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

Diff of /code/trunk/ChangeLog

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

revision 23 by nigel, Sat Feb 24 21:38:41 2007 UTC revision 1357 by ph10, Tue Aug 27 15:49:16 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    3.  Make a small performance improvement in strlen16() and strlen32() in
19        pcretest.
20    
21    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23        these cases). There was also one in pcretest.c.
24    
25    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26    
27    6.  In UTF mode, the code for checking whether a group could match an empty
28        string (which is used for indefinitely repeated groups to allow for
29        breaking an infinite loop) was broken when the group contained a repeated
30        negated single-character class with a character that occupied more than one
31        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32        UTF-8 mode). The effect was undefined: the group might or might not be
33        deemed as matching an empty string, or the program might have crashed.
34    
35    7.  The code for checking whether a group could match an empty string was not
36        recognizing that \h, \H, \v, \V, and \R must match a character.
37    
38    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39        an empty string. If it can, pcretest shows this in its information output.
40    
41    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
42        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43        matched by pcre_exec() without using JIT:
44    
45        (a) If the rest of the pattern did not match after a maximal run of
46            grapheme clusters, the code for backing up to try with fewer of them
47            did not always back up over a full grapheme when characters that do not
48            have the modifier quality were involved, e.g. Hangul syllables.
49    
50        (b) If the match point in a subject started with modifier character, and
51            there was no match, the code could incorrectly back up beyond the match
52            point, and potentially beyond the first character in the subject,
53            leading to a segfault or an incorrect match result.
54    
55    10. A conditional group with an assertion condition could lead to PCRE
56        recording an incorrect first data item for a match if no other first data
57        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58        first data item, and therefore matched "ca" after "c" instead of at the
59        start.
60    
61    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63    
64    12. The source of pcregrep now includes z/OS-specific code so that it can be
65        compiled for z/OS as part of the special z/OS distribution.
66    
67    13. Added the -T and -TM options to pcretest.
68    
69    
70    Version 8.33 28-May-2013
71    ------------------------
72    
73    1.  Added 'U' to some constants that are compared to unsigned integers, to
74        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
75        variables that are added to signed variables, to ensure the result is
76        signed and can be negated.
77    
78    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
79        CMake config files.
80    
81    3.  Revise the creation of config.h.generic so that all boolean macros are
82        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
83        overriding via -D on the command line possible.
84    
85    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
86        to unsigned int is reported to make a quite noticeable speed difference in
87        a specific Windows environment. Testing on Linux did also appear to show
88        some benefit (and it is clearly not harmful). Also fixed the definition of
89        Xop which should be unsigned.
90    
91    5.  Related to (4), changing the definition of the intermediate variable cc
92        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
93        improvements.
94    
95    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
96        unnecessary spaces.
97    
98    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
99        and later.
100    
101    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
102    
103    9.  Optimizing fast_forward_start_bits in JIT.
104    
105    10. Adding support for callouts in JIT, and fixing some issues revealed
106        during this work. Namely:
107    
108        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
109    
110        (b) Minimum length was not checked before the matching is started.
111    
112    11. The value of capture_last that is passed to callouts was incorrect in some
113        cases when there was a capture on one path that was subsequently abandoned
114        after a backtrack. Also, the capture_last value is now reset after a
115        recursion, since all captures are also reset in this case.
116    
117    12. The interpreter no longer returns the "too many substrings" error in the
118        case when an overflowing capture is in a branch that is subsequently
119        abandoned after a backtrack.
120    
121    13. In the pathological case when an offset vector of size 2 is used, pcretest
122        now prints out the matched string after a yield of 0 or 1.
123    
124    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
125        Only supported by the JIT compiler at the moment.
126    
127    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
128    
129    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
130        offset of the starting point of the matching process, provided the offsets
131        vector is large enough.
132    
133    17. The \A escape now records a lookbehind value of 1, though its execution
134        does not actually inspect the previous character. This is to ensure that,
135        in partial multi-segment matching, at least one character from the old
136        segment is retained when a new segment is processed. Otherwise, if there
137        are no lookbehinds in the pattern, \A might match incorrectly at the start
138        of a new segment.
139    
140    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
141    
142    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
143        gives some modest performance improvement in 8-bit mode.
144    
145    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
146        be expressed in certain programming languages using Universal Character
147        Names.
148    
149    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
150        which points out that "non characters" are not "characters that may not
151        appear in Unicode strings" but rather "characters that are reserved for
152        internal use and have only local meaning".
