. This could lead to incorrect results. For example, with the pattern
3726        /1234/, partially matching against "123" and then "a4" gave a match.
3727    
3728     2. Changes to pcregrep:
3729    
3730        (a) All non-match returns from pcre_exec() were being treated as failures
3731            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3732            error message is output. Some extra information is given for the
3733            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3734            probably the only errors that are likely to be caused by users (by
3735            specifying a regex that has nested indefinite repeats, for instance).
3736            If there are more than 20 of these errors, pcregrep is abandoned.
3737    
3738        (b) A binary zero was treated as data while matching, but terminated the
3739            output line if it was written out. This has been fixed: binary zeroes
3740            are now no different to any other data bytes.
3741    
3742        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3743            used to set a locale for matching. The --locale=xxxx long option has
3744            been added (no short equivalent) to specify a locale explicitly on the
3745            pcregrep command, overriding the environment variables.
3746    
3747        (d) When -B was used with -n, some line numbers in the output were one less
3748            than they should have been.
3749    
3750        (e) Added the -o (--only-matching) option.
3751    
3752        (f) If -A or -C was used with -c (count only), some lines of context were
3753            accidentally printed for the final match.
3754    
3755        (g) Added the -H (--with-filename) option.
3756    
3757        (h) The combination of options -rh failed to suppress file names for files
3758            that were found from directory arguments.
3759    
3760        (i) Added the -D (--devices) and -d (--directories) options.
3761    
3762        (j) Added the -F (--fixed-strings) option.
3763    
3764        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3765    
3766        (l) Added the --colo(u)r option.
3767    
3768        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3769            is not present by default.
3770    
3771     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3772        items such as (?R) or (?1), when the recursion could match a number of
3773        alternatives. If it matched one of the alternatives, but subsequently,
3774        outside the recursion, there was a failure, the code tried to back up into
3775        the recursion. However, because of the way PCRE is implemented, this is not
3776        possible, and the result was an incorrect result from the match.
3777    
3778        In order to prevent this happening, the specification of recursion has
3779        been changed so that all such subpatterns are automatically treated as
3780        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3781    
3782     4. I had overlooked the fact that, in some locales, there are characters for
3783        which isalpha() is true but neither isupper() nor islower() are true. In
3784        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3785        and ordfeminine) are like this. This affected the treatment of \w and \W
3786        when they appeared in character classes, but not when they appeared outside
3787        a character class. The bit map for "word" characters is now created
3788        separately from the results of isalnum() instead of just taking it from the
3789        upper, lower, and digit maps. (Plus the underscore character, of course.)