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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1284 by ph10, Fri Mar 15 11:54:58 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46    11. The value of capture_last that is passed to callouts was incorrect in some
47        cases when there was a capture on one path that was subsequently abandoned
48        after a backtrack. Also, the capture_last value is now reset after a
49        recursion, since all captures are also reset in this case.
50    
51    12. The interpreter no longer returns the "too many substrings" error in the
52        case when an overflowing capture is in a branch that is subsequently
53        abandoned after a backtrack.
54    
55    13. In the pathological case when an offset vector of size 2 is used, pcretest
56        now prints out the matched string after a yield of 0 or 1.
57    
58    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59        Only supported by the JIT compiler at the moment.
60    
61    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62    
63    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64        offset of the starting point of the matching process, provided the offsets
65        vector is large enough.
66    
67    17. The \A escape now records a lookbehind value of 1, though its execution
68        does not actually inspect the previous character. This is to ensure that,
69        in partial multi-segment matching, at least one character from the old
70        segment is retained when a new segment is processed. Otherwise, if there
71        are no lookbehinds in the pattern, \A might match incorrectly at the start
72        of a new segment.
73    
74    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75    
76    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77        gives some modest performance improvement in 8-bit mode.
78    
79    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80        be expressed in certain programming languages using Universal Character
81        Names.
82    
83    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84        which points out that "non characters" are not "characters that may not
85        appear in Unicode strings" but rather "characters that are reserved for
86        internal use and have only local meaning".
87    
88    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89        there was a conditional group that depended on an assertion, if the
90        assertion was false, the callout that immediately followed the alternation
91        in the condition was skipped when pcre_exec() was used for matching.
92    
93    23. Allow an explicit callout to be inserted before an assertion that is the
94        condition for a conditional group, for compatibility with automatic
95        callouts, which always insert a callout at this point.
96    
97    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99    
100    25. (*PRUNE) is now supported by the JIT compiler.
101    
102    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103    
104    27. Fix the case where there are two or more SKIPs with arguments that may be
105        ignored.
106    
107    28. Experimental support of (*SKIP) backtracking verb in the JIT compiler.
108    
109    29. Experimental support of (*THEN) backtracking verb in the JIT compiler.
110    
111    30. Update RunTest with additional test selector options.
112    
113    31. PCRE has been changed to be more compatible with Perl when there is more
114        than one backtracking verb present. Previously, in something like
115        (*COMMIT)(*SKIP), COMMIT would override SKIP. Apart from one anomaly (which
116        has been reported), Perl seems to act on whichever backtracking verb is
117        reached first, so PCRE has been changed to follow this behaviour.
118    
119    
120    Version 8.32 30-November-2012
121    -----------------------------
122    
123    1.  Improved JIT compiler optimizations for first character search and single
124        character iterators.
125    
126    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
127        Patch by Daniel Richard G.
128    
129    3.  Single character iterator optimizations in the JIT compiler.
130    
131    4.  Improved JIT compiler optimizations for character ranges.
132    
133    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
134        Reported by Giuseppe D'Angelo.
135    
136    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
137        of a line, was being set incorrectly in cases where .* appeared inside
138        atomic brackets at the start of a pattern, or where there was a subsequent
139        *PRUNE or *SKIP.
140    
141    7.  Improved instruction cache flush for POWER/PowerPC.
142        Patch by Daniel Richard G.
143    
144    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
145        grep:
146    
147        (a) There is now no limit to the number of patterns to be matched.
148    
149        (b) An error is given if a pattern is too long.
150    
151        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
152            are now supported.
153    
154        (d) --exclude-from and --include-from (multiple use) have been added.
155    
156        (e) Exclusions and inclusions now apply to all files and directories, not
157            just to those obtained from scanning a directory recursively.
158    
159        (f) Multiple uses of -f and --file-list are now supported.
160    
161        (g) In a Windows environment, the default for -d has been changed from
162            "read" (the GNU grep default) to "skip", because otherwise the presence
163            of a directory in the file list provokes an error.
164    
165        (h) The documentation has been revised and clarified in places.
166    
167    9.  Improve the matching speed of capturing brackets.
168    
169    10. Changed the meaning of \X so that it now matches a Unicode extended
170        grapheme cluster.
171    
172    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
173        out POSIX threads when JIT support is configured.
174    
175    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
176    
177    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
178        structure could go wrong in environments where size_t is not the same size
179        as int.
180    
181    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
182        set.
183    
184    15. The EBCDIC support had decayed; later updates to the code had included
185        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
186        general tidy up of EBCDIC-related issues, and the documentation was also
187        not quite right. There is now a test that can be run on ASCII systems to
188        check some of the EBCDIC-related things (but is it not a full test).
189    
190    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
191        in a small tidy to the code.
192    
193    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
194    
195    18. If the --only-matching (-o) option in pcregrep is specified multiple
196        times, each one causes appropriate output. For example, -o1 -o2 outputs the
197        substrings matched by the 1st and 2nd capturing parentheses. A separating
198        string can be specified by --om-separator (default empty).
199    
200    19. Improving the first n character searches.
201    
202    20. Turn case lists for horizontal and vertical white space into macros so that
203        they are defined only once.
204    
205    21. This set of changes together give more compatible Unicode case-folding
206        behaviour for characters that have more than one other case when UCP
207        support is available.
208    
209        (a) The Unicode property table now has offsets into a new table of sets of
210            three or more characters that are case-equivalent. The MultiStage2.py
211            script that generates these tables (the pcre_ucd.c file) now scans
212            CaseFolding.txt instead of UnicodeData.txt for character case
213            information.
214    
215        (b) The code for adding characters or ranges of characters to a character
216            class has been abstracted into a generalized function that also handles
217            case-independence. In UTF-mode with UCP support, this uses the new data
218            to handle characters with more than one other case.
219    
220        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
221            whose other case is greater than 256 are now correctly matched
222            caselessly. Previously, the high codepoint matched the low one, but not
223            vice versa.
224    
225        (d) The processing of \h, \H, \v, and \ in character classes now makes use
226            of the new class addition function, using character lists defined as
227            macros alongside the case definitions of 20 above.
228    
229        (e) Caseless back references now work with characters that have more than
230            one other case.
231    
232        (f) General caseless matching of characters with more than one other case
233            is supported.
234    
235    22. Unicode character properties were updated from Unicode 6.2.0
236    
237    23. Improved CMake support under Windows. Patch by Daniel Richard G.
238    
239    24. Add support for 32-bit character strings, and UTF-32
240    
241    25. Major JIT compiler update (code refactoring and bugfixing).
242        Experimental Sparc 32 support is added.
243    
244    26. Applied a modified version of Daniel Richard G's patch to create
245        pcre.h.generic and config.h.generic by "make" instead of in the
246        PrepareRelease script.
247    
248    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
249        pcre_compile.c when checking for a zero character.
250    
251    28. Introducing a native interface for JIT. Through this interface, the compiled
252        machine code can be directly executed. The purpose of this interface is to
253        provide fast pattern matching, so several sanity checks are not performed.
254        However, feature tests are still performed. The new interface provides
255        1.4x speedup compared to the old one.
256    
257    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
258        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
259        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
260    
261    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
262        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
263        the "old" RFC 2279). Instead, it ended up passing a negative length to
264        pcre_exec().
265    
266    31. Add support for GCC's visibility feature to hide internal functions.
267    
268    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
269        "unknown -C option" after outputting 0 or 1.
270    
271    33. There is now support for generating a code coverage report for the test
272        suite in environments where gcc is the compiler and lcov is installed. This
273        is mainly for the benefit of the developers.
274    
275    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
276        unaddressable using valgrind annotations, allowing valgrind to detect
277        invalid memory accesses. This is mainly for the benefit of the developers.
278    
279    25. (*UTF) can now be used to start a pattern in any of the three libraries.
280    
281    26. Give configure error if --enable-cpp but no C++ compiler found.
282    
283    
284    Version 8.31 06-July-2012
285    -------------------------
286    
287    1.  Fixing a wrong JIT test case and some compiler warnings.
288    
289    2.  Removed a bashism from the RunTest script.
290    
291    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
292        to unsigned type, result still unsigned" that was given by an MS compiler
293        on encountering the code "-sizeof(xxx)".
294    
295    4.  Partial matching support is added to the JIT compiler.
296    
297    5.  Fixed several bugs concerned with partial matching of items that consist
298        of more than one character:
299    
300        (a) /^(..)\1/ did not partially match "aba" because checking references was
301            done on an "all or nothing" basis. This also applied to repeated
302            references.
303    
304        (b) \R did not give a hard partial match if \r was found at the end of the
305            subject.
306    
307        (c) \X did not give a hard partial match after matching one or more
308            characters at the end of the subject.
309    
310        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
311            a partial match for the string "\r".
312    
313        (e) When newline was set to CRLF, the metacharacter "." did not recognize
314            a partial match for a CR character at the end of the subject string.
315    
316    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
317        running pcretest, the text "(JIT)" added to the output whenever JIT is
318        actually used to run the match.
319    
320    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
321        or /S+[+] with a digit between 1 and 7.
322    
323    8.  OP_NOT now supports any UTF character not just single-byte ones.
324    
325    9.  (*MARK) control verb is now supported by the JIT compiler.
326    
327    10. The command "./RunTest list" lists the available tests without actually
328        running any of them. (Because I keep forgetting what they all are.)
329    
330    11. Add PCRE_INFO_MAXLOOKBEHIND.
331    
332    12. Applied a (slightly modified) user-supplied patch that improves performance
333        when the heap is used for recursion (compiled with --disable-stack-for-
334        recursion). Instead of malloc and free for each heap frame each time a
335        logical recursion happens, frames are retained on a chain and re-used where
336        possible. This sometimes gives as much as 30% improvement.
337    
338    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
339        call.
340    
341    14. As documented, (*COMMIT) is now confined to within a positive assertion.
342    
343    15. It is now possible to link pcretest with libedit as an alternative to
344        libreadline.
345    
346    16. (*COMMIT) control verb is now supported by the JIT compiler.
347    
348    17. The Unicode data tables have been updated to Unicode 6.1.0.
349    
350    18. Added --file-list option to pcregrep.
351    
352    19. Added binary file support to pcregrep, including the -a, --binary-files,
353        -I, and --text options.
354    
355    20. The madvise function is renamed for posix_madvise for QNX compatibility
356        reasons. Fixed by Giuseppe D'Angelo.
357    
358    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
359    
360    22. Changed the option for creating symbolic links for 16-bit man pages from
361        -s to -sf so that re-installing does not cause issues.
362    
363    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
364    
365    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
366        matches in certain environments (the workspace was not being correctly
367        retained). Also added to pcre_dfa_exec() a simple plausibility check on
368        some of the workspace data at the beginning of a restart.
369    
370    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
371        was not doing so when it should - probably a typo introduced by SVN 528
372        (change 8.10/14).
373    
374    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
375        \w+ when the character tables indicated that \x{c4} was a word character.
376        There were several related cases, all because the tests for doing a table
377        lookup were testing for characters less than 127 instead of 255.
378    
379    27. If a pattern contains capturing parentheses that are not used in a match,
380        their slots in the ovector are set to -1. For those that are higher than
381        any matched groups, this happens at the end of processing. In the case when
382        there were back references that the ovector was too small to contain
383        (causing temporary malloc'd memory to be used during matching), and the
384        highest capturing number was not used, memory off the end of the ovector
385        was incorrectly being set to -1. (It was using the size of the temporary
386        memory instead of the true size.)
387    
388    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
389        ovector size, it uses memory at the end of the block that it has got.
390    
391    29. Check for an overlong MARK name and give an error at compile time. The
392        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
393    
394    30. JIT compiler update.
395    
396    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
397        Rill for the patch.
398    
399    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
400    
401    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
402    
403    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
404        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
405        was enabled.
406    
407    35. Improve JIT code generation for greedy plus quantifier.
408    
409    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
410        1 to "aa" instead of to an empty string. The bug affected repeated groups
411        that could potentially match an empty string.
412    
413    37. Optimizing single character iterators in JIT.
414    
415    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
416        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
417        codepoints that are too big for the mode are faulted, and in a UTF mode,
418        disallowed codepoints are also faulted.
419    
420    39. If PCRE was compiled with UTF support, in three places in the DFA
421        matcher there was code that should only have been obeyed in UTF mode, but
422        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
423        processing when bytes with values greater than 127 were present. In 16-bit
424        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
425        both cases the values are those that cannot be the first data item in a UTF
426        character. The three items that might have provoked this were recursions,
427        possessively repeated groups, and atomic groups.
428    
429    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
430        and pcregrep, because some OS require shared objects to be explicitly
431        passed to ld, causing the link step to fail if they are not.
432    
433    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
434        mode, patterns that started with \h* or \R* might be incorrectly matched.
435    
436    
437    Version 8.30 04-February-2012
438    -----------------------------
439    
440    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
441        name is defined in ctype.h.
442    
443    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
444        only in quite long subpatterns.
445    
446    3.  Removed the function pcre_info(), which has been obsolete and deprecated
447        since it was replaced by pcre_fullinfo() in February 2000.
448    
449    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
450        match a (*MARK), and the match failed at the start of the subject, a
451        reference to memory before the start of the subject could occur. This bug
452        was introduced by fix 17 of release 8.21.
453    
454    5.  A reference to an unset group with zero minimum repetition was giving
455        totally wrong answers (in non-JavaScript-compatibility mode). For example,
456        /(another)?(\1?)test/ matched against "hello world test". This bug was
457        introduced in release 8.13.
458    
459    6.  Add support for 16-bit character strings (a large amount of work involving
460        many changes and refactorings).
461    
462    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
463        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
464        from a file.
465    
466    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
467        rounding is not applied in this particular case).
468    
469    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
470        if they appear, or are escaped, in patterns.
471    
472    10. Get rid of a number of -Wunused-but-set-variable warnings.
473    
474    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
475        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
476        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
477        also returns the mark "x". This bug applied to capturing parentheses,
478        non-capturing parentheses, and atomic parentheses. It also applied to some
479        assertions.
480    
481    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
482        information out of configure.ac instead of relying on pcre.h.generic, which
483        is not stored in the repository.
484    
485    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
486        -lreadline.
487    
488    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
489    
490    15. Applied Graycode's patch to put the top-level frame on the stack rather
491        than the heap when not using the stack for recursion. This gives a
492        performance improvement in many cases when recursion is not deep.
493    
494    16. Experimental code added to "pcretest -C" to output the stack frame size.
495    
496    
497    Version 8.21 12-Dec-2011
498    ------------------------
499    
500    1.  Updating the JIT compiler.
501    
502    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
503        are added as well.
504    
505    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
506        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
507        calling _pcre_jit_exec. Some extra comments are added.
508    
509    4.  (*MARK) settings inside atomic groups that do not contain any capturing
510        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
511        was introduced by change 18 for 8.20.
512    
513    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
514        ECMA-262 standard.
515    
516    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
517        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
518        This bug was probably introduced by change 9 of 8.13.
519    
520    7.  While fixing 6 above, I noticed that a number of other items were being
521        incorrectly rejected as "not fixed length". This arose partly because newer
522        opcodes had not been added to the fixed-length checking code. I have (a)
523        corrected the bug and added tests for these items, and (b) arranged for an
524        error to occur if an unknown opcode is encountered while checking for fixed
525        length instead of just assuming "not fixed length". The items that were
526        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
527        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
528        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
529    
530    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
531        being incorrectly compiled and would have given unpredicatble results.