153    
154    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
155        there was a conditional group that depended on an assertion, if the
156        assertion was false, the callout that immediately followed the alternation
157        in the condition was skipped when pcre_exec() was used for matching.
158    
159    23. Allow an explicit callout to be inserted before an assertion that is the
160        condition for a conditional group, for compatibility with automatic
161        callouts, which always insert a callout at this point.
162    
163    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
164        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
165    
166    25. (*PRUNE) is now supported by the JIT compiler.
167    
168    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
169    
170    27. Fix the case where there are two or more SKIPs with arguments that may be
171        ignored.
172    
173    28. (*SKIP) is now supported by the JIT compiler.
174    
175    29. (*THEN) is now supported by the JIT compiler.
176    
177    30. Update RunTest with additional test selector options.
178    
179    31. The way PCRE handles backtracking verbs has been changed in two ways.
180    
181        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
182        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
183        backtracking. In some cases this makes it more Perl-compatible, but Perl's
184        rather obscure rules do not always do the same thing.
185    
186        (2) Previously, backtracking verbs were confined within assertions. This is
187        no longer the case for positive assertions, except for (*ACCEPT). Again,
188        this sometimes improves Perl compatibility, and sometimes does not.
189    
190    32. A number of tests that were in test 2 because Perl did things differently
191        have been moved to test 1, because either Perl or PCRE has changed, and
192        these tests are now compatible.
193    
194    32. Backtracking control verbs are now handled in the same way in JIT and
195        interpreter.
196    
197    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
198        contained a forward subroutine reference caused a compile error.
199    
200    34. Auto-detect and optimize limited repetitions in JIT.
201    
202    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
203        blocking (*UTF) etc.
204    
205    36. In the interpreter, maximizing pattern repetitions for characters and
206        character types now use tail recursion, which reduces stack usage.
207    
208    37. The value of the max lookbehind was not correctly preserved if a compiled
209        and saved regex was reloaded on a host of different endianness.
210    
211    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
212        of the compiled pattern block, expand the flags field from 16 to 32 bits
213        because it was almost full.
214    
215    39. Try madvise first before posix_madvise.
216    
217    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
218        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
219        original change locked out all matches of empty strings. This has been
220        changed so that one match of an empty string per line is recognized.
221        Subsequent searches on the same line (for colouring or for --only-matching,
222        for example) do not recognize empty strings.
223    
224    41. Applied a user patch to fix a number of spelling mistakes in comments.
225    
226    42. Data lines longer than 65536 caused pcretest to crash.
227    
228    43. Clarified the data type for length and startoffset arguments for pcre_exec
229        and pcre_dfa_exec in the function-specific man pages, where they were
230        explicitly stated to be in bytes, never having been updated. I also added
231        some clarification to the pcreapi man page.
232    
233    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
234        a segmentation fault.
235    
236    
237    Version 8.32 30-November-2012
238    -----------------------------
239    
240    1.  Improved JIT compiler optimizations for first character search and single
241        character iterators.
242    
243    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
244        Patch by Daniel Richard G.
245    
246    3.  Single character iterator optimizations in the JIT compiler.
247    
248    4.  Improved JIT compiler optimizations for character ranges.
249    
250    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
251        Reported by Giuseppe D'Angelo.
252    
253    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
254        of a line, was being set incorrectly in cases where .* appeared inside
255        atomic brackets at the start of a pattern, or where there was a subsequent
256        *PRUNE or *SKIP.
257    
258    7.  Improved instruction cache flush for POWER/PowerPC.
259        Patch by Daniel Richard G.
260    
261    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
262        grep:
263    
264        (a) There is now no limit to the number of patterns to be matched.
265    
266        (b) An error is given if a pattern is too long.
267    
268        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
269            are now supported.
270    
271        (d) --exclude-from and --include-from (multiple use) have been added.
272    
273        (e) Exclusions and inclusions now apply to all files and directories, not
274            just to those obtained from scanning a directory recursively.
275    
276        (f) Multiple uses of -f and --file-list are now supported.
277    
278        (g) In a Windows environment, the default for -d has been changed from
279            "read" (the GNU grep default) to "skip", because otherwise the presence
280            of a directory in the file list provokes an error.
281    
282        (h) The documentation has been revised and clarified in places.
283    
284    9.  Improve the matching speed of capturing brackets.
285    
286    10. Changed the meaning of \X so that it now matches a Unicode extended
287        grapheme cluster.
288    
289    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
290        out POSIX threads when JIT support is configured.
291    
292    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
293    
294    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
295        structure could go wrong in environments where size_t is not the same size
296        as int.
297    
298    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
299        set.
300    
301    15. The EBCDIC support had decayed; later updates to the code had included
302        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
303        general tidy up of EBCDIC-related issues, and the documentation was also
304        not quite right. There is now a test that can be run on ASCII systems to
305        check some of the EBCDIC-related things (but is it not a full test).
306    
307    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
308        in a small tidy to the code.
309    
310    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
311    
312    18. If the --only-matching (-o) option in pcregrep is specified multiple
313        times, each one causes appropriate output. For example, -o1 -o2 outputs the
314        substrings matched by the 1st and 2nd capturing parentheses. A separating
315        string can be specified by --om-separator (default empty).
316    
317    19. Improving the first n character searches.
318    
319    20. Turn case lists for horizontal and vertical white space into macros so that
320        they are defined only once.
321    
322    21. This set of changes together give more compatible Unicode case-folding
323        behaviour for characters that have more than one other case when UCP
324        support is available.
325    
326        (a) The Unicode property table now has offsets into a new table of sets of
327            three or more characters that are case-equivalent. The MultiStage2.py
328            script that generates these tables (the pcre_ucd.c file) now scans
329            CaseFolding.txt instead of UnicodeData.txt for character case
330            information.
331    
332        (b) The code for adding characters or ranges of characters to a character
333            class has been abstracted into a generalized function that also handles
334            case-independence. In UTF-mode with UCP support, this uses the new data
335            to handle characters with more than one other case.
336    
337        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
338            whose other case is greater than 256 are now correctly matched
339            caselessly. Previously, the high codepoint matched the low one, but not
340            vice versa.
341    
342        (d) The processing of \h, \H, \v, and \ in character classes now makes use
343            of the new class addition function, using character lists defined as
344            macros alongside the case definitions of 20 above.
345    
346        (e) Caseless back references now work with characters that have more than
347            one other case.
348    
349        (f) General caseless matching of characters with more than one other case
350            is supported.
351    
352    22. Unicode character properties were updated from Unicode 6.2.0
353    
354    23. Improved CMake support under Windows. Patch by Daniel Richard G.
355    
356    24. Add support for 32-bit character strings, and UTF-32
357    
358    25. Major JIT compiler update (code refactoring and bugfixing).
359        Experimental Sparc 32 support is added.
360    
361    26. Applied a modified version of Daniel Richard G's patch to create
362        pcre.h.generic and config.h.generic by "make" instead of in the
363        PrepareRelease script.
364    
365    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
366        pcre_compile.c when checking for a zero character.
367    
368    28. Introducing a native interface for JIT. Through this interface, the compiled
369        machine code can be directly executed. The purpose of this interface is to
370        provide fast pattern matching, so several sanity checks are not performed.
371        However, feature tests are still performed. The new interface provides
372        1.4x speedup compared to the old one.
373    
374    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
375        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
376        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
377    
378    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
379        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
380        the "old" RFC 2279). Instead, it ended up passing a negative length to
381        pcre_exec().
382    
383    31. Add support for GCC's visibility feature to hide internal functions.
384    
385    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
386        "unknown -C option" after outputting 0 or 1.
387    
388    33. There is now support for generating a code coverage report for the test
389        suite in environments where gcc is the compiler and lcov is installed. This
390        is mainly for the benefit of the developers.
391    
392    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
393        unaddressable using valgrind annotations, allowing valgrind to detect
394        invalid memory accesses. This is mainly for the benefit of the developers.
395    
396    25. (*UTF) can now be used to start a pattern in any of the three libraries.
397    
398    26. Give configure error if --enable-cpp but no C++ compiler found.
399    
400    
401    Version 8.31 06-July-2012
402    -------------------------
403    
404    1.  Fixing a wrong JIT test case and some compiler warnings.
405    
406    2.  Removed a bashism from the RunTest script.
407    
408    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
409        to unsigned type, result still unsigned" that was given by an MS compiler
410        on encountering the code "-sizeof(xxx)".
411    
412    4.  Partial matching support is added to the JIT compiler.