532    
533    9.  A possessively repeated subpattern with minimum repeat count greater than
534        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
535        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
536        the first (A) could occur when it should not.
537    
538    10. Add a cast and remove a redundant test from the code.
539    
540    11. JIT should use pcre_malloc/pcre_free for allocation.
541    
542    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
543        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
544        is anything other than /usr, -L is still shown).
545    
546    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
547    
548    14. Perl does not support \N without a following name in a [] class; PCRE now
549        also gives an error.
550    
551    15. If a forward reference was repeated with an upper limit of around 2000,
552        it caused the error "internal error: overran compiling workspace". The
553        maximum number of forward references (including repeats) was limited by the
554        internal workspace, and dependent on the LINK_SIZE. The code has been
555        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
556        the default depends on LINK_SIZE. There is a new upper limit (for safety)
557        of around 200,000 forward references. While doing this, I also speeded up
558        the filling in of repeated forward references.
559    
560    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
561        incorrectly expecting the subject to contain another "a" after the start.
562    
563    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
564        in the match, the SKIP should be ignored. This was not happening; instead
565        the SKIP was being treated as NOMATCH. For patterns such as
566        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
567        tested.
568    
569    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
570        now much more compatible with Perl, in particular in cases where the result
571        is a non-match for a non-anchored pattern. For example, if
572        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
573        "m", where previously it did not return a name. A side effect of this
574        change is that for partial matches, the last encountered mark name is
575        returned, as for non matches. A number of tests that were previously not
576        Perl-compatible have been moved into the Perl-compatible test files. The
577        refactoring has had the pleasing side effect of removing one argument from
578        the match() function, thus reducing its stack requirements.
579    
580    19. If the /S+ option was used in pcretest to study a pattern using JIT,
581        subsequent uses of /S (without +) incorrectly behaved like /S+.
582    
583    21. Retrieve executable code size support for the JIT compiler and fixing
584        some warnings.
585    
586    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
587        not work when the shorter character appeared right at the end of the
588        subject string.
589    
590    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
591        systems.
592    
593    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
594        output it when the /M option is used in pcretest.
595    
596    25. The CheckMan script was not being included in the distribution. Also, added
597        an explicit "perl" to run Perl scripts from the PrepareRelease script
598        because this is reportedly needed in Windows.
599    
600    26. If study data was being save in a file and studying had not found a set of
601        "starts with" bytes for the pattern, the data written to the file (though
602        never used) was taken from uninitialized memory and so caused valgrind to
603        complain.
604    
605    27. Updated RunTest.bat as provided by Sheri Pierce.
606    
607    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
608    
609    29. Computation of memory usage for the table of capturing group names was
610        giving an unnecessarily large value.
611    
612    
613    Version 8.20 21-Oct-2011
614    ------------------------
615    
616    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
617        a POSIX class. After further experiments with Perl, which convinced me that
618        Perl has bugs and confusions, a closing square bracket is no longer allowed
619        in a POSIX name. This bug also affected patterns with classes that started
620        with full stops.
621    
622    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
623        captured substring, but while checking the failing first alternative,
624        substring 1 is temporarily captured. If the output vector supplied to
625        pcre_exec() was not big enough for this capture, the yield of the function
626        was still zero ("insufficient space for captured substrings"). This cannot
627        be totally fixed without adding another stack variable, which seems a lot
628        of expense for a edge case. However, I have improved the situation in cases
629        such as /(a)(b)x|abc/ matched against "abc", where the return code
630        indicates that fewer than the maximum number of slots in the ovector have
631        been set.
632    
633    3.  Related to (2) above: when there are more back references in a pattern than
634        slots in the output vector, pcre_exec() uses temporary memory during
635        matching, and copies in the captures as far as possible afterwards. It was
636        using the entire output vector, but this conflicts with the specification
637        that only 2/3 is used for passing back captured substrings. Now it uses
638        only the first 2/3, for compatibility. This is, of course, another edge
639        case.
640    
641    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
642        main code base, and can be used by building with --enable-jit. When this is
643        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
644        runtime --no-jit option is given.
645    
646    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
647        ovector, the return from the function was zero, implying that there were
648        other matches that did not fit. The correct "exactly full" value is now
649        returned.
650    
651    6.  If a subpattern that was called recursively or as a subroutine contained
652        (*PRUNE) or any other control that caused it to give a non-standard return,
653        invalid errors such as "Error -26 (nested recursion at the same subject
654        position)" or even infinite loops could occur.
655    
656    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
657        computing the minimum length on reaching *ACCEPT, and so ended up with the
658        wrong value of 1 rather than 0. Further investigation indicates that
659        computing a minimum subject length in the presence of *ACCEPT is difficult
660        (think back references, subroutine calls), and so I have changed the code
661        so that no minimum is registered for a pattern that contains *ACCEPT.
662    
663    8.  If (*THEN) was present in the first (true) branch of a conditional group,
664        it was not handled as intended. [But see 16 below.]
665    
666    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
667        Sheri Pierce.
668    
669    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
670        the first byte in a match must be "a".
671    
672    11. Change 17 for 8.13 increased the recursion depth for patterns like
673        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
674        pattern contains any instances of (*THEN). If it does not, the old
675        optimizations are restored. It would be nice to do this on a per-group
676        basis, but at the moment that is not feasible.
677    
678    12. In some environments, the output of pcretest -C is CRLF terminated. This
679        broke RunTest's code that checks for the link size. A single white space
680        character after the value is now allowed for.
681    
682    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
683        For "fr", it uses the Windows-specific input and output files.
684    
685    14. If (*THEN) appeared in a group that was called recursively or as a
686        subroutine, it did not work as intended. [But see next item.]
687    
688    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
689        pattern fragments (but not containing any | characters). If A and B are
690        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
691        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
692        D. In other words, Perl considers parentheses that do not contain any |
693        characters to be part of a surrounding alternative, whereas PCRE was
694        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
695        differently. PCRE now behaves in the same way as Perl, except in the case
696        of subroutine/recursion calls such as (?1) which have in any case always
697        been different (but PCRE had them first :-).
698    
699    16. Related to 15 above: Perl does not treat the | in a conditional group as
700        creating alternatives. Such a group is treated in the same way as an
701        ordinary group without any | characters when processing (*THEN). PCRE has
702        been changed to match Perl's behaviour.
703    
704    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
705        RunGrepTest script failed.
706    
707    18. Change 22 for version 13 caused atomic groups to use more stack. This is
708        inevitable for groups that contain captures, but it can lead to a lot of
709        stack use in large patterns. The old behaviour has been restored for atomic
710        groups that do not contain any capturing parentheses.
711    
712    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
713        suppress the check for a minimum subject length at run time. (If it was
714        given to pcre_exec() or pcre_dfa_exec() it did work.)
715    
716    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
717        fail to work when decoding hex characters in data strings in EBCDIC
718        environments.
719    
720    21. It appears that in at least one Mac OS environment, the isxdigit() function
721        is implemented as a macro that evaluates to its argument more than once,
722        contravening the C 90 Standard (I haven't checked a later standard). There
723        was an instance in pcretest which caused it to go wrong when processing
724        \x{...} escapes in subject strings. The has been rewritten to avoid using
725        things like p++ in the argument of isxdigit().
726    
727    
728    Version 8.13 16-Aug-2011
729    ------------------------
730    
731    1.  The Unicode data tables have been updated to Unicode 6.0.0.
732    
733    2.  Two minor typos in pcre_internal.h have been fixed.
734    
735    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
736        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
737        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
738    
739    4.  There were a number of related bugs in the code for matching backrefences
740        caselessly in UTF-8 mode when codes for the characters concerned were
741        different numbers of bytes. For example, U+023A and U+2C65 are an upper
742        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
743        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
744        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
745        2-byte code at the end of the subject (it thought there wasn't enough data
746        left).
747    
748    5.  Comprehensive information about what went wrong is now returned by
749        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
750        as the output vector has at least 2 elements. The offset of the start of
751        the failing character and a reason code are placed in the vector.
752    
753    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
754        now returned is for the first byte of the failing character, instead of the
755        last byte inspected. This is an incompatible change, but I hope it is small
756        enough not to be a problem. It makes the returned offset consistent with
757        pcre_exec() and pcre_dfa_exec().
758    
759    7.  pcretest now gives a text phrase as well as the error number when
760        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
761        failure, the offset and reason code are output.
762    
763    8.  When \R was used with a maximizing quantifier it failed to skip backwards
764        over a \r\n pair if the subsequent match failed. Instead, it just skipped
765        back over a single character (\n). This seems wrong (because it treated the
766        two characters as a single entity when going forwards), conflicts with the
767        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
768        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
769        has been changed.
770    
771    9.  Some internal refactoring has changed the processing so that the handling
772        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
773        time (the PCRE_DOTALL option was changed this way some time ago: version
774        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
775        which was always a bit of a fudge. It also means that there is one less
776        argument for the match() function, which reduces its stack requirements
777        slightly. This change also fixes an incompatibility with Perl: the pattern
778        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
779    
780    10. More internal refactoring has drastically reduced the number of recursive
781        calls to match() for possessively repeated groups such as (abc)++ when
782        using pcre_exec().
783    
784    11. While implementing 10, a number of bugs in the handling of groups were
785        discovered and fixed:
786    
787        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
788        (a|)*(?1) gave a compile-time internal error.
789        ((a|)+)+  did not notice that the outer group could match an empty string.
790        (^a|^)+   was not marked as anchored.
791        (.*a|.*)+ was not marked as matching at start or after a newline.
792    
793    12. Yet more internal refactoring has removed another argument from the match()
794        function. Special calls to this function are now indicated by setting a
795        value in a variable in the "match data" data block.
796    
797    13. Be more explicit in pcre_study() instead of relying on "default" for
798        opcodes that mean there is no starting character; this means that when new
799        ones are added and accidentally left out of pcre_study(), testing should
800        pick them up.
801    
802    14. The -s option of pcretest has been documented for ages as being an old
803        synonym of -m (show memory usage). I have changed it to mean "force study
804        for every regex", that is, assume /S for every regex. This is similar to -i
805        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
806        using it. It makes it easier to run collections of tests with and without
807        study enabled, and thereby test pcre_study() more easily. All the standard
808        tests are now run with and without -s (but some patterns can be marked as
809        "never study" - see 20 below).
810    
811    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
812        restoration of the capturing data to the outer values was not happening
813        correctly.
814    
815    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
816        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
817        pattern had matched an empty string, and so incorrectly returned a no
818        match.
819    
820    17. There was optimizing code for the last branch of non-capturing parentheses,
821        and also for the obeyed branch of a conditional subexpression, which used
822        tail recursion to cut down on stack usage. Unfortunately, now that there is
823        the possibility of (*THEN) occurring in these branches, tail recursion is
824        no longer possible because the return has to be checked for (*THEN). These
825        two optimizations have therefore been removed. [But see 8.20/11 above.]
826    
827    18. If a pattern containing \R was studied, it was assumed that \R always
828        matched two bytes, thus causing the minimum subject length to be
829        incorrectly computed because \R can also match just one byte.
830    
831    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
832        was incorrectly computed.
833    
834    20. If /S is present twice on a test pattern in pcretest input, it now
835        *disables* studying, thereby overriding the use of -s on the command line
836        (see 14 above). This is necessary for one or two tests to keep the output
837        identical in both cases.
838    
839    21. When (*ACCEPT) was used in an assertion that matched an empty string and
840        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
841    
842    22. When an atomic group that contained a capturing parenthesis was
843        successfully matched, but the branch in which it appeared failed, the
844        capturing was not being forgotten if a higher numbered group was later
845        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
846        group 1 to "a", when in fact it should be unset. This applied to multi-
847        branched capturing and non-capturing groups, repeated or not, and also to
848        positive assertions (capturing in negative assertions does not happen
849        in PCRE) and also to nested atomic groups.
850    
851    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
852        subject after a captured substring, to make it easier to tell which of a
853        number of identical substrings has been captured.
854    
855    24. The way atomic groups are processed by pcre_exec() has been changed so that
856        if they are repeated, backtracking one repetition now resets captured
857        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
858        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
859        "aaa". Previously, it would have been "a". As part of this code
860        refactoring, the way recursive calls are handled has also been changed.
861    
862    25. If an assertion condition captured any substrings, they were not passed
863        back unless some other capturing happened later. For example, if
864        (?(?=(a))a) was matched against "a", no capturing was returned.
865    
866    26. When studying a pattern that contained subroutine calls or assertions,
867        the code for finding the minimum length of a possible match was handling
868        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
869        group 1 called group 2 while simultaneously a separate group 2 called group
870        1). A stack overflow occurred in this case. I have fixed this by limiting
871        the recursion depth to 10.
872    
873    27. Updated RunTest.bat in the distribution to the version supplied by Tom
874        Fortmann. This supports explicit test numbers on the command line, and has
875        argument validation and error reporting.
876    
877    28. An instance of \X with an unlimited repeat could fail if at any point the
878        first character it looked at was a mark character.
879    
880    29. Some minor code refactoring concerning Unicode properties and scripts
881        should reduce the stack requirement of match() slightly.
882    
883    30. Added the '=' option to pcretest to check the setting of unused capturing
884        slots at the end of the pattern, which are documented as being -1, but are
885        not included in the return count.
886    
887    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
888        compiled something random. Now it gives a compile-time error (as does
889        Perl).
890    
891    32. A *MARK encountered during the processing of a positive assertion is now
892        recorded and passed back (compatible with Perl).
893    
894    33. If --only-matching or --colour was set on a pcregrep call whose pattern
895        had alternative anchored branches, the search for a second match in a line
896        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
897        matched the line "0102" twice. The same bug affected patterns that started
898        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
899        twice.
900    
901    34. Previously, PCRE did not allow quantification of assertions. However, Perl
902        does, and because of capturing effects, quantifying parenthesized
903        assertions may at times be useful. Quantifiers are now allowed for
904        parenthesized assertions.
905    
906    35. A minor code tidy in pcre_compile() when checking options for \R usage.
907    
908    36. \g was being checked for fancy things in a character class, when it should
909        just be a literal "g".
910    
911    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
912        appearance of a nested POSIX class supersedes an apparent external class.
913        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
914        unescaped square brackets may also appear as part of class names. For
915        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
916        more like Perl. (But see 8.20/1 above.)
917    
918    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
919        was because it thought it was \N{name}, which is not supported).
920    
921    39. Add minix to OS list not supporting the -S option in pcretest.
922    
923    40. PCRE tries to detect cases of infinite recursion at compile time, but it
924        cannot analyze patterns in sufficient detail to catch mutual recursions
925        such as ((?1))((?2)). There is now a runtime test that gives an error if a
926        subgroup is called recursively as a subpattern for a second time at the
927        same position in the subject string. In previous releases this might have
928        been caught by the recursion limit, or it might have run out of stack.
929    
930    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
931        happen only once. PCRE was, however incorrectly giving a compile time error
932        "recursive call could loop indefinitely" because it cannot analyze the
933        pattern in sufficient detail. The compile time test no longer happens when
934        PCRE is compiling a conditional subpattern, but actual runaway loops are
935        now caught at runtime (see 40 above).
936    
937    42. It seems that Perl allows any characters other than a closing parenthesis
938        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
939        has been changed to be the same.
940    
941    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
942        as not to get warnings when autogen.sh is called. Also changed
943        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
944    
945    44. To help people who use pcregrep to scan files containing exceedingly long
946        lines, the following changes have been made:
947    
948        (a) The default value of the buffer size parameter has been increased from
949            8K to 20K. (The actual buffer used is three times this size.)