413    
414    5.  Fixed several bugs concerned with partial matching of items that consist
415        of more than one character:
416    
417        (a) /^(..)\1/ did not partially match "aba" because checking references was
418            done on an "all or nothing" basis. This also applied to repeated
419            references.
420    
421        (b) \R did not give a hard partial match if \r was found at the end of the
422            subject.
423    
424        (c) \X did not give a hard partial match after matching one or more
425            characters at the end of the subject.
426    
427        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
428            a partial match for the string "\r".
429    
430        (e) When newline was set to CRLF, the metacharacter "." did not recognize
431            a partial match for a CR character at the end of the subject string.
432    
433    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
434        running pcretest, the text "(JIT)" added to the output whenever JIT is
435        actually used to run the match.
436    
437    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
438        or /S+[+] with a digit between 1 and 7.
439    
440    8.  OP_NOT now supports any UTF character not just single-byte ones.
441    
442    9.  (*MARK) control verb is now supported by the JIT compiler.
443    
444    10. The command "./RunTest list" lists the available tests without actually
445        running any of them. (Because I keep forgetting what they all are.)
446    
447    11. Add PCRE_INFO_MAXLOOKBEHIND.
448    
449    12. Applied a (slightly modified) user-supplied patch that improves performance
450        when the heap is used for recursion (compiled with --disable-stack-for-
451        recursion). Instead of malloc and free for each heap frame each time a
452        logical recursion happens, frames are retained on a chain and re-used where
453        possible. This sometimes gives as much as 30% improvement.
454    
455    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
456        call.
457    
458    14. As documented, (*COMMIT) is now confined to within a positive assertion.
459    
460    15. It is now possible to link pcretest with libedit as an alternative to
461        libreadline.
462    
463    16. (*COMMIT) control verb is now supported by the JIT compiler.
464    
465    17. The Unicode data tables have been updated to Unicode 6.1.0.
466    
467    18. Added --file-list option to pcregrep.
468    
469    19. Added binary file support to pcregrep, including the -a, --binary-files,
470        -I, and --text options.
471    
472    20. The madvise function is renamed for posix_madvise for QNX compatibility
473        reasons. Fixed by Giuseppe D'Angelo.
474    
475    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
476    
477    22. Changed the option for creating symbolic links for 16-bit man pages from
478        -s to -sf so that re-installing does not cause issues.
479    
480    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
481    
482    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
483        matches in certain environments (the workspace was not being correctly
484        retained). Also added to pcre_dfa_exec() a simple plausibility check on
485        some of the workspace data at the beginning of a restart.
486    
487    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
488        was not doing so when it should - probably a typo introduced by SVN 528
489        (change 8.10/14).
490    
491    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
492        \w+ when the character tables indicated that \x{c4} was a word character.
493        There were several related cases, all because the tests for doing a table
494        lookup were testing for characters less than 127 instead of 255.
495    
496    27. If a pattern contains capturing parentheses that are not used in a match,
497        their slots in the ovector are set to -1. For those that are higher than
498        any matched groups, this happens at the end of processing. In the case when
499        there were back references that the ovector was too small to contain
500        (causing temporary malloc'd memory to be used during matching), and the
501        highest capturing number was not used, memory off the end of the ovector
502        was incorrectly being set to -1. (It was using the size of the temporary
503        memory instead of the true size.)
504    
505    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
506        ovector size, it uses memory at the end of the block that it has got.
507    
508    29. Check for an overlong MARK name and give an error at compile time. The
509        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
510    
511    30. JIT compiler update.
512    
513    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
514        Rill for the patch.
515    
516    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
517    
518    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
519    
520    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
521        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
522        was enabled.
523    
524    35. Improve JIT code generation for greedy plus quantifier.
525    
526    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
527        1 to "aa" instead of to an empty string. The bug affected repeated groups
528        that could potentially match an empty string.
529    
530    37. Optimizing single character iterators in JIT.
531    
532    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
533        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
534        codepoints that are too big for the mode are faulted, and in a UTF mode,
535        disallowed codepoints are also faulted.
536    
537    39. If PCRE was compiled with UTF support, in three places in the DFA
538        matcher there was code that should only have been obeyed in UTF mode, but
539        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
540        processing when bytes with values greater than 127 were present. In 16-bit
541        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
542        both cases the values are those that cannot be the first data item in a UTF
543        character. The three items that might have provoked this were recursions,
544        possessively repeated groups, and atomic groups.