950    
951        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
952            PCRE is built.
953    
954        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
955            to be set at run time.
956    
957        (d) Numerical values in pcregrep options can be followed by K or M, for
958            example --buffer-size=50K.
959    
960        (e) If a line being scanned overflows pcregrep's buffer, an error is now
961            given and the return code is set to 2.
962    
963    45. Add a pointer to the latest mark to the callout data block.
964    
965    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
966        partial match of an empty string instead of no match. This was specific to
967        the use of ".".
968    
969    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
970        complete match instead of a partial match. This bug was dependent on both
971        the PCRE_UTF8 and PCRE_DOTALL options being set.
972    
973    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
974        starting byte set, because \b was not being ignored.
975    
976    
977    Version 8.12 15-Jan-2011
978    ------------------------
979    
980    1.  Fixed some typos in the markup of the man pages, and wrote a script that
981        checks for such things as part of the documentation building process.
982    
983    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
984        --match-limit and --recursion-limit options (added for 8.11). In
985        particular, this made one of the standard tests fail. (The integer value
986        went into the wrong half of a long int.)
987    
988    3.  If the --colour option was given to pcregrep with -v (invert match), it
989        did strange things, either producing crazy output, or crashing. It should,
990        of course, ignore a request for colour when reporting lines that do not
991        match.
992    
993    4.  Another pcregrep bug caused similar problems if --colour was specified with
994        -M (multiline) and the pattern match finished with a line ending.
995    
996    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
997        matched in multiline mode, the following line was shown as part of the
998        match. This seems wrong, so I have changed it.
999    
1000    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1001        the check for further matches in the same line (so they could be coloured)
1002        to overrun the end of the current line. If another match was found, it was
1003        incorrectly shown (and then shown again when found in the next line).
1004    
1005    7.  If pcregrep was compiled under Windows, there was a reference to the
1006        function pcregrep_exit() before it was defined. I am assuming this was
1007        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1008        reported by a user. I've moved the definition above the reference.
1009    
1010    
1011    Version 8.11 10-Dec-2010
1012    ------------------------
1013    
1014    1.  (*THEN) was not working properly if there were untried alternatives prior
1015        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1016        backtracked to try for "b" instead of moving to the next alternative branch
1017        at the same level (in this case, to look for "c"). The Perl documentation
1018        is clear that when (*THEN) is backtracked onto, it goes to the "next
1019        alternative in the innermost enclosing group".
1020    
1021    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1022        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1023        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1024        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1025        (*THEN).
1026    
1027    3.  If \s appeared in a character class, it removed the VT character from
1028        the class, even if it had been included by some previous item, for example
1029        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1030        of \s, but is part of the POSIX "space" class.)
1031    
1032    4.  A partial match never returns an empty string (because you can always
1033        match an empty string at the end of the subject); however the checking for
1034        an empty string was starting at the "start of match" point. This has been
1035        changed to the "earliest inspected character" point, because the returned
1036        data for a partial match starts at this character. This means that, for
1037        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1038        (previously it gave "no match").
1039    
1040    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1041        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1042        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1043        has an implication that the given string is incomplete (because a partial
1044        match is preferred over a full match). For this reason, these items now
1045        give a partial match in this situation. [Aside: previously, the one case
1046        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1047        match rather than a full match, which was wrong by the old rules, but is
1048        now correct.]
1049    
1050    6.  There was a bug in the handling of #-introduced comments, recognized when
1051        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1052        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1053        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1054        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1055        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1056        places in pcre_compile().
1057    
1058    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1059        comments when looking ahead for named forward references to subpatterns,
1060        the only newline sequence it recognized was NL. It now handles newlines
1061        according to the set newline convention.
1062    
1063    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1064        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1065        cater for a lack of strerror(). These oversights have been fixed.
1066    
1067    9.  Added --match-limit and --recursion-limit to pcregrep.
1068    
1069    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1070    
1071    11. When the -o option was used, pcregrep was setting a return code of 1, even
1072        when matches were found, and --line-buffered was not being honoured.
1073    
1074    12. Added an optional parentheses number to the -o and --only-matching options
1075        of pcregrep.
1076    
1077    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1078        can match an empty string. The code to do it in pcretest and pcredemo
1079        needed fixing:
1080    
1081        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1082            only one byte after an empty string match just before CRLF (this case
1083            just got forgotten; "any" and "anycrlf" were OK).
1084    
1085        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1086            mode when an empty string match preceded an ASCII character followed by
1087            a non-ASCII character. (The code for advancing by one character rather
1088            than one byte was nonsense.)
1089    
1090        (c) The pcredemo.c sample program did not have any code at all to handle
1091            the cases when CRLF is a valid newline sequence.
1092    
1093    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1094        as a starting offset was within the subject string. There is now a new
1095        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1096        negative or greater than the length of the string. In order to test this,
1097        pcretest is extended to allow the setting of negative starting offsets.
1098    
1099    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1100        starting offset points to the beginning of a UTF-8 character was
1101        unnecessarily clumsy. I tidied it up.
1102    
1103    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1104        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1105    
1106    17. Nobody had reported that the --include_dir option, which was added in
1107        release 7.7 should have been called --include-dir (hyphen, not underscore)
1108        for compatibility with GNU grep. I have changed it to --include-dir, but
1109        left --include_dir as an undocumented synonym, and the same for
1110        --exclude-dir, though that is not available in GNU grep, at least as of
1111        release 2.5.4.
1112    
1113    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1114        characters from a string of bytes) have been redefined so as not to use
1115        loops, in order to improve performance in some environments. At the same
1116        time, I abstracted some of the common code into auxiliary macros to save
1117        repetition (this should not affect the compiled code).
1118    
1119    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1120        compile-time error is now given if \c is not followed by an ASCII
1121        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1122        different, and any byte value is allowed.)
1123    
1124    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1125        START_OPTIMIZE option, which is now allowed at compile time - but just
1126        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1127        to pcregrep and other applications that have no direct access to PCRE
1128        options. The new /Y option in pcretest sets this option when calling
1129        pcre_compile().
1130    
1131    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1132        back references. Groups containing recursive back references were forced to
1133        be atomic by that change, but in the case of named groups, the amount of
1134        memory required was incorrectly computed, leading to "Failed: internal
1135        error: code overflow". This has been fixed.
1136    
1137    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1138        pcretest.c, to avoid build problems in some Borland environments.
1139    
1140    
1141    Version 8.10 25-Jun-2010
1142    ------------------------
1143    
1144    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1145        THEN.
1146    
1147    2.  (*ACCEPT) was not working when inside an atomic group.
1148    
1149    3.  Inside a character class, \B is treated as a literal by default, but
1150        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1151        causes the error). The code is unchanged, but I tidied the documentation.
1152    
1153    4.  Inside a character class, PCRE always treated \R and \X as literals,
1154        whereas Perl faults them if its -w option is set. I have changed PCRE so
1155        that it faults them when PCRE_EXTRA is set.
1156    
1157    5.  Added support for \N, which always matches any character other than
1158        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1159    
1160    6.  When compiling pcregrep with newer versions of gcc which may have
1161        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1162        declared with attribute warn_unused_result" were given. Just casting the
1163        result to (void) does not stop the warnings; a more elaborate fudge is
1164        needed. I've used a macro to implement this.
1165    
1166    7.  Minor change to pcretest.c to avoid a compiler warning.
1167    
1168    8.  Added four artifical Unicode properties to help with an option to make
1169        \s etc use properties (see next item). The new properties are: Xan
1170        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1171    
1172    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1173        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1174        this option. Modified pcretest to add /W to test this facility. Added
1175        REG_UCP to make it available via the POSIX interface.
1176    
1177    10. Added --line-buffered to pcregrep.
1178    
1179    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1180        studied, and the match started with a letter with a code point greater than
1181        127 whose first byte was different to the first byte of the other case of
1182        the letter, the other case of this starting letter was not recognized
1183        (#976).
1184    
1185    12. If a pattern that was studied started with a repeated Unicode property
1186        test, for example, \p{Nd}+, there was the theoretical possibility of
1187        setting up an incorrect bitmap of starting bytes, but fortunately it could
1188        not have actually happened in practice until change 8 above was made (it
1189        added property types that matched character-matching opcodes).
1190    
1191    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1192        possible starting bytes for non-anchored patterns.
1193    
1194    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1195        \R, and also a number of cases that involve Unicode properties, both
1196        explicit and implicit when PCRE_UCP is set.
1197    
1198    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1199        input, it could crash or give wrong results if characters with values
1200        greater than 0xc0 were present in the subject string. (Detail: it assumed
1201        UTF-8 input when processing these items.)
1202    
1203    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1204        size_t is 64-bit (#991).
1205    
1206    17. Added a check for running out of memory when PCRE is compiled with
1207        --disable-stack-for-recursion (#990).
1208    
1209    18. If the last data line in a file for pcretest does not have a newline on
1210        the end, a newline was missing in the output.
1211    
1212    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1213        less than 128) in its various bitmaps. However, there is a facility for
1214        generating tables according to the current locale when PCRE is compiled. It
1215        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1216        characters, are recognized by isspace() and therefore were getting set in
1217        these tables, and indeed these tables seem to approximate to ISO 8859. This
1218        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1219        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1220        which of course cannot start UTF-8 characters. I have changed the code so
1221        that only real ASCII characters (less than 128) and the correct starting
1222        bytes for UTF-8 encodings are set for characters greater than 127 when in
1223        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1224        altogether.)
1225    
1226    20. Added the /T option to pcretest so as to be able to run tests with non-
1227        standard character tables, thus making it possible to include the tests
1228        used for 19 above in the standard set of tests.
1229    
1230    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1231        reference to a subpattern the other side of a comment that contains an
1232        opening parenthesis caused either an internal compiling error, or a
1233        reference to the wrong subpattern.
1234    
1235    
1236    Version 8.02 19-Mar-2010
1237    ------------------------
1238    
1239    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1240    
1241    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1242        configured.
1243    
1244    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1245        original author of that file, following a query about its status.
1246    
1247    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1248        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1249    
1250    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1251        quantifier applied to a forward-referencing subroutine call, could compile
1252        incorrect code or give the error "internal error: previously-checked
1253        referenced subpattern not found".
1254    
1255    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1256        variables to point to external functions. For these systems, therefore,
1257        pcre_malloc etc. are now initialized to local functions that call the
1258        relevant global functions.
1259    
1260    7.  There were two entries missing in the vectors called coptable and poptable
1261        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1262        I've fixed the data, and added a kludgy way of testing at compile time that
1263        the lengths are correct (equal to the number of opcodes).
1264    
1265    8.  Following on from 7, I added a similar kludge to check the length of the
1266        eint vector in pcreposix.c.
1267    
1268    9.  Error texts for pcre_compile() are held as one long string to avoid too
1269        much relocation at load time. To find a text, the string is searched,
1270        counting zeros. There was no check for running off the end of the string,
1271        which could happen if a new error number was added without updating the
1272        string.
1273    
1274    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1275    
1276    11. \K was not working if it appeared in an atomic group or in a group that
1277        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1278        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1279        the assertion is positive, but not if it is negative.
1280    
1281    12. Change 11 fortuitously reduced the size of the stack frame used in the
1282        "match()" function of pcre_exec.c by one pointer. Forthcoming
1283        implementation of support for (*MARK) will need an extra pointer on the
1284        stack; I have reserved it now, so that the stack frame size does not
1285        decrease.
1286    
1287    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1288        item in branch that calls a recursion is a subroutine call - as in the
1289        second branch in the above example - was incorrectly given the compile-
1290        time error "recursive call could loop indefinitely" because pcre_compile()
1291        was not correctly checking the subroutine for matching a non-empty string.
1292    
1293    14. The checks for overrunning compiling workspace could trigger after an
1294        overrun had occurred. This is a "should never occur" error, but it can be
1295        triggered by pathological patterns such as hundreds of nested parentheses.
1296        The checks now trigger 100 bytes before the end of the workspace.
1297    
1298    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1299    
1300    
1301    Version 8.01 19-Jan-2010
1302    ------------------------
1303    
1304    1.  If a pattern contained a conditional subpattern with only one branch (in
1305        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1306        computed the wrong minimum data length (which is of course zero for such
1307        subpatterns). This could cause incorrect "no match" results.
1308    
1309    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1310        the pattern is reset in the first branch, pcre_compile() failed with
1311        "internal error: code overflow at offset...". This happened only when
1312        the reset was to the original external option setting. (An optimization
1313        abstracts leading options settings into an external setting, which was the
1314        cause of this.)
1315    
1316    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1317        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1318        assertion pattern did not match (meaning that the assertion was true), it
1319        was incorrectly treated as false if the SKIP had been reached during the
1320        matching. This also applied to assertions used as conditions.
1321    
1322    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1323        assertion subpattern, including such a pattern used as a condition,
1324        unpredictable results occurred, instead of the error return
1325        PCRE_ERROR_DFA_UITEM.
1326    
1327    5.  The C++ GlobalReplace function was not working like Perl for the special
1328        situation when an empty string is matched. It now does the fancy magic
1329        stuff that is necessary.
1330    
1331    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1332        removed. (These were left over from very, very early versions of PCRE.)
1333    
1334    7.  Some cosmetic changes to the code to make life easier when compiling it
1335        as part of something else:
1336    
1337        (a) Change DEBUG to PCRE_DEBUG.
1338    
1339        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1340            called "current" as "current_branch", to prevent a collision with the
1341            Linux macro when compiled as a kernel module.
1342    
1343        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1344            prevent a collision with the Linux macro when compiled as a kernel
1345            module.
1346    
1347    8.  In pcre_compile() there are some checks for integer overflows that used to
1348        cast potentially large values to (double). This has been changed to that
1349        when building, a check for int64_t is made, and if it is found, it is used
1350        instead, thus avoiding the use of floating point arithmetic. (There is no
1351        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1352        double.
1353    
1354    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1355        2005 (difference between two addresses compared to an unsigned value).
1356    
1357    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1358        custom one, because of the following reported problem in Windows:
1359    
1360          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1361              under Win32.
1362          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1363              therefore missing the function definition.
1364          - The compiler thus generates a "C" signature for the test function.
1365          - The linker fails to find the "C" function.
1366          - PCRE fails to configure if asked to do so against libbz2.
1367    
1368    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1369        messages were output:
1370    
1371          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1372          rerunning libtoolize, to keep the correct libtool macros in-tree.
1373          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1374    
1375        I have done both of these things.
1376    
1377    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1378        most of the time, it *can* run out if it is given a pattern that contains a
1379        runaway infinite recursion. I updated the discussion in the pcrestack man
1380        page.
1381    
1382    13. Now that we have gone to the x.xx style of version numbers, the minor
1383        version may start with zero. Using 08 or 09 is a bad idea because users
1384        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1385        interpreted as invalid octal numbers. I've updated the previous comment in
1386        configure.ac, and also added a check that gives an error if 08 or 09 are
1387        used.
1388    
1389    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1390        causing partial matching to fail when the end of the subject matched \W
1391        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1392    
1393    15. There were some discrepancies between the declarations in pcre_internal.h
1394        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1395        their definitions. The declarations used "const uschar *" and the
1396        definitions used USPTR. Even though USPTR is normally defined as "const
1397        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1398        reported that: "This difference in casting confuses some C++ compilers, for
1399        example, SunCC recognizes above declarations as different functions and
1400        generates broken code for hbpcre." I have changed the declarations to use
1401        USPTR.
1402    
1403    16. GNU libtool is named differently on some systems. The autogen.sh script now
1404        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1405        (FreeBSD).
1406    
1407    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1408        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1409        comment: "Figure out how to create a longlong from a string: strtoll and
1410        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1411        instance, but it only takes 2 args instead of 3!"
1412    
1413    18. A subtle bug concerned with back references has been fixed by a change of
1414        specification, with a corresponding code fix. A pattern such as
1415        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1416        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1417        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1418        same bug. Such groups have to be quantified to be useful, or contained
1419        inside another quantified group. (If there's no repetition, the reference
1420        can never match.) The problem arises because, having left the group and
1421        moved on to the rest of the pattern, a later failure that backtracks into
1422        the group uses the captured value from the final iteration of the group
1423        rather than the correct earlier one. I have fixed this in PCRE by forcing
1424        any group that contains a reference to itself to be an atomic group; that
1425        is, there cannot be any backtracking into it once it has completed. This is
1426        similar to recursive and subroutine calls.
1427    
1428    
1429    Version 8.00 19-Oct-09
1430    ----------------------
1431    
1432    1.  The table for translating pcre_compile() error codes into POSIX error codes
1433        was out-of-date, and there was no check on the pcre_compile() error code
1434        being within the table. This could lead to an OK return being given in
1435        error.
1436    
1437    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1438        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1439        in a Windows environment.
1440    
1441    3.  The pcregrep --count option prints the count for each file even when it is
1442        zero, as does GNU grep. However, pcregrep was also printing all files when
1443        --files-with-matches was added. Now, when both options are given, it prints
1444        counts only for those files that have at least one match. (GNU grep just
1445        prints the file name in this circumstance, but including the count seems
1446        more useful - otherwise, why use --count?) Also ensured that the
1447        combination -clh just lists non-zero counts, with no names.
1448    
1449    4.  The long form of the pcregrep -F option was incorrectly implemented as
1450        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1451        but it seems right to fix it, and I didn't think it was worth preserving
1452        the old behaviour.
1453    
1454    5.  The command line items --regex=pattern and --regexp=pattern were not
1455        recognized by pcregrep, which required --regex pattern or --regexp pattern
1456        (with a space rather than an '='). The man page documented the '=' forms,
1457        which are compatible with GNU grep; these now work.
1458    
1459    6.  No libpcreposix.pc file was created for pkg-config; there was just
1460        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1461    
1462    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1463        when UCP support is not needed, by modifying the Python script that
1464        generates it from Unicode data files. This should not matter if the module
1465        is correctly used as a library, but I received one complaint about 50K of
1466        unwanted data. My guess is that the person linked everything into his
1467        program rather than using a library. Anyway, it does no harm.
1468    
1469    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1470        was a minimum greater than 1 for a wide character in a possessive
1471        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1472        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1473        character. Chaos in the form of incorrect output or a compiling loop could
1474        result.
1475    
1476    9.  The restrictions on what a pattern can contain when partial matching is
1477        requested for pcre_exec() have been removed. All patterns can now be
1478        partially matched by this function. In addition, if there are at least two
1479        slots in the offset vector, the offset of the earliest inspected character
1480        for the match and the offset of the end of the subject are set in them when
1481        PCRE_ERROR_PARTIAL is returned.
1482    
1483    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1484        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1485        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1486        and may be more useful for multi-segment matching.
1487    
1488    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1489        used to be given if ever the end of the subject was reached; now it is
1490        given only if matching could not proceed because another character was
1491        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1492        string "Z", which now yields "no match" instead of "partial match". In the
1493        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1494        final character ended with (*FAIL).
1495    
1496    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1497        if the pattern had a "must contain" character that was already found in the
1498        earlier partial match, unless partial matching was again requested. For
1499        example, with the pattern /dog.(body)?/, the "must contain" character is
1500        "g". If the first part-match was for the string "dog", restarting with
1501        "sbody" failed. This bug has been fixed.
1502    
1503    13. The string returned by pcre_dfa_exec() after a partial match has been
1504        changed so that it starts at the first inspected character rather than the
1505        first character of the match. This makes a difference only if the pattern
1506        starts with a lookbehind assertion or \b or \B (\K is not supported by
1507        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1508        matching functions compatible, and I think it's the right thing to do.
1509    
1510    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1511        so that the demonstration program is easily available in environments where
1512        PCRE has not been installed from source.
1513    
1514    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1515        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1516        library.
1517    
1518    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1519        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1520        is not the first non-POSIX option to be added. Clearly some people find
1521        these options useful.
1522    
1523    17. If a caller to the POSIX matching function regexec() passes a non-zero
1524        value for nmatch with a NULL value for pmatch, the value of
1525        nmatch is forced to zero.
1526    
1527    18. RunGrepTest did not have a test for the availability of the -u option of
1528        the diff command, as RunTest does. It now checks in the same way as
1529        RunTest, and also checks for the -b option.
1530    
1531    19. If an odd number of negated classes containing just a single character
1532        interposed, within parentheses, between a forward reference to a named
1533        subpattern and the definition of the subpattern, compilation crashed with
1534        an internal error, complaining that it could not find the referenced
1535        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1536        [The bug was that it was starting one character too far in when skipping
1537        over the character class, thus treating the ] as data rather than
1538        terminating the class. This meant it could skip too much.]
1539    
1540    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1541        /g option in pcretest when the pattern contains \K, which makes it possible
1542        to have an empty string match not at the start, even when the pattern is
1543        anchored. Updated pcretest and pcredemo to use this option.
1544    
1545    21. If the maximum number of capturing subpatterns in a recursion was greater
1546        than the maximum at the outer level, the higher number was returned, but
1547        with unset values at the outer level. The correct (outer level) value is
1548        now given.
1549    
1550    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1551        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1552        make it do so. The string so far is captured, making this feature
1553        compatible with Perl.
1554    
1555    23. The tests have been re-organized, adding tests 11 and 12, to make it
1556        possible to check the Perl 5.10 features against Perl 5.10.
1557    
1558    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1559        pattern matches a fixed length string. PCRE did not allow this; now it
1560        does. Neither allows recursion.
1561    
1562    25. I finally figured out how to implement a request to provide the minimum
1563        length of subject string that was needed in order to match a given pattern.
1564        (It was back references and recursion that I had previously got hung up
1565        on.) This code has now been added to pcre_study(); it finds a lower bound
1566        to the length of subject needed. It is not necessarily the greatest lower
1567        bound, but using it to avoid searching strings that are too short does give
1568        some useful speed-ups. The value is available to calling programs via
1569        pcre_fullinfo().
1570    
1571    26. While implementing 25, I discovered to my embarrassment that pcretest had
1572        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1573        study optimizations had never been tested with that matching function.
1574        Oops. What is worse, even when it was passed study data, there was a bug in
1575        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1576        were also very few tests of studied patterns with pcre_dfa_exec().
1577    
1578    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1579        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1580        on the other side of the coin, they are no longer allowed to have different
1581        names, because these cannot be distinguished in PCRE, and this has caused
1582        confusion. (This is a difference from Perl.)
1583    
1584    28. When duplicate subpattern names are present (necessarily with different
1585        numbers, as required by 27 above), and a test is made by name in a
1586        conditional pattern, either for a subpattern having been matched, or for
1587        recursion in such a pattern, all the associated numbered subpatterns are
1588        tested, and the overall condition is true if the condition is true for any
1589        one of them. This is the way Perl works, and is also more like the way
1590        testing by number works.
1591    
1592    
1593    Version 7.9 11-Apr-09
1594    ---------------------
1595    
1596    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1597        (pcretest), all targets were linked against these libraries. This included
1598        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1599        libraries. This caused unwanted dependencies to be created. This problem
1600        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1601        pcretest is linked with readline.
1602    
1603    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1604        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1605        moved outside it again, because FALSE and TRUE are already defined in AIX,
1606        but BOOL is not.
1607    
1608    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1609        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1610    
1611    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1612        hyphens) following file names and line numbers when outputting matching
1613        lines. This is not true; no spaces are inserted. I have also clarified the
1614        wording for the --colour (or --color) option.
1615    
1616    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1617        was not coloured; this is different to GNU grep, so I have changed it to be
1618        the same.
1619    
1620    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1621        each matching line was coloured. Now it goes on to look for further matches
1622        of any of the test patterns, which is the same behaviour as GNU grep.
1623    
1624    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1625        doesn't make sense to accept an empty string match in pcregrep, so I have
1626        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1627        seems to be how GNU grep behaves.
1628    
1629    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1630        start or after a newline", because the conditional assertion was not being
1631        correctly handled. The rule now is that both the assertion and what follows
1632        in the first alternative must satisfy the test.
1633    
1634    9.  If auto-callout was enabled in a pattern with a conditional group whose
1635        condition was an assertion, PCRE could crash during matching, both with
1636        pcre_exec() and pcre_dfa_exec().
1637    
1638    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1639        used for matching.
1640    
1641    11. Unicode property support in character classes was not working for
1642        characters (bytes) greater than 127 when not in UTF-8 mode.
1643    
1644    12. Added the -M command line option to pcretest.
1645    
1646    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1647    
1648    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1649    
1650    16. Added comments and documentation about mis-use of no_arg in the C++
1651        wrapper.
1652    
1653    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1654        from Martin Jerabek that uses macro names for all relevant character and
1655        string constants.
1656    
1657    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1658        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1659        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1660        these, but not everybody uses configure.
1661    
1662    19. A conditional group that had only one branch was not being correctly
1663        recognized as an item that could match an empty string. This meant that an
1664        enclosing group might also not be so recognized, causing infinite looping
1665        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1666        with the subject "ab", where knowledge that the repeated group can match
1667        nothing is needed in order to break the loop.
1668    
1669    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1670        exec(), but without supplying a callout function, matching went wrong.
1671    
1672    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1673        leak if the size of the offset vector was greater than 30. When the vector
1674        is smaller, the saved offsets during recursion go onto a local stack
1675        vector, but for larger vectors malloc() is used. It was failing to free
1676        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1677        error, in fact).
1678    
1679    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1680        heapframe that is used only when UTF-8 support is enabled. This caused no
1681        problem, but was untidy.
1682    
1683    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1684        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1685        included within another project.
1686    
1687    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1688        slightly modified by me:
1689    
1690          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1691              not building pcregrep.
1692    
1693          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1694              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1695    
1696    25. Forward references, both numeric and by name, in patterns that made use of
1697        duplicate group numbers, could behave incorrectly or give incorrect errors,
1698        because when scanning forward to find the reference group, PCRE was not
1699        taking into account the duplicate group numbers. A pattern such as
1700        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1701    
1702    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1703        the feature of a custom pointer more persuasive (as requested by a user).
1704    
1705    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1706        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1707        pre-defined.
1708    
1709    28. Added support for (*UTF8) at the start of a pattern.
1710    
1711    29. Arrange for flags added by the "release type" setting in CMake to be shown
1712        in the configuration summary.
1713    
1714    
1715    Version 7.8 05-Sep-08
1716    ---------------------
1717    
1718    1.  Replaced UCP searching code with optimized version as implemented for Ad
1719        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1720        stage table and inline lookup instead of a function, giving speed ups of 2
1721        to 5 times on some simple patterns that I tested. Permission was given to
1722        distribute the MultiStage2.py script that generates the tables (it's not in
1723        the tarball, but is in the Subversion repository).
1724    
1725    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1726        scripts.
1727    
1728    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1729        a group with a zero qualifier. The result of the study could be incorrect,
1730        or the function might crash, depending on the pattern.
1731    
1732    4.  Caseless matching was not working for non-ASCII characters in back
1733        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1734        It now works when Unicode Property Support is available.
1735    
1736    5.  In pcretest, an escape such as \x{de} in the data was always generating
1737        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1738        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1739        truncation.
1740    
1741    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1742    
1743    7.  Added two (int) casts to pcregrep when printing the difference of two
1744        pointers, in case they are 64-bit values.
1745    
1746    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1747        test 2 if it fails.
1748    
1749    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1750        and a #define of that name to empty if it is not externally set. This is to
1751        allow users of MSVC to set it if necessary.
1752    
1753    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1754        the convenience functions in the pcre_get.c source file.
1755    
1756    11. An option change at the start of a pattern that had top-level alternatives
1757        could cause overwriting and/or a crash. This command provoked a crash in
1758        some environments:
1759    
1760          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1761    
1762        This potential security problem was recorded as CVE-2008-2371.
1763    
1764    12. For a pattern where the match had to start at the beginning or immediately
1765        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1766        pcre_dfa_exec() could read past the end of the passed subject if there was
1767        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1768        pcretest so that it places the subject at the end of its malloc-ed buffer.
1769    
1770    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1771        exec() might read past the end of the data buffer in UTF-8 mode.
1772    
1773    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1774        the data contained the byte 0x85 as part of a UTF-8 character within its
1775        first line. This applied both to normal and DFA matching.
1776    
1777    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1778        /^[^d]*?$/8 failed to match "abc".
1779    
1780    16. Added a missing copyright notice to pcrecpp_internal.h.
1781    
1782    17. Make it more clear in the documentation that values returned from
1783        pcre_exec() in ovector are byte offsets, not character counts.
1784    
1785    18. Tidied a few places to stop certain compilers from issuing warnings.
1786    
1787    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1788        supplied by Stefan Weber. I made a further small update for 7.8 because
1789        there is a change of source arrangements: the pcre_searchfuncs.c module is
1790        replaced by pcre_ucd.c.
1791    
1792    
1793    Version 7.7 07-May-08
1794    ---------------------
1795    
1796    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1797        a string to a long long, pretend we don't even have a long long." This is
1798        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1799    
1800    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1801        pre-7.6 versions, which defined a global no_arg variable instead of putting
1802        it in the RE class. (See also #8 below.)
1803    
1804    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1805        Lopes.
1806    
1807    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1808    
1809        (1) The include/exclude patterns were being applied to the whole pathnames
1810            of files, instead of just to the final components.
1811    
1812        (2) If there was more than one level of directory, the subdirectories were
1813            skipped unless they satisfied the include/exclude conditions. This is
1814            inconsistent with GNU grep (and could even be seen as contrary to the
1815            pcregrep specification - which I improved to make it absolutely clear).
1816            The action now is always to scan all levels of directory, and just
1817            apply the include/exclude patterns to regular files.
1818    
1819    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1820        --exclude_dir in the tests to avoid scanning .svn directories.
1821    
1822    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1823        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1824        doesn't support NULs in patterns.
1825    
1826    7.  Added some missing "const"s to declarations of static tables in
1827        pcre_compile.c and pcre_dfa_exec.c.
1828    
1829    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1830        caused by fix #2  above. (Subsequently also a second patch to fix the
1831        first patch. And a third patch - this was a messy problem.)
1832    
1833    9.  Applied Craig's patch to remove the use of push_back().
1834    
1835    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1836        matching function regexec().
1837    
1838    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1839        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1840        references. PCRE supports relative numbers with this syntax (I don't think
1841        Oniguruma does).
1842    
1843    12. Previously, a group with a zero repeat such as (...){0} was completely
1844        omitted from the compiled regex. However, this means that if the group
1845        was called as a subroutine from elsewhere in the pattern, things went wrong
1846        (an internal error was given). Such groups are now left in the compiled
1847        pattern, with a new opcode that causes them to be skipped at execution
1848        time.
1849    
1850    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1851        to the way PCRE behaves:
1852    
1853        (a) A lone ] character is dis-allowed (Perl treats it as data).
1854    
1855        (b) A back reference to an unmatched subpattern matches an empty string
1856            (Perl fails the current match path).