545    
546    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
547        and pcregrep, because some OS require shared objects to be explicitly
548        passed to ld, causing the link step to fail if they are not.
549    
550    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
551        mode, patterns that started with \h* or \R* might be incorrectly matched.
552    
553    
554    Version 8.30 04-February-2012
555    -----------------------------
556    
557    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
558        name is defined in ctype.h.
559    
560    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
561        only in quite long subpatterns.
562    
563    3.  Removed the function pcre_info(), which has been obsolete and deprecated
564        since it was replaced by pcre_fullinfo() in February 2000.
565    
566    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
567        match a (*MARK), and the match failed at the start of the subject, a
568        reference to memory before the start of the subject could occur. This bug
569        was introduced by fix 17 of release 8.21.
570    
571    5.  A reference to an unset group with zero minimum repetition was giving
572        totally wrong answers (in non-JavaScript-compatibility mode). For example,
573        /(another)?(\1?)test/ matched against "hello world test". This bug was
574        introduced in release 8.13.
575    
576    6.  Add support for 16-bit character strings (a large amount of work involving
577        many changes and refactorings).
578    
579    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
580        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
581        from a file.
582    
583    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
584        rounding is not applied in this particular case).
585    
586    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
587        if they appear, or are escaped, in patterns.
588    
589    10. Get rid of a number of -Wunused-but-set-variable warnings.
590    
591    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
592        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
593        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
594        also returns the mark "x". This bug applied to capturing parentheses,
595        non-capturing parentheses, and atomic parentheses. It also applied to some
596        assertions.
597    
598    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
599        information out of configure.ac instead of relying on pcre.h.generic, which
600        is not stored in the repository.
601    
602    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
603        -lreadline.
604    
605    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
606    
607    15. Applied Graycode's patch to put the top-level frame on the stack rather
608        than the heap when not using the stack for recursion. This gives a
609        performance improvement in many cases when recursion is not deep.
610    
611    16. Experimental code added to "pcretest -C" to output the stack frame size.
612    
613    
614    Version 8.21 12-Dec-2011
615    ------------------------
616    
617    1.  Updating the JIT compiler.
618    
619    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
620        are added as well.
621    
622    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
623        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
624        calling _pcre_jit_exec. Some extra comments are added.
625    
626    4.  (*MARK) settings inside atomic groups that do not contain any capturing
627        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
628        was introduced by change 18 for 8.20.
629    
630    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
631        ECMA-262 standard.
632    
633    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
634        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
635        This bug was probably introduced by change 9 of 8.13.
636    
637    7.  While fixing 6 above, I noticed that a number of other items were being
638        incorrectly rejected as "not fixed length". This arose partly because newer
639        opcodes had not been added to the fixed-length checking code. I have (a)
640        corrected the bug and added tests for these items, and (b) arranged for an
641        error to occur if an unknown opcode is encountered while checking for fixed
642        length instead of just assuming "not fixed length". The items that were
643        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
644        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
645        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
646    
647    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
648        being incorrectly compiled and would have given unpredicatble results.
649    
650    9.  A possessively repeated subpattern with minimum repeat count greater than
651        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
652        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
653        the first (A) could occur when it should not.
654    
655    10. Add a cast and remove a redundant test from the code.
656    
657    11. JIT should use pcre_malloc/pcre_free for allocation.
658    
659    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
660        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
661        is anything other than /usr, -L is still shown).
662    
663    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
664    
665    14. Perl does not support \N without a following name in a [] class; PCRE now
666        also gives an error.
667    
668    15. If a forward reference was repeated with an upper limit of around 2000,
669        it caused the error "internal error: overran compiling workspace". The
670        maximum number of forward references (including repeats) was limited by the
671        internal workspace, and dependent on the LINK_SIZE. The code has been
672        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
673        the default depends on LINK_SIZE. There is a new upper limit (for safety)
674        of around 200,000 forward references. While doing this, I also speeded up
675        the filling in of repeated forward references.
676    
677    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
678        incorrectly expecting the subject to contain another "a" after the start.
679    
680    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
681        in the match, the SKIP should be ignored. This was not happening; instead
682        the SKIP was being treated as NOMATCH. For patterns such as
683        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
684        tested.