1857    
1858        (c) A data ] in a character class must be notated as \] because if the
1859            first data character in a class is ], it defines an empty class. (In
1860            Perl it is not possible to have an empty class.) The empty class []
1861            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1862            The negative empty class [^] matches any one character, independently
1863            of the DOTALL setting.
1864    
1865    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1866        non-existent subpattern following a character class starting with ']' and
1867        containing () gave an internal compiling error instead of "reference to
1868        non-existent subpattern". Fortunately, when the pattern did exist, the
1869        compiled code was correct. (When scanning forwards to check for the
1870        existencd of the subpattern, it was treating the data ']' as terminating
1871        the class, so got the count wrong. When actually compiling, the reference
1872        was subsequently set up correctly.)
1873    
1874    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1875        it was being rejected as not supported by pcre_dfa_exec(), even though
1876        other assertions are supported. I have made pcre_dfa_exec() support
1877        (*FAIL).
1878    
1879    16. The implementation of 13c above involved the invention of a new opcode,
1880        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1881        cannot be changed at match time, I realized I could make a small
1882        improvement to matching performance by compiling OP_ALLANY instead of
1883        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1884        on the OP_ANY path.
1885    
1886    17. Compiling pcretest on Windows with readline support failed without the
1887        following two fixes: (1) Make the unistd.h include conditional on
1888        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1889    
1890    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1891        ncurses library to be included for pcretest when ReadLine support is
1892        requested, but also to allow for it to be overridden. This patch came from
1893        Daniel Bergström.
1894    
1895    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1896        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1897        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1898        spotting this.
1899    
1900    
1901    Version 7.6 28-Jan-08
1902    ---------------------
1903    
1904    1.  A character class containing a very large number of characters with
1905        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1906        overflow.
1907    
1908    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1909        HAVE_LONG_LONG is not defined.
1910    
1911    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1912        bring them up to date and include new features. This patch includes:
1913    
1914        - Fixed PH's badly added libz and libbz2 support.
1915        - Fixed a problem with static linking.
1916        - Added pcredemo. [But later removed - see 7 below.]
1917        - Fixed dftables problem and added an option.
1918        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1919            HAVE_LONG_LONG.
1920        - Added readline support for pcretest.
1921        - Added an listing of the option settings after cmake has run.
1922    
1923    4.  A user submitted a patch to Makefile that makes it easy to create
1924        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1925        Makefile.am that cause it to include this special target, without
1926        affecting anything else. Note that the same mingw target plus all
1927        the other distribution libraries and programs are now supported
1928        when configuring with CMake (see 6 below) instead of with
1929        Configure/Make.
1930    
1931    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1932        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1933        exported in the Windows port". It has not yet been confirmed that the patch
1934        solves the problem, but it does no harm.
1935    
1936    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1937        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1938        with CMake, and also correct the comment about stack recursion.
1939    
1940    7.  Remove the automatic building of pcredemo from the ./configure system and
1941        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1942        of a program that users should build themselves after PCRE is installed, so
1943        building it automatically is not really right. What is more, it gave
1944        trouble in some build environments.
1945    
1946    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1947    
1948    
1949    Version 7.5 10-Jan-08
1950    ---------------------
1951    
1952    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1953        values in parens when parsing an RE using the C++ wrapper."
1954    
1955    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1956        Characters greater than 255 were excluded from the class instead of being
1957        included.
1958    
1959    3.  The same bug as (2) above applied to negated POSIX classes such as
1960        [:^space:].
1961    
1962    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1963        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1964        I have changed it.
1965    
1966    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1967        first named subpattern) and a construct such as (?&a) would reference the
1968        first named subpattern whose name started with "a" (in other words, the
1969        length check was missing). Both these problems are fixed. "Subpattern name
1970        expected" is now given for (?&) (a zero-length name), and this patch also
1971        makes it give the same error for \k'' (previously it complained that that
1972        was a reference to a non-existent subpattern).
1973    
1974    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1975        this is right because (?- can be followed by option settings as well as by
1976        digits. I have, however, made the messages clearer.
1977    
1978    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1979        than the number used in the conditional) now cause a compile-time error.
1980        This is actually not compatible with Perl, which accepts such patterns, but
1981        treats the conditional as always being FALSE (as PCRE used to), but it
1982        seems to me that giving a diagnostic is better.
1983    
1984    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1985        and messages.
1986    
1987    9.  Fix two occurrences of "backslash" in comments that should have been
1988        "backspace".
1989    
1990    10. Remove two redundant lines of code that can never be obeyed (their function
1991        was moved elsewhere).
1992    
1993    11. The program that makes PCRE's Unicode character property table had a bug
1994        which caused it to generate incorrect table entries for sequences of
1995        characters that have the same character type, but are in different scripts.
1996        It amalgamated them into a single range, with the script of the first of
1997        them. In other words, some characters were in the wrong script. There were
1998        thirteen such cases, affecting characters in the following ranges:
1999    
2000          U+002b0 - U+002c1
2001          U+0060c - U+0060d
2002          U+0061e - U+00612
2003          U+0064b - U+0065e
2004          U+0074d - U+0076d
2005          U+01800 - U+01805
2006          U+01d00 - U+01d77
2007          U+01d9b - U+01dbf
2008          U+0200b - U+0200f
2009          U+030fc - U+030fe
2010          U+03260 - U+0327f
2011          U+0fb46 - U+0fbb1
2012          U+10450 - U+1049d
2013    
2014    12. The -o option (show only the matching part of a line) for pcregrep was not
2015        compatible with GNU grep in that, if there was more than one match in a
2016        line, it showed only the first of them. It now behaves in the same way as
2017        GNU grep.
2018    
2019    13. If the -o and -v options were combined for pcregrep, it printed a blank
2020        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2021        does the same. The return code can be used to tell if there were any
2022        non-matching lines.
2023    
2024    14. Added --file-offsets and --line-offsets to pcregrep.
2025    
2026    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2027        infinitely looping recursion. The bug was that positive lookaheads were not
2028        being skipped when checking for a possible empty match (negative lookaheads
2029        and both kinds of lookbehind were skipped).
2030    
2031    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2032        inclusion of <windows.h> to before rather than after the definition of
2033        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2034    
2035    17. Specifying a possessive quantifier with a specific limit for a Unicode
2036        character property caused pcre_compile() to compile bad code, which led at
2037        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2038        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2039        caused the error; without that there was no problem.
2040    
2041    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2042    
2043    19. Added --enable-pcretest-libreadline.
2044    
2045    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2046        RE::GlobalReplace(). As a result, the number of replacements returned was
2047        double what it should be. I removed one of the increments, but Craig sent a
2048        later patch that removed the other one (the right fix) and added unit tests
2049        that check the return values (which was not done before).
2050    
2051    21. Several CMake things:
2052    
2053        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2054            the names libpcre and libpcreposix, not just pcre and pcreposix.
2055    
2056        (2) The above change means that pcretest and pcregrep are now correctly
2057            linked with the newly-built libraries, not previously installed ones.
2058    
2059        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2060    
2061    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2062        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2063        UTF-8 newline character). The key issue is that the pattern starts .*;
2064        this means that the match must be either at the beginning, or after a
2065        newline. The bug was in the code for advancing after a failed match and
2066        checking that the new position followed a newline. It was not taking
2067        account of UTF-8 characters correctly.
2068    
2069    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2070        character classes. PCRE was not treating the sequence [:...:] as a
2071        character class unless the ... were all letters. Perl, however, seems to
2072        allow any characters between [: and :], though of course it rejects as
2073        unknown any "names" that contain non-letters, because all the known class
2074        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2075        for example, whereas PCRE did not - it did not recognize a POSIX character
2076        class. This seemed a bit dangerous, so the code has been changed to be
2077        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2078        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2079        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2080        Perl does, and where it didn't before.
2081    
2082    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2083        Windows environments %n is disabled by default.
2084    
2085    
2086    Version 7.4 21-Sep-07
2087    ---------------------
2088    
2089    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2090        means that a class such as [\s] counted as "explicit reference to CR or
2091        LF". That isn't really right - the whole point of the change was to try to
2092        help when there was an actual mention of one of the two characters. So now
2093        the change happens only if \r or \n (or a literal CR or LF) character is
2094        encountered.
2095    
2096    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2097        of both had grown to the point where there were only 3 bits left.
2098        Fortunately, there was spare space in the data structure, and so I have
2099        moved the internal flags into a new 16-bit field to free up more option
2100        bits.
2101    
2102    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2103        but did not set the internal JCHANGED flag - either of these is enough to
2104        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2105        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2106        start sets both bits.
2107    
2108    4.  Added options (at build time, compile time, exec time) to change \R from
2109        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2110    
2111    5.  doc/pcresyntax.html was missing from the distribution.
2112    
2113    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2114        compatibility, even though it is no longer used.
2115    
2116    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2117        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2118        windows.h file is present (where different names are used). [This was
2119        reversed later after testing - see 16 below.]
2120    
2121    8.  Changed all #include <config.h> to #include "config.h". There were also
2122        some further <pcre.h> cases that I changed to "pcre.h".
2123    
2124    9.  When pcregrep was used with the --colour option, it missed the line ending
2125        sequence off the lines that it output.
2126    
2127    10. It was pointed out to me that arrays of string pointers cause lots of
2128        relocations when a shared library is dynamically loaded. A technique of
2129        using a single long string with a table of offsets can drastically reduce
2130        these. I have refactored PCRE in four places to do this. The result is
2131        dramatic:
2132    
2133          Originally:                          290
2134          After changing UCP table:            187
2135          After changing error message table:   43
2136          After changing table of "verbs"       36
2137          After changing table of Posix names   22
2138    
2139        Thanks to the folks working on Gregex for glib for this insight.
2140    
2141    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2142        unicode-properties was also set.
2143    
2144    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2145    
2146    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2147        checked only for CRLF.
2148    
2149    14. Added casts to pcretest.c to avoid compiler warnings.
2150    
2151    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2152    
2153    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2154        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2155        entirely. This removes changes made in 7 above.
2156    
2157    17. The CMake files have been updated, and there is now more information about
2158        building with CMake in the NON-UNIX-USE document.
2159    
2160    
2161    Version 7.3 28-Aug-07
2162    ---------------------
2163    
2164     1. In the rejigging of the build system that eventually resulted in 7.1, the
2165        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2166        brackets there is not right, since it causes compilers to look for an
2167        installed pcre.h, not the version that is in the source that is being
2168        compiled (which of course may be different). I have changed it back to:
2169    
2170          #include "pcre.h"
2171    
2172        I have a vague recollection that the change was concerned with compiling in
2173        different directories, but in the new build system, that is taken care of
2174        by the VPATH setting the Makefile.
2175    
2176     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2177        when the subject happened to end in the byte 0x85 (e.g. if the last
2178        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2179        characters but of course it shouldn't be taken as a newline when it is part
2180        of another character. The bug was that, for an unlimited repeat of . in
2181        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2182        characters when looking for a newline.
2183    
2184     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2185    
2186     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2187        in debug output.
2188    
2189     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2190        long printing in the pcrecpp unittest when running under MinGW.
2191    
2192     6. ESC_K was left out of the EBCDIC table.
2193    
2194     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2195        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2196        limit also applies to "virtual nesting" when a pattern is recursive, and in
2197        this case 1000 isn't so big. I have been able to remove this limit at the
2198        expense of backing off one optimization in certain circumstances. Normally,
2199        when pcre_exec() would call its internal match() function recursively and
2200        immediately return the result unconditionally, it uses a "tail recursion"
2201        feature to save stack. However, when a subpattern that can match an empty
2202        string has an unlimited repetition quantifier, it no longer makes this
2203        optimization. That gives it a stack frame in which to save the data for
2204        checking that an empty string has been matched. Previously this was taken
2205        from the 1000-entry workspace that had been reserved. So now there is no
2206        explicit limit, but more stack is used.
2207    
2208     8. Applied Daniel's patches to solve problems with the import/export magic
2209        syntax that is required for Windows, and which was going wrong for the
2210        pcreposix and pcrecpp parts of the library. These were overlooked when this
2211        problem was solved for the main library.
2212    
2213     9. There were some crude static tests to avoid integer overflow when computing
2214        the size of patterns that contain repeated groups with explicit upper
2215        limits. As the maximum quantifier is 65535, the maximum group length was
2216        set at 30,000 so that the product of these two numbers did not overflow a
2217        32-bit integer. However, it turns out that people want to use groups that
2218        are longer than 30,000 bytes (though not repeat them that many times).
2219        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2220        made it possible to implement the integer overflow checks in a much more
2221        dynamic way, which I have now done. The artificial limitation on group
2222        length has been removed - we now have only the limit on the total length of
2223        the compiled pattern, which depends on the LINK_SIZE setting.
2224    
2225    10. Fixed a bug in the documentation for get/copy named substring when
2226        duplicate names are permitted. If none of the named substrings are set, the
2227        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2228        empty string.
2229    
2230    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2231        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2232        because the ] is interpreted as the first data character and the
2233        terminating ] is not found. PCRE has been made compatible with Perl in this
2234        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2235        cause memory overwriting.
2236    
2237    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2238        string has been matched (to stop an infinite loop). It was not recognizing
2239        a conditional subpattern that could match an empty string if that
2240        subpattern was within another subpattern. For example, it looped when
2241        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2242        condition was not nested. This bug has been fixed.
2243    
2244    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2245        past the start of the subject in the presence of bytes with the top bit
2246        set, for example "\x8aBCD".
2247    
2248    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2249        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2250    
2251    14. Optimized (?!) to (*FAIL).
2252    
2253    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2254        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2255        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2256        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2257        does: it's just the validity check that is more restrictive.
2258    
2259    16. Inserted checks for integer overflows during escape sequence (backslash)
2260        processing, and also fixed erroneous offset values for syntax errors during
2261        backslash processing.
2262    
2263    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2264        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2265    
2266    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2267        caused an overrun.
2268    
2269    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2270        something other than just ASCII characters) inside a group that had an
2271        unlimited repeat caused a loop at compile time (while checking to see
2272        whether the group could match an empty string).
2273    
2274    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2275        [\P{Any}] did so. (Error in the code for printing property names.)
2276    
2277    21. An orphan \E inside a character class could cause a crash.
2278    
2279    22. A repeated capturing bracket such as (A)? could cause a wild memory
2280        reference during compilation.
2281    
2282    23. There are several functions in pcre_compile() that scan along a compiled
2283        expression for various reasons (e.g. to see if it's fixed length for look
2284        behind). There were bugs in these functions when a repeated \p or \P was
2285        present in the pattern. These operators have additional parameters compared
2286        with \d, etc, and these were not being taken into account when moving along
2287        the compiled data. Specifically:
2288    
2289        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2290            length.
2291    
2292        (b) An item such as \pL+ within a repeated group could cause crashes or
2293            loops.
2294    
2295        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2296            "reference to non-existent subpattern" error.
2297    
2298        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2299    
2300    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2301        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2302    
2303    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2304    
2305    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2306        character were causing crashes (broken optimization).
2307    
2308    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2309        \p or \P) caused a compile-time loop.
2310    
2311    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2312        break. For example, the unstudied pattern [\r\n]A does not match the string
2313        "\r\nA" because change 7.0/46 below moves the current point on by two
2314        characters after failing to match at the start. However, the pattern \nA
2315        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2316        the same is true. There doesn't seem any very clean way out of this, but
2317        what I have chosen to do makes the common cases work: PCRE now takes note
2318        of whether there can be an explicit match for \r or \n anywhere in the
2319        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2320        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2321        pattern has explicit CR or LF references.
2322    
2323    29. Added (*CR) etc for changing newline setting at start of pattern.
2324    
2325    
2326    Version 7.2 19-Jun-07
2327    ---------------------
2328    
2329     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2330        which is apparently normally available under Windows.
2331    
2332     2. Re-jig the pcregrep tests with different newline settings in an attempt
2333        to make them independent of the local environment's newline setting.
2334    
2335     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2336    
2337     4. Some of the "internals" tests were previously cut out when the link size
2338        was not 2, because the output contained actual offsets. The recent new
2339        "Z" feature of pcretest means that these can be cut out, making the tests
2340        usable with all link sizes.
2341    
2342     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2343        stack recursion. This gives a massive performance boost under BSD, but just
2344        a small improvement under Linux. However, it saves one field in the frame
2345        in all cases.
2346    
2347     6. Added more features from the forthcoming Perl 5.10:
2348    
2349        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2350            recursion call. It refers to the nth most recently opened parentheses.
2351    
2352        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2353            to be opened parentheses.
2354    
2355        (c) Conditions that refer to capturing parentheses can be specified
2356            relatively, for example, (?(-2)... or (?(+3)...
2357    
2358        (d) \K resets the start of the current match so that everything before
2359            is not part of it.
2360    
2361        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2362    
2363        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2364            reference syntax.
2365    
2366        (g) (?| introduces a group in which the numbering of parentheses in each
2367            alternative starts with the same number.
2368    
2369        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2370    
2371     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2372        PCRE_INFO_JCHANGED.
2373    
2374     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2375        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2376        for detecting groups that can match an empty string.
2377    
2378     9. A pattern with a very large number of alternatives (more than several
2379        hundred) was running out of internal workspace during the pre-compile
2380        phase, where pcre_compile() figures out how much memory will be needed. A
2381        bit of new cunning has reduced the workspace needed for groups with
2382        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2383        workspace instead of running out of the 4096 that are available.
2384    
2385    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2386    
2387    11. Applied patch from Google to remove an optimization that didn't quite work.
2388        The report of the bug said:
2389    
2390          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2391          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2392          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2393    
2394    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2395        it matched the wrong number of bytes.
2396    
2397    
2398    Version 7.1 24-Apr-07
2399    ---------------------
2400    
2401     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2402        that is more "standard", making use of automake and other Autotools. There
2403        is some re-arrangement of the files and adjustment of comments consequent
2404        on this.
2405    
2406     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2407        for recursive directory scanning broke on some systems because the files
2408        are not scanned in any specific order and on different systems the order
2409        was different. A call to "sort" has been inserted into RunGrepTest for the
2410        approprate test as a short-term fix. In the longer term there may be an
2411        alternative.
2412    
2413     3. I had an email from Eric Raymond about problems translating some of PCRE's
2414        man pages to HTML (despite the fact that I distribute HTML pages, some
2415        people do their own conversions for various reasons). The problems
2416        concerned the use of low-level troff macros .br and .in. I have therefore
2417        removed all such uses from the man pages (some were redundant, some could
2418        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2419        HTML has been updated to handle .nf/.fi and to complain if it encounters
2420        .br or .in.
2421    
2422     4. Updated comments in configure.ac that get placed in config.h.in and also
2423        arranged for config.h to be included in the distribution, with the name
2424        config.h.generic, for the benefit of those who have to compile without
2425        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2426    
2427     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2428        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2429        makevp.bat for the current PCRE, using the additional files
2430        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2431    
2432     6. A Windows user reported a minor discrepancy with test 2, which turned out
2433        to be caused by a trailing space on an input line that had got lost in his
2434        copy. The trailing space was an accident, so I've just removed it.
2435    
2436     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2437        that is needed.
2438    
2439     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2440        as "const" (a) because they are and (b) because it helps the PHP
2441        maintainers who have recently made a script to detect big data structures
2442        in the php code that should be moved to the .rodata section. I remembered
2443        to update Builducptable as well, so it won't revert if ucptable.h is ever
2444        re-created.
2445    
2446     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2447        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2448        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2449        support is not required. This saves 1.5-2K of code, which is important in
2450        some applications.
2451    
2452        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2453        so as not to refer to the tables, even though these functions will never be
2454        called when UTF-8 support is disabled. Otherwise there are problems with a
2455        shared library.
2456    
2457    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2458    
2459        (a) It was defining its arguments as char * instead of void *.
2460    
2461        (b) It was assuming that all moves were upwards in memory; this was true
2462            a long time ago when I wrote it, but is no longer the case.
2463    
2464        The emulated memove() is provided for those environments that have neither
2465        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2466        is clearly not the case, as these two bugs were recently reported.
2467    
2468    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2469        and Detrail to create the HTML documentation, the .txt form of the man
2470        pages, and it removes trailing spaces from listed files. It also creates
2471        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2472        case, it wraps all the #defines with #ifndefs. This script should be run
2473        before "make dist".
2474    
2475    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2476        with Unicode property support.
2477    
2478        (a) For a maximizing quantifier, if the two different cases of the
2479            character were of different lengths in their UTF-8 codings (there are
2480            some cases like this - I found 11), and the matching function had to
2481            back up over a mixture of the two cases, it incorrectly assumed they
2482            were both the same length.
2483    
2484        (b) When PCRE was configured to use the heap rather than the stack for
2485            recursion during matching, it was not correctly preserving the data for
2486            the other case of a UTF-8 character when checking ahead for a match
2487            while processing a minimizing repeat. If the check also involved
2488            matching a wide character, but failed, corruption could cause an
2489            erroneous result when trying to check for a repeat of the original
2490            character.
2491    
2492    13. Some tidying changes to the testing mechanism:
2493    
2494        (a) The RunTest script now detects the internal link size and whether there
2495            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2496            values substituted by "configure". (The RunGrepTest script already did
2497            this for UTF-8.) The configure.ac script no longer substitutes the
2498            relevant variables.
2499    
2500        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2501            with length and offset values. This means that the output is different
2502            for different internal link sizes. Test 2 is skipped for link sizes
2503            other than 2 because of this, bypassing the problem. Unfortunately,
2504            there was also a test in test 3 (the locale tests) that used /B and
2505            failed for link sizes other than 2. Rather than cut the whole test out,
2506            I have added a new /Z option to pcretest that replaces the length and
2507            offset values with spaces. This is now used to make test 3 independent
2508            of link size. (Test 2 will be tidied up later.)
2509    
2510    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2511        segmentation fault instead of returning the appropriate error message.
2512    
2513    15. In multiline mode when the newline sequence was set to "any", the pattern
2514        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2515        This doesn't seem right; it now treats the CRLF combination as the line
2516        ending, and so does not match in that case. It's only a pattern such as ^$
2517        that would hit this one: something like ^ABC$ would have failed after \r
2518        and then tried again after \r\n.
2519    
2520    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2521        in an attempt to make files that differ only in their line terminators
2522        compare equal. This works on Linux.
2523    
2524    17. Under certain error circumstances pcregrep might try to free random memory
2525        as it exited. This is now fixed, thanks to valgrind.
2526    
2527    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2528        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2529        was because its rules for how to advance for /g after matching an empty
2530        string at the end of a line did not allow for this case. They now check for
2531        it specially.
2532    
2533    20. pcretest is supposed to handle patterns and data of any length, by
2534        extending its buffers when necessary. It was getting this wrong when the
2535        buffer for a data line had to be extended.
2536    
2537    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2538        CRLF as a newline sequence.
2539    
2540    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2541        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2542        I have nevertheless tidied it up.
2543    
2544    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2545    
2546    24. Added a man page for pcre-config.
2547    
2548    
2549    Version 7.0 19-Dec-06
2550    ---------------------
2551    
2552     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2553        moving to gcc 4.1.1.
2554    
2555     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2556        sys/time.h, which is documented as needed for this function. It doesn't
2557        seem to matter on Linux, but it showed up on some releases of OS X.
2558    
2559     3. It seems that there are systems where bytes whose values are greater than
2560        127 match isprint() in the "C" locale. The "C" locale should be the
2561        default when a C program starts up. In most systems, only ASCII printing
2562        characters match isprint(). This difference caused the output from pcretest
2563        to vary, making some of the tests fail. I have changed pcretest so that:
2564    
2565        (a) When it is outputting text in the compiled version of a pattern, bytes
2566            other than 32-126 are always shown as hex escapes.
2567    
2568        (b) When it is outputting text that is a matched part of a subject string,
2569            it does the same, unless a different locale has been set for the match
2570            (using the /L modifier). In this case, it uses isprint() to decide.
2571    
2572     4. Fixed a major bug that caused incorrect computation of the amount of memory
2573        required for a compiled pattern when options that changed within the
2574        pattern affected the logic of the preliminary scan that determines the
2575        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2576        that the computed length was too small. The symptoms of this bug were
2577        either the PCRE error "internal error: code overflow" from pcre_compile(),
2578        or a glibc crash with a message such as "pcretest: free(): invalid next
2579        size (fast)". Examples of patterns that provoked this bug (shown in
2580        pcretest format) are:
2581    
2582          /(?-x: )/x
2583          /(?x)(?-x: \s*#\s*)/
2584          /((?i)[\x{c0}])/8
2585          /(?i:[\x{c0}])/8
2586    
2587        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2588        is now done differently.
2589    
2590     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2591        wrapper classes; (b) implement a new function in the C++ scanner that is
2592        more efficient than the old way of doing things because it avoids levels of
2593        recursion in the regex matching; (c) add a paragraph to the documentation
2594        for the FullMatch() function.
2595    
2596     6. The escape sequence \n was being treated as whatever was defined as
2597        "newline". Not only was this contrary to the documentation, which states
2598        that \n is character 10 (hex 0A), but it also went horribly wrong when
2599        "newline" was defined as CRLF. This has been fixed.
2600    
2601     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2602        was being set to -1 for the "end of line" case (supposedly a value that no
2603        character can have). Though this value is never used (the check for end of
2604        line is "zero bytes in current character"), it caused compiler complaints.
2605        I've changed it to 0xffffffff.
2606    
2607     8. In pcre_version.c, the version string was being built by a sequence of
2608        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2609        string (as it is for production releases) called a macro with an empty
2610        argument. The C standard says the result of this is undefined. The gcc
2611        compiler treats it as an empty string (which was what was wanted) but it is
2612        reported that Visual C gives an error. The source has been hacked around to
2613        avoid this problem.
2614    
2615     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2616        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2617        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2618        of them did).
2619    
2620    10. Originally, pcretest opened its input and output without "b"; then I was
2621        told that "b" was needed in some environments, so it was added for release
2622        5.0 to both the input and output. (It makes no difference on Unix-like
2623        systems.) Later I was told that it is wrong for the input on Windows. I've
2624        now abstracted the modes into two macros, to make it easier to fiddle with
2625        them, and removed "b" from the input mode under Windows.
2626    
2627    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2628    
2629    12. Added -help and --help to pcretest as an official way of being reminded
2630        of the options.
2631    
2632    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2633        and pcrecpp.cc because they annoy compilers at high warning levels.
2634    
2635    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2636    
2637    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2638        scripts are not C programs :-) and which was not noticed because it works
2639        on Linux.
2640    
2641    16. pcretest is supposed to handle any length of pattern and data line (as one
2642        line or as a continued sequence of lines) by extending its input buffer if
2643        necessary. This feature was broken for very long pattern lines, leading to
2644        a string of junk being passed to pcre_compile() if the pattern was longer
2645        than about 50K.
2646    
2647    17. I have done a major re-factoring of the way pcre_compile() computes the
2648        amount of memory needed for a compiled pattern. Previously, there was code
2649        that made a preliminary scan of the pattern in order to do this. That was
2650        OK when PCRE was new, but as the facilities have expanded, it has become
2651        harder and harder to keep it in step with the real compile phase, and there
2652        have been a number of bugs (see for example, 4 above). I have now found a
2653        cunning way of running the real compile function in a "fake" mode that
2654        enables it to compute how much memory it would need, while actually only
2655        ever using a few hundred bytes of working memory and without too many
2656        tests of the mode. This should make future maintenance and development
2657        easier. A side effect of this work is that the limit of 200 on the nesting
2658        depth of parentheses has been removed (though this was never a serious
2659        limitation, I suspect). However, there is a downside: pcre_compile() now
2660        runs more slowly than before (30% or more, depending on the pattern). I
2661        hope this isn't a big issue. There is no effect on runtime performance.
2662    
2663    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2664        newline (only possible for the last line of a file) and it was a
2665        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2666    
2667    19. Added additional timing features to pcretest. (1) The -tm option now times
2668        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2669        separate command line item, by a number that specifies the number of
2670        repeats to use when timing. The default is 50000; this gives better
2671        precision, but takes uncomfortably long for very large patterns.
2672    
2673    20. Extended pcre_study() to be more clever in cases where a branch of a
2674        subpattern has no definite first character. For example, (a*|b*)[cd] would
2675        previously give no result from pcre_study(). Now it recognizes that the
2676        first character must be a, b, c, or d.
2677    
2678    21. There was an incorrect error "recursive call could loop indefinitely" if
2679        a subpattern (or the entire pattern) that was being tested for matching an
2680        empty string contained only one non-empty item after a nested subpattern.
2681        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2682        incorrectly, because the \d was being skipped in the check.
2683    
2684    22. The pcretest program now has a new pattern option /B and a command line
2685        option -b, which is equivalent to adding /B to every pattern. This causes
2686        it to show the compiled bytecode, without the additional information that
2687        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2688        is the same as /B/I).
2689    
2690    23. A new optimization is now able automatically to treat some sequences such
2691        as a*b as a*+b. More specifically, if something simple (such as a character
2692        or a simple class like \d) has an unlimited quantifier, and is followed by
2693        something that cannot possibly match the quantified thing, the quantifier
2694        is automatically "possessified".
2695    
2696    24. A recursive reference to a subpattern whose number was greater than 39
2697        went wrong under certain circumstances in UTF-8 mode. This bug could also
2698        have affected the operation of pcre_study().
2699    
2700    25. Realized that a little bit of performance could be had by replacing
2701        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2702    
2703    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2704    
2705    27. Possessive quantifiers such as a++ were previously implemented by turning
2706        them into atomic groups such as ($>a+). Now they have their own opcodes,
2707        which improves performance. This includes the automatically created ones
2708        from 23 above.
2709    
2710    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2711        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2712        the first matched character to be a colon. This applied both to named and
2713        numbered groups.
2714    
2715    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2716    
2717    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2718        building PCRE on Windows easier, so I have included it in the distribution.
2719    
2720    31. There is now a check in pcretest against a ridiculously large number being
2721        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2722        loop, the loop is abandoned.
2723    
2724    32. Forward references to subpatterns in conditions such as (?(2)...) where
2725        subpattern 2 is defined later cause pcre_compile() to search forwards in
2726        the pattern for the relevant set of parentheses. This search went wrong
2727        when there were unescaped parentheses in a character class, parentheses
2728        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2729    
2730    33. "Subroutine" calls and backreferences were previously restricted to
2731        referencing subpatterns earlier in the regex. This restriction has now
2732        been removed.
2733    
2734    34. Added a number of extra features that are going to be in Perl 5.10. On the
2735        whole, these are just syntactic alternatives for features that PCRE had
2736        previously implemented using the Python syntax or my own invention. The
2737        other formats are all retained for compatibility.
2738    
2739        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2740            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2741            also .NET compatible.
2742    
2743        (b) A recursion or subroutine call to a named group can now be defined as
2744            (?&name) as well as (?P>name).