685    
686    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
687        now much more compatible with Perl, in particular in cases where the result
688        is a non-match for a non-anchored pattern. For example, if
689        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
690        "m", where previously it did not return a name. A side effect of this
691        change is that for partial matches, the last encountered mark name is
692        returned, as for non matches. A number of tests that were previously not
693        Perl-compatible have been moved into the Perl-compatible test files. The
694        refactoring has had the pleasing side effect of removing one argument from
695        the match() function, thus reducing its stack requirements.
696    
697    19. If the /S+ option was used in pcretest to study a pattern using JIT,
698        subsequent uses of /S (without +) incorrectly behaved like /S+.
699    
700    21. Retrieve executable code size support for the JIT compiler and fixing
701        some warnings.
702    
703    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
704        not work when the shorter character appeared right at the end of the
705        subject string.
706    
707    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
708        systems.
709    
710    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
711        output it when the /M option is used in pcretest.
712    
713    25. The CheckMan script was not being included in the distribution. Also, added
714        an explicit "perl" to run Perl scripts from the PrepareRelease script
715        because this is reportedly needed in Windows.
716    
717    26. If study data was being save in a file and studying had not found a set of
718        "starts with" bytes for the pattern, the data written to the file (though
719        never used) was taken from uninitialized memory and so caused valgrind to
720        complain.
721    
722    27. Updated RunTest.bat as provided by Sheri Pierce.
723    
724    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
725    
726    29. Computation of memory usage for the table of capturing group names was
727        giving an unnecessarily large value.
728    
729    
730    Version 8.20 21-Oct-2011
731    ------------------------
732    
733    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
734        a POSIX class. After further experiments with Perl, which convinced me that
735        Perl has bugs and confusions, a closing square bracket is no longer allowed
736        in a POSIX name. This bug also affected patterns with classes that started
737        with full stops.
738    
739    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
740        captured substring, but while checking the failing first alternative,
741        substring 1 is temporarily captured. If the output vector supplied to
742        pcre_exec() was not big enough for this capture, the yield of the function
743        was still zero ("insufficient space for captured substrings"). This cannot
744        be totally fixed without adding another stack variable, which seems a lot
745        of expense for a edge case. However, I have improved the situation in cases
746        such as /(a)(b)x|abc/ matched against "abc", where the return code
747        indicates that fewer than the maximum number of slots in the ovector have
748        been set.
749    
750    3.  Related to (2) above: when there are more back references in a pattern than
751        slots in the output vector, pcre_exec() uses temporary memory during
752        matching, and copies in the captures as far as possible afterwards. It was
753        using the entire output vector, but this conflicts with the specification
754        that only 2/3 is used for passing back captured substrings. Now it uses
755        only the first 2/3, for compatibility. This is, of course, another edge
756        case.
757    
758    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
759        main code base, and can be used by building with --enable-jit. When this is
760        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
761        runtime --no-jit option is given.
762    
763    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
764        ovector, the return from the function was zero, implying that there were
765        other matches that did not fit. The correct "exactly full" value is now
766        returned.
767    
768    6.  If a subpattern that was called recursively or as a subroutine contained
769        (*PRUNE) or any other control that caused it to give a non-standard return,
770        invalid errors such as "Error -26 (nested recursion at the same subject
771        position)" or even infinite loops could occur.
772    
773    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
774        computing the minimum length on reaching *ACCEPT, and so ended up with the
775        wrong value of 1 rather than 0. Further investigation indicates that
776        computing a minimum subject length in the presence of *ACCEPT is difficult
777        (think back references, subroutine calls), and so I have changed the code
778        so that no minimum is registered for a pattern that contains *ACCEPT.
779    
780    8.  If (*THEN) was present in the first (true) branch of a conditional group,
781        it was not handled as intended. [But see 16 below.]
782    
783    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
784        Sheri Pierce.
785    
786    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
787        the first byte in a match must be "a".
788    
789    11. Change 17 for 8.13 increased the recursion depth for patterns like
790        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
791        pattern contains any instances of (*THEN). If it does not, the old
792        optimizations are restored. It would be nice to do this on a per-group
793        basis, but at the moment that is not feasible.
794    
795    12. In some environments, the output of pcretest -C is CRLF terminated. This
796        broke RunTest's code that checks for the link size. A single white space
797        character after the value is now allowed for.
798    
799    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
800        For "fr", it uses the Windows-specific input and output files.
801    
802    14. If (*THEN) appeared in a group that was called recursively or as a
803        subroutine, it did not work as intended. [But see next item.]