2745    
2746        (c) A backreference to a named group can now be defined as \k<name> or
2747            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2748            5.10, are also .NET compatible.
2749    
2750        (d) A conditional reference to a named group can now use the syntax
2751            (?(<name>) or (?('name') as well as (?(name).
2752    
2753        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2754            groups (named and numbered) that are never evaluated inline, but can be
2755            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2756            is always false. There may be only one alternative in such a group.
2757    
2758        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2759            as the simple (?(R). The condition is true only if the most recent
2760            recursion is that of the given number or name. It does not search out
2761            through the entire recursion stack.
2762    
2763        (g) The escape \gN or \g{N} has been added, where N is a positive or
2764            negative number, specifying an absolute or relative reference.
2765    
2766    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2767        some "unreachable code" warnings.
2768    
2769    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2770        things, this adds five new scripts.
2771    
2772    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2773        There were also incompatibilities regarding the handling of \Q..\E inside
2774        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2775        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2776    
2777    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2778        matches an empty string, and forcibly breaks the loop. There were bugs in
2779        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2780        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2781        separate and independent bugs (that affected different cases) have been
2782        fixed.
2783    
2784    39. Refactored the code to abolish the use of different opcodes for small
2785        capturing bracket numbers. This is a tidy that I avoided doing when I
2786        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2787        The new approach is not only tidier, it makes it possible to reduce the
2788        memory needed to fix the previous bug (38).
2789    
2790    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2791        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2792        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2793        mode.
2794    
2795    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2796        report.
2797    
2798    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2799        copy construction and assignment in the C++ wrapper.
2800    
2801    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2802        couple of bugs that could have given wrong results in the "--newline=crlf"
2803        case.
2804    
2805    44. Added a number of casts and did some reorganization of signed/unsigned int
2806        variables following suggestions from Dair Grant. Also renamed the variable
2807        "this" as "item" because it is a C++ keyword.
2808    
2809    45. Arranged for dftables to add
2810    
2811          #include "pcre_internal.h"
2812    
2813        to pcre_chartables.c because without it, gcc 4.x may remove the array
2814        definition from the final binary if PCRE is built into a static library and
2815        dead code stripping is activated.
2816    
2817    46. For an unanchored pattern, if a match attempt fails at the start of a
2818        newline sequence, and the newline setting is CRLF or ANY, and the next two
2819        characters are CRLF, advance by two characters instead of one.
2820    
2821    
2822    Version 6.7 04-Jul-06
2823    ---------------------
2824    
2825     1. In order to handle tests when input lines are enormously long, pcretest has
2826        been re-factored so that it automatically extends its buffers when
2827        necessary. The code is crude, but this _is_ just a test program. The
2828        default size has been increased from 32K to 50K.
2829    
2830     2. The code in pcre_study() was using the value of the re argument before
2831        testing it for NULL. (Of course, in any sensible call of the function, it
2832        won't be NULL.)
2833    
2834     3. The memmove() emulation function in pcre_internal.h, which is used on
2835        systems that lack both memmove() and bcopy() - that is, hardly ever -
2836        was missing a "static" storage class specifier.
2837    
2838     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2839        containing an extended class (one that cannot be represented by a bitmap
2840        because it contains high-valued characters or Unicode property items, e.g.
2841        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2842        pattern, but PCRE should not loop if you do not (it no longer does).
2843        [Detail: two cases were found: (a) a repeated subpattern containing an
2844        extended class; (b) a recursive reference to a subpattern that followed a
2845        previous extended class. It wasn't skipping over the extended class
2846        correctly when UTF-8 mode was not set.]
2847    
2848     5. A negated single-character class was not being recognized as fixed-length
2849        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2850        compile error "lookbehind assertion is not fixed length".
2851    
2852     6. The RunPerlTest auxiliary script was showing an unexpected difference
2853        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2854        write a Perl script that can interpret lines of an input file either as
2855        byte characters or as UTF-8, which is what "perltest" was being required to
2856        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2857        can't do is switch easily at run time between having the "use utf8;" pragma
2858        or not. In the end, I fudged it by using the RunPerlTest script to insert
2859        "use utf8;" explicitly for the UTF-8 tests.
2860    
2861     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2862        the end of the subject string, contrary to the documentation and to what
2863        Perl does. This was true of both matching functions. Now it matches only at
2864        the start of the subject and immediately after *internal* newlines.
2865    
2866     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2867        a pointer to an int instead of a pointer to an unsigned long int. This
2868        caused problems on 64-bit systems.
2869    
2870     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2871        instance of the 'standard' template library not being so standard".
2872    
2873    10. There was no check on the number of named subpatterns nor the maximum
2874        length of a subpattern name. The product of these values is used to compute
2875        the size of the memory block for a compiled pattern. By supplying a very
2876        long subpattern name and a large number of named subpatterns, the size
2877        computation could be caused to overflow. This is now prevented by limiting
2878        the length of names to 32 characters, and the number of named subpatterns
2879        to 10,000.
2880    
2881    11. Subpatterns that are repeated with specific counts have to be replicated in
2882        the compiled pattern. The size of memory for this was computed from the
2883        length of the subpattern and the repeat count. The latter is limited to
2884        65535, but there was no limit on the former, meaning that integer overflow
2885        could in principle occur. The compiled length of a repeated subpattern is
2886        now limited to 30,000 bytes in order to prevent this.
2887    
2888    12. Added the optional facility to have named substrings with the same name.
2889    
2890    13. Added the ability to use a named substring as a condition, using the
2891        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2892        are numbers (not recommended). Forward references are permitted.
2893    
2894    14. Added forward references in named backreferences (if you see what I mean).
2895    
2896    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2897        pattern could run off the end of the subject. For example, the pattern
2898        "(?s)(.{1,5})"8 did this with the subject "ab".
2899    
2900    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2901        PCRE_CASELESS was set when matching characters that were quantified with ?
2902        or *.
2903    
2904    17. A character class other than a single negated character that had a minimum
2905        but no maximum quantifier - for example [ab]{6,} - was not handled
2906        correctly by pce_dfa_exec(). It would match only one character.
2907    
2908    18. A valid (though odd) pattern that looked like a POSIX character
2909        class but used an invalid character after [ (for example [[,abc,]]) caused
2910        pcre_compile() to give the error "Failed: internal error: code overflow" or
2911        in some cases to crash with a glibc free() error. This could even happen if
2912        the pattern terminated after [[ but there just happened to be a sequence of
2913        letters, a binary zero, and a closing ] in the memory that followed.
2914    
2915    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2916        over the years. Originally (before any Unicode support), just the bottom 8
2917        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2918        output from "man perlunicode" includes this:
2919    
2920          The regular expression compiler produces polymorphic opcodes.  That
2921          is, the pattern adapts to the data and automatically switches to
2922          the Unicode character scheme when presented with Unicode data--or
2923          instead uses a traditional byte scheme when presented with byte
2924          data.
2925    
2926        Sadly, a wide octal escape does not cause a switch, and in a string with
2927        no other multibyte characters, these octal escapes are treated as before.
2928        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2929        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2930        Unicode string.
2931    
2932        I have not perpetrated such confusion in PCRE. Up till now, it took just
2933        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2934        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2935        translate to the appropriate multibyte character.
2936    
2937    29. Applied some refactoring to reduce the number of warnings from Microsoft
2938        and Borland compilers. This has included removing the fudge introduced
2939        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2940        a warning about an unused variable.
2941    
2942    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2943        characters since release 4.0, because Perl (from release 5.004) does not.
2944        [Or at least, is documented not to: some releases seem to be in conflict
2945        with the documentation.] However, when a pattern was studied with
2946        pcre_study() and all its branches started with \s, PCRE still included VT
2947        as a possible starting character. Of course, this did no harm; it just
2948        caused an unnecessary match attempt.
2949    
2950    22. Removed a now-redundant internal flag bit that recorded the fact that case
2951        dependency changed within the pattern. This was once needed for "required
2952        byte" processing, but is no longer used. This recovers a now-scarce options
2953        bit. Also moved the least significant internal flag bit to the most-
2954        significant bit of the word, which was not previously used (hangover from
2955        the days when it was an int rather than a uint) to free up another bit for
2956        the future.
2957    
2958    23. Added support for CRLF line endings as well as CR and LF. As well as the
2959        default being selectable at build time, it can now be changed at runtime
2960        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2961        specify that it is scanning data with non-default line endings.
2962    
2963    24. Changed the definition of CXXLINK to make it agree with the definition of
2964        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2965    
2966    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2967        recursions. This makes a big different to stack usage for some patterns.
2968    
2969    26. If a subpattern containing a named recursion or subroutine reference such
2970        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2971        the space required for the compiled pattern went wrong and gave too small a
2972        value. Depending on the environment, this could lead to "Failed: internal
2973        error: code overflow at offset 49" or "glibc detected double free or
2974        corruption" errors.
2975    
2976    27. Applied patches from Google (a) to support the new newline modes and (b) to
2977        advance over multibyte UTF-8 characters in GlobalReplace.
2978    
2979    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2980        difference for some implementation of PCRE in some Windows version.
2981    
2982    29. Added some extra testing facilities to pcretest:
2983    
2984        \q<number>   in a data line sets the "match limit" value
2985        \Q<number>   in a data line sets the "match recursion limt" value
2986        -S <number>  sets the stack size, where <number> is in megabytes
2987    
2988        The -S option isn't available for Windows.
2989    
2990    
2991    Version 6.6 06-Feb-06
2992    ---------------------
2993    
2994     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2995        in pcreposix.h. I have copied the definition from pcre.h.
2996    
2997     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2998        because pcre.h is no longer a built file.
2999    
3000     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3001        not normally included in the compiled code.
3002    
3003    
3004    Version 6.5 01-Feb-06
3005    ---------------------
3006    
3007     1. When using the partial match feature with pcre_dfa_exec(), it was not
3008        anchoring the second and subsequent partial matches at the new starting
3009        point. This could lead to incorrect results. For example, with the pattern
3010        /1234/, partially matching against "123" and then "a4" gave a match.
3011    
3012     2. Changes to pcregrep:
3013    
3014        (a) All non-match returns from pcre_exec() were being treated as failures
3015            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3016            error message is output. Some extra information is given for the
3017            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3018            probably the only errors that are likely to be caused by users (by
3019            specifying a regex that has nested indefinite repeats, for instance).
3020            If there are more than 20 of these errors, pcregrep is abandoned.
3021    
3022        (b) A binary zero was treated as data while matching, but terminated the
3023            output line if it was written out. This has been fixed: binary zeroes
3024            are now no different to any other data bytes.
3025    
3026        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3027            used to set a locale for matching. The --locale=xxxx long option has
3028            been added (no short equivalent) to specify a locale explicitly on the
3029            pcregrep command, overriding the environment variables.
3030    
3031        (d) When -B was used with -n, some line numbers in the output were one less
3032            than they should have been.
3033    
3034        (e) Added the -o (--only-matching) option.
3035    
3036        (f) If -A or -C was used with -c (count only), some lines of context were
3037            accidentally printed for the final match.
3038    
3039        (g) Added the -H (--with-filename) option.
3040    
3041        (h) The combination of options -rh failed to suppress file names for files
3042            that were found from directory arguments.
3043    
3044        (i) Added the -D (--devices) and -d (--directories) options.
3045    
3046        (j) Added the -F (--fixed-strings) option.
3047    
3048        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3049    
3050        (l) Added the --colo(u)r option.
3051    
3052        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3053            is not present by default.
3054    
3055     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3056        items such as (?R) or (?1), when the recursion could match a number of
3057        alternatives. If it matched one of the alternatives, but subsequently,
3058        outside the recursion, there was a failure, the code tried to back up into
3059        the recursion. However, because of the way PCRE is implemented, this is not
3060        possible, and the result was an incorrect result from the match.
3061    
3062        In order to prevent this happening, the specification of recursion has
3063        been changed so that all such subpatterns are automatically treated as
3064        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3065    
3066     4. I had overlooked the fact that, in some locales, there are characters for
3067        which isalpha() is true but neither isupper() nor islower() are true. In
3068        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3069        and ordfeminine) are like this. This affected the treatment of \w and \W
3070        when they appeared in character classes, but not when they appeared outside
3071        a character class. The bit map for "word" characters is now created
3072        separately from the results of isalnum() instead of just taking it from the
3073        upper, lower, and digit maps. (Plus the underscore character, of course.)
3074    
3075     5. The above bug also affected the handling of POSIX character classes such as
3076        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3077        permanent tables. Instead, the bit maps for such a class were previously
3078        created as the appropriate unions of the upper, lower, and digit bitmaps.
3079        Now they are created by subtraction from the [[:word:]] class, which has
3080        its own bitmap.
3081    
3082     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3083        It is created by subtracting the vertical space characters (\x09, \x0a,
3084        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3085        subtraction was done in the overall bitmap for a character class, meaning
3086        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3087        be recognized. This bug has been fixed.
3088    
3089     7. Patches from the folks at Google:
3090    
3091          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3092          real life, but is still worth protecting against".
3093    
3094          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3095          regular expressions".
3096    
3097          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3098          have it.
3099    
3100          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3101          "configure" and the latter not, in order to fix a problem somebody had
3102          with compiling the Arg class on HP-UX.
3103    
3104          (e) Improve the error-handling of the C++ wrapper a little bit.
3105    
3106          (f) New tests for checking recursion limiting.
3107    
3108     8. The pcre_memmove() function, which is used only if the environment does not
3109        have a standard memmove() function (and is therefore rarely compiled),
3110        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3111        returning a result (though PCRE never actually uses the result).
3112    
3113     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3114        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3115        returned instead of calling malloc() with an overflowing number that would
3116        most likely cause subsequent chaos.
3117    
3118    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3119    
3120    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3121        with this option is matched, the nmatch and pmatch options of regexec() are
3122        ignored.
3123    
3124    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3125        provided in case anyone wants to the the POSIX interface with UTF-8
3126        strings.
3127    
3128    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3129        C++ linking (needed for some HP-UX environments).
3130    
3131    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3132        (unused parameter) and in the pcre_printint() function (omitted "default"
3133        switch label when the default is to do nothing).
3134    
3135    15. Added some code to make it possible, when PCRE is compiled as a C++
3136        library, to replace subject pointers for pcre_exec() with a smart pointer
3137        class, thus making it possible to process discontinuous strings.
3138    
3139    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3140        much the same function. They were added by different people who were trying
3141        to make PCRE easy to compile on non-Unix systems. It has been suggested
3142        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3143        for compiling on Windows systems. I have therefore replaced it with
3144        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3145        defaults to "extern" for C or "extern C" for C++, which works fine on
3146        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3147        SCOPE with something explicit in config.h. In addition:
3148    
3149        (a) pcreposix.h still had just "extern" instead of either of these macros;
3150            I have replaced it with PCRE_DATA_SCOPE.
3151    
3152        (b) Functions such as _pcre_xclass(), which are internal to the library,
3153            but external in the C sense, all had PCRE_EXPORT in their definitions.
3154            This is apparently wrong for the Windows case, so I have removed it.
3155            (It makes no difference on Unix-like systems.)
3156    
3157    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3158        of recursive calls to match(). This is different to MATCH_LIMIT because
3159        that limits the total number of calls to match(), not all of which increase
3160        the depth of recursion. Limiting the recursion depth limits the amount of
3161        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3162        when PCRE is compiled, and changed at run time. A patch from Google adds
3163        this functionality to the C++ interface.