804    
805    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
806        pattern fragments (but not containing any | characters). If A and B are
807        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
808        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
809        D. In other words, Perl considers parentheses that do not contain any |
810        characters to be part of a surrounding alternative, whereas PCRE was
811        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
812        differently. PCRE now behaves in the same way as Perl, except in the case
813        of subroutine/recursion calls such as (?1) which have in any case always
814        been different (but PCRE had them first :-).
815    
816    16. Related to 15 above: Perl does not treat the | in a conditional group as
817        creating alternatives. Such a group is treated in the same way as an
818        ordinary group without any | characters when processing (*THEN). PCRE has
819        been changed to match Perl's behaviour.
820    
821    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
822        RunGrepTest script failed.
823    
824    18. Change 22 for version 13 caused atomic groups to use more stack. This is
825        inevitable for groups that contain captures, but it can lead to a lot of
826        stack use in large patterns. The old behaviour has been restored for atomic
827        groups that do not contain any capturing parentheses.
828    
829    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
830        suppress the check for a minimum subject length at run time. (If it was
831        given to pcre_exec() or pcre_dfa_exec() it did work.)
832    
833    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
834        fail to work when decoding hex characters in data strings in EBCDIC
835        environments.
836    
837    21. It appears that in at least one Mac OS environment, the isxdigit() function
838        is implemented as a macro that evaluates to its argument more than once,
839        contravening the C 90 Standard (I haven't checked a later standard). There
840        was an instance in pcretest which caused it to go wrong when processing
841        \x{...} escapes in subject strings. The has been rewritten to avoid using
842        things like p++ in the argument of isxdigit().
843    
844    
845    Version 8.13 16-Aug-2011
846    ------------------------
847    
848    1.  The Unicode data tables have been updated to Unicode 6.0.0.
849    
850    2.  Two minor typos in pcre_internal.h have been fixed.
851    
852    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
853        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
854        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
855    
856    4.  There were a number of related bugs in the code for matching backrefences
857        caselessly in UTF-8 mode when codes for the characters concerned were
858        different numbers of bytes. For example, U+023A and U+2C65 are an upper
859        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
860        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
861        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
862        2-byte code at the end of the subject (it thought there wasn't enough data
863        left).
864    
865    5.  Comprehensive information about what went wrong is now returned by
866        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
867        as the output vector has at least 2 elements. The offset of the start of
868        the failing character and a reason code are placed in the vector.
869    
870    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
871        now returned is for the first byte of the failing character, instead of the
872        last byte inspected. This is an incompatible change, but I hope it is small
873        enough not to be a problem. It makes the returned offset consistent with
874        pcre_exec() and pcre_dfa_exec().
875    
876    7.  pcretest now gives a text phrase as well as the error number when
877        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
878        failure, the offset and reason code are output.
879    
880    8.  When \R was used with a maximizing quantifier it failed to skip backwards
881        over a \r\n pair if the subsequent match failed. Instead, it just skipped
882        back over a single character (\n). This seems wrong (because it treated the
883        two characters as a single entity when going forwards), conflicts with the
884        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
885        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
886        has been changed.
887    
888    9.  Some internal refactoring has changed the processing so that the handling
889        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
890        time (the PCRE_DOTALL option was changed this way some time ago: version
891        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
892        which was always a bit of a fudge. It also means that there is one less
893        argument for the match() function, which reduces its stack requirements
894        slightly. This change also fixes an incompatibility with Perl: the pattern
895        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
896    
897    10. More internal refactoring has drastically reduced the number of recursive
898        calls to match() for possessively repeated groups such as (abc)++ when
899        using pcre_exec().
900    
901    11. While implementing 10, a number of bugs in the handling of groups were
902        discovered and fixed:
903    
904        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
905        (a|)*(?1) gave a compile-time internal error.
906        ((a|)+)+  did not notice that the outer group could match an empty string.
907        (^a|^)+   was not marked as anchored.
908        (.*a|.*)+ was not marked as matching at start or after a newline.
909    
910    12. Yet more internal refactoring has removed another argument from the match()
911        function. Special calls to this function are now indicated by setting a
912        value in a variable in the "match data" data block.
913    
914    13. Be more explicit in pcre_study() instead of relying on "default" for
915        opcodes that mean there is no starting character; this means that when new
916        ones are added and accidentally left out of pcre_study(), testing should
917        pick them up.