3164    
3165    18. Changes to the handling of Unicode character properties:
3166    
3167        (a) Updated the table to Unicode 4.1.0.
3168    
3169        (b) Recognize characters that are not in the table as "Cn" (undefined).
3170    
3171        (c) I revised the way the table is implemented to a much improved format
3172            which includes recognition of ranges. It now supports the ranges that
3173            are defined in UnicodeData.txt, and it also amalgamates other
3174            characters into ranges. This has reduced the number of entries in the
3175            table from around 16,000 to around 3,000, thus reducing its size
3176            considerably. I realized I did not need to use a tree structure after
3177            all - a binary chop search is just as efficient. Having reduced the
3178            number of entries, I extended their size from 6 bytes to 8 bytes to
3179            allow for more data.
3180    
3181        (d) Added support for Unicode script names via properties such as \p{Han}.
3182    
3183    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3184        matching that character.
3185    
3186    20. When matching a repeated Unicode property with a minimum greater than zero,
3187        (for example \pL{2,}), PCRE could look past the end of the subject if it
3188        reached it while seeking the minimum number of characters. This could
3189        happen only if some of the characters were more than one byte long, because
3190        there is a check for at least the minimum number of bytes.
3191    
3192    21. Refactored the implementation of \p and \P so as to be more general, to
3193        allow for more different types of property in future. This has changed the
3194        compiled form incompatibly. Anybody with saved compiled patterns that use
3195        \p or \P will have to recompile them.
3196    
3197    22. Added "Any" and "L&" to the supported property types.
3198    
3199    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3200        but give a compile time error if the value is greater than 0xff.
3201    
3202    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3203        accidentally not being installed or uninstalled.
3204    
3205    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3206        made were to insert the current release number. This seemed silly, because
3207        it made things harder for people building PCRE on systems that don't run
3208        "configure". I have turned pcre.h into a distributed file, no longer built
3209        by "configure", with the version identification directly included. There is
3210        no longer a pcre.h.in file.
3211    
3212        However, this change necessitated a change to the pcre-config script as
3213        well. It is built from pcre-config.in, and one of the substitutions was the
3214        release number. I have updated configure.ac so that ./configure now finds
3215        the release number by grepping pcre.h.
3216    
3217    26. Added the ability to run the tests under valgrind.
3218    
3219    
3220    Version 6.4 05-Sep-05
3221    ---------------------
3222    
3223     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3224        "--" to be printed when multiple files were scanned, even when none of the
3225        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3226        consider it to be a bug, and have restored the previous behaviour.
3227    
3228     2. A couple of code tidies to get rid of compiler warnings.
3229    
3230     3. The pcretest program used to cheat by referring to symbols in the library
3231        whose names begin with _pcre_. These are internal symbols that are not
3232        really supposed to be visible externally, and in some environments it is
3233        possible to suppress them. The cheating is now confined to including
3234        certain files from the library's source, which is a bit cleaner.
3235    
3236     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3237        file's purpose clearer.
3238    
3239     5. Reorganized pcre_ucp_findchar().
3240    
3241    
3242    Version 6.3 15-Aug-05
3243    ---------------------
3244    
3245     1. The file libpcre.pc.in did not have general read permission in the tarball.
3246    
3247     2. There were some problems when building without C++ support:
3248    
3249        (a) If C++ support was not built, "make install" and "make test" still
3250            tried to test it.
3251    
3252        (b) There were problems when the value of CXX was explicitly set. Some
3253            changes have been made to try to fix these, and ...
3254    
3255        (c) --disable-cpp can now be used to explicitly disable C++ support.
3256    
3257        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3258            backslash in a target when C++ was disabled. This confuses some
3259            versions of "make", apparently. Using an intermediate variable solves
3260            this. (Same for CPP_LOBJ.)
3261    
3262     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3263        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3264        necessary on certain architectures.
3265    
3266     4. Added a setting of -export-symbols-regex to the link command to remove
3267        those symbols that are exported in the C sense, but actually are local
3268        within the library, and not documented. Their names all begin with
3269        "_pcre_". This is not a perfect job, because (a) we have to except some
3270        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3271        available (and never for static libraries). I have made a note to try to
3272        find a way round (a) in the future.
3273    
3274    
3275    Version 6.2 01-Aug-05
3276    ---------------------
3277    
3278     1. There was no test for integer overflow of quantifier values. A construction
3279        such as {1111111111111111} would give undefined results. What is worse, if
3280        a minimum quantifier for a parenthesized subpattern overflowed and became
3281        negative, the calculation of the memory size went wrong. This could have
3282        led to memory overwriting.
3283    
3284     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3285    
3286     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3287        operating environments where this matters.
3288    
3289     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3290        PCRE options from within the C++ wrapper.
3291    
3292     5. Named capturing subpatterns were not being correctly counted when a pattern
3293        was compiled. This caused two problems: (a) If there were more than 100
3294        such subpatterns, the calculation of the memory needed for the whole
3295        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3296        back references of the form \12, where the number was greater than 9, were
3297        not recognized as back references, even though there were sufficient
3298        previous subpatterns.
3299    
3300     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3301        versions of gcc, e.g. 2.95.4.
3302    
3303    
3304    Version 6.1 21-Jun-05
3305    ---------------------
3306    
3307     1. There was one reference to the variable "posix" in pcretest.c that was not
3308        surrounded by "#if !defined NOPOSIX".
3309    
3310     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3311        the cross-check on the old pcre_info() function, for the benefit of the
3312        cut-down version of PCRE that is currently imported into Exim.
3313    
3314     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3315        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3316        patterns that start (?i) but I don't think that matters. The use of (?i) is
3317        just an example; this all applies to the other options as well.
3318    
3319     4. Since libtool seems to echo the compile commands it is issuing, the output
3320        from "make" can be reduced a bit by putting "@" in front of each libtool
3321        compile command.
3322    
3323     5. Patch from the folks at Google for configure.in to be a bit more thorough
3324        in checking for a suitable C++ installation before trying to compile the
3325        C++ stuff. This should fix a reported problem when a compiler was present,
3326        but no suitable headers.
3327    
3328     6. The man pages all had just "PCRE" as their title. I have changed them to
3329        be the relevant file name. I have also arranged that these names are
3330        retained in the file doc/pcre.txt, which is a concatenation in text format
3331        of all the man pages except the little individual ones for each function.
3332    
3333     7. The NON-UNIX-USE file had not been updated for the different set of source
3334        files that come with release 6. I also added a few comments about the C++
3335        wrapper.
3336    
3337    
3338    Version 6.0 07-Jun-05
3339    ---------------------
3340    
3341     1. Some minor internal re-organization to help with my DFA experiments.
3342    
3343     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3344        didn't matter for the library itself when fully configured, but did matter
3345        when compiling without UCP support, or within Exim, where the ucp files are
3346        not imported.
3347    
3348     3. Refactoring of the library code to split up the various functions into
3349        different source modules. The addition of the new DFA matching code (see
3350        below) to a single monolithic source would have made it really too
3351        unwieldy, quite apart from causing all the code to be include in a
3352        statically linked application, when only some functions are used. This is
3353        relevant even without the DFA addition now that patterns can be compiled in
3354        one application and matched in another.
3355    
3356        The downside of splitting up is that there have to be some external
3357        functions and data tables that are used internally in different modules of
3358        the library but which are not part of the API. These have all had their
3359        names changed to start with "_pcre_" so that they are unlikely to clash
3360        with other external names.
3361    
3362     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3363        a different (DFA) algorithm. Although it is slower than the original
3364        function, it does have some advantages for certain types of matching
3365        problem.
3366    
3367     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3368        including restarting after a partial match.
3369    
3370     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3371        defined when compiling for Windows was sent to me. I have put it into the
3372        code, though I have no means of testing or verifying it.
3373    
3374     7. Added the pcre_refcount() auxiliary function.
3375    
3376     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3377        match before or at the first newline in the subject string. In pcretest,
3378        the /f option on a pattern can be used to set this.
3379    
3380     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3381        would behave wrongly. This has been present in PCRE since release 4.0.
3382    
3383    10. A number of changes to the pcregrep command:
3384    
3385        (a) Refactored how -x works; insert ^(...)$ instead of setting
3386            PCRE_ANCHORED and checking the length, in preparation for adding
3387            something similar for -w.
3388    
3389        (b) Added the -w (match as a word) option.
3390    
3391        (c) Refactored the way lines are read and buffered so as to have more
3392            than one at a time available.
3393    
3394        (d) Implemented a pcregrep test script.
3395    
3396        (e) Added the -M (multiline match) option. This allows patterns to match
3397            over several lines of the subject. The buffering ensures that at least
3398            8K, or the rest of the document (whichever is the shorter) is available
3399            for matching (and similarly the previous 8K for lookbehind assertions).
3400    
3401        (f) Changed the --help output so that it now says
3402    
3403              -w, --word-regex(p)
3404    
3405            instead of two lines, one with "regex" and the other with "regexp"
3406            because that confused at least one person since the short forms are the
3407            same. (This required a bit of code, as the output is generated
3408            automatically from a table. It wasn't just a text change.)
3409    
3410        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3411            option but starts with a hyphen. Could be a pattern or a path name
3412            starting with a hyphen, for instance.
3413    
3414        (h) "-" can be given as a file name to represent stdin.
3415    
3416        (i) When file names are being printed, "(standard input)" is used for
3417            the standard input, for compatibility with GNU grep. Previously
3418            "<stdin>" was used.
3419    
3420        (j) The option --label=xxx can be used to supply a name to be used for
3421            stdin when file names are being printed. There is no short form.
3422    
3423        (k) Re-factored the options decoding logic because we are going to add
3424            two more options that take data. Such options can now be given in four
3425            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3426    
3427        (l) Added the -A, -B, and -C options for requesting that lines of context
3428            around matches be printed.
3429    
3430        (m) Added the -L option to print the names of files that do not contain
3431            any matching lines, that is, the complement of -l.
3432    
3433        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3434            continue to scan other files.
3435    
3436        (o) The -s option was incorrectly implemented. For compatibility with other
3437            greps, it now suppresses the error message for a non-existent or non-
3438            accessible file (but not the return code). There is a new option called
3439            -q that suppresses the output of matching lines, which was what -s was
3440            previously doing.
3441    
3442        (p) Added --include and --exclude options to specify files for inclusion
3443            and exclusion when recursing.
3444    
3445    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3446        Hopefully, it now does.
3447    
3448    12. Missing cast in pcre_study().
3449    
3450    13. Added an "uninstall" target to the makefile.
3451    
3452    14. Replaced "extern" in the function prototypes in Makefile.in with
3453        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3454        world, but is set differently for Windows.
3455    
3456    15. Added a second compiling function called pcre_compile2(). The only
3457        difference is that it has an extra argument, which is a pointer to an
3458        integer error code. When there is a compile-time failure, this is set
3459        non-zero, in addition to the error test pointer being set to point to an
3460        error message. The new argument may be NULL if no error number is required
3461        (but then you may as well call pcre_compile(), which is now just a
3462        wrapper). This facility is provided because some applications need a
3463        numeric error indication, but it has also enabled me to tidy up the way
3464        compile-time errors are handled in the POSIX wrapper.
3465    
3466    16. Added VPATH=.libs to the makefile; this should help when building with one
3467        prefix path and installing with another. (Or so I'm told by someone who
3468        knows more about this stuff than I do.)
3469    
3470    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3471        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3472        match everything, including newlines. This is not POSIX-compatible, but
3473        somebody wanted the feature. From pcretest it can be activated by using
3474        both the P and the s flags.
3475    
3476    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3477    
3478    19. libpcre.pc was being incorrectly installed as executable.
3479    
3480    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3481        it now also looks for '\r' so that it will work unmodified on Windows.
3482    
3483    21. Added Google's contributed C++ wrapper to the distribution.
3484    
3485    22. Added some untidy missing memory free() calls in pcretest, to keep
3486        Electric Fence happy when testing.
3487    
3488    
3489    
3490    Version 5.0 13-Sep-04
3491    ---------------------
3492    
3493     1. Internal change: literal characters are no longer packed up into items
3494        containing multiple characters in a single byte-string. Each character
3495        is now matched using a separate opcode. However, there may be more than one
3496        byte in the character in UTF-8 mode.
3497    
3498     2. The pcre_callout_block structure has two new fields: pattern_position and
3499        next_item_length. These contain the offset in the pattern to the next match
3500        item, and its length, respectively.
3501    
3502     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3503        insertion of callouts before each pattern item. Added the /C option to
3504        pcretest to make use of this.
3505    
3506     4. On the advice of a Windows user, the lines
3507    
3508          #if defined(_WIN32) || defined(WIN32)
3509          _setmode( _fileno( stdout ), 0x8000 );
3510          #endif  /* defined(_WIN32) || defined(WIN32) */
3511    
3512        have been added to the source of pcretest. This apparently does useful
3513        magic in relation to line terminators.
3514    
3515     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3516        for the benefit of those environments where the "b" makes a difference.
3517    
3518     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3519        to know about it. I have put a hack into configure.in that adds in code
3520        to set GCC=yes if CC=icc. This seems to end up at a point in the
3521        generated configure script that is early enough to affect the setting of
3522        compiler options, which is what is needed, but I have no means of testing
3523        whether it really works. (The user who reported this had patched the
3524        generated configure script, which of course I cannot do.)
3525    
3526        LATER: After change 22 below (new libtool files), the configure script
3527        seems to know about icc (and also ecc). Therefore, I have commented out
3528        this hack in configure.in.
3529    
3530     7. Added support for pkg-config (2 patches were sent in).
3531    
3532     8. Negated POSIX character classes that used a combination of internal tables
3533        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3534        [[:^ascii]]. Typically, they would match almost any characters. The other
3535        POSIX classes were not broken in this way.
3536    
3537     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3538        to find the match, as PCRE was deluded into thinking that the match had to
3539        start at the start point or following a newline. The same bug applied to
3540        patterns with negative forward assertions or any backward assertions
3541        preceding ".*" at the start, unless the pattern required a fixed first
3542        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3543    
3544    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3545        starting at the last subject character, bytes beyond the end of the subject
3546        string were read.
3547    
3548    11. Renamed the variable "class" as "classbits" to make life easier for C++
3549        users. (Previously there was a macro definition, but it apparently wasn't
3550        enough.)
3551    
3552    12. Added the new field "tables" to the extra data so that tables can be passed
3553        in at exec time, or the internal tables can be re-selected. This allows
3554        a compiled regex to be saved and re-used at a later time by a different
3555        program that might have everything at different addresses.
3556    
3557    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3558        -R library as well as a -L library.
3559    
3560    14. The debugging options of pcretest (-d on the command line or D on a
3561        pattern) showed incorrect output for anything following an extended class
3562        that contained multibyte characters and which was followed by a quantifier.
3563    
3564    15. Added optional support for general category Unicode character properties
3565        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3566        support. It adds about 90K to the size of the library. The meanings of the
3567        inbuilt class escapes such as \d and \s have NOT been changed.
3568    
3569    16. Updated pcredemo.c to include calls to free() to release the memory for the
3570        compiled pattern.
3571    
3572    17. The generated file chartables.c was being created in the source directory
3573        instead of in the building directory. This caused the build to fail if the
3574        source directory was different from the building directory, and was
3575        read-only.
3576    
3577    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3578        file. No doubt somebody will tell me if they don't make sense... Also added
3579        Dan Mooney's comments about building on OpenVMS.
3580    
3581    19. Added support for partial matching via the PCRE_PARTIAL option for
3582        pcre_exec() and the \P