918    
919    14. The -s option of pcretest has been documented for ages as being an old
920        synonym of -m (show memory usage). I have changed it to mean "force study
921        for every regex", that is, assume /S for every regex. This is similar to -i
922        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
923        using it. It makes it easier to run collections of tests with and without
924        study enabled, and thereby test pcre_study() more easily. All the standard
925        tests are now run with and without -s (but some patterns can be marked as
926        "never study" - see 20 below).
927    
928    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
929        restoration of the capturing data to the outer values was not happening
930        correctly.
931    
932    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
933        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
934        pattern had matched an empty string, and so incorrectly returned a no
935        match.
936    
937    17. There was optimizing code for the last branch of non-capturing parentheses,
938        and also for the obeyed branch of a conditional subexpression, which used
939        tail recursion to cut down on stack usage. Unfortunately, now that there is
940        the possibility of (*THEN) occurring in these branches, tail recursion is
941        no longer possible because the return has to be checked for (*THEN). These
942        two optimizations have therefore been removed. [But see 8.20/11 above.]
943    
944    18. If a pattern containing \R was studied, it was assumed that \R always
945        matched two bytes, thus causing the minimum subject length to be
946        incorrectly computed because \R can also match just one byte.
947    
948    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
949        was incorrectly computed.
950    
951    20. If /S is present twice on a test pattern in pcretest input, it now
952        *disables* studying, thereby overriding the use of -s on the command line
953        (see 14 above). This is necessary for one or two tests to keep the output
954        identical in both cases.
955    
956    21. When (*ACCEPT) was used in an assertion that matched an empty string and
957        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
958    
959    22. When an atomic group that contained a capturing parenthesis was
960        successfully matched, but the branch in which it appeared failed, the
961        capturing was not being forgotten if a higher numbered group was later
962        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
963        group 1 to "a", when in fact it should be unset. This applied to multi-
964        branched capturing and non-capturing groups, repeated or not, and also to
965        positive assertions (capturing in negative assertions does not happen
966        in PCRE) and also to nested atomic groups.
967    
968    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
969        subject after a captured substring, to make it easier to tell which of a
970        number of identical substrings has been captured.
971    
972    24. The way atomic groups are processed by pcre_exec() has been changed so that
973        if they are repeated, backtracking one repetition now resets captured
974        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
975        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
976        "aaa". Previously, it would have been "a". As part of this code
977        refactoring, the way recursive calls are handled has also been changed.
978    
979    25. If an assertion condition captured any substrings, they were not passed
980        back unless some other capturing happened later. For example, if
981        (?(?=(a))a) was matched against "a", no capturing was returned.
982    
983    26. When studying a pattern that contained subroutine calls or assertions,
984        the code for finding the minimum length of a possible match was handling
985        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
986        group 1 called group 2 while simultaneously a separate group 2 called group
987        1). A stack overflow occurred in this case. I have fixed this by limiting
988        the recursion depth to 10.
989    
990    27. Updated RunTest.bat in the distribution to the version supplied by Tom
991        Fortmann. This supports explicit test numbers on the command line, and has
992        argument validation and error reporting.
993    
994    28. An instance of \X with an unlimited repeat could fail if at any point the
995        first character it looked at was a mark character.
996    
997    29. Some minor code refactoring concerning Unicode properties and scripts
998        should reduce the stack requirement of match() slightly.
999    
1000    30. Added the '=' option to pcretest to check the setting of unused capturing
1001        slots at the end of the pattern, which are documented as being -1, but are
1002        not included in the return count.
1003    
1004    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1005        compiled something random. Now it gives a compile-time error (as does
1006        Perl).
1007    
1008    32. A *MARK encountered during the processing of a positive assertion is now
1009        recorded and passed back (compatible with Perl).
1010    
1011    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1012        had alternative anchored branches, the search for a second match in a line
1013        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1014        matched the line "0102" twice. The same bug affected patterns that started
1015        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1016        twice.
1017    
1018    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1019        does, and because of capturing effects, quantifying parenthesized
1020        assertions may at times be useful. Quantifiers are now allowed for
1021        parenthesized assertions.
1022    
1023    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1024    
1025    36. \g was being checked for fancy things in a character class, when it should
1026        just be a literal "g".
1027    
1028    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1029        appearance of a nested POSIX class supersedes an apparent external class.
1030        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1031        unescaped square brackets may also appear as part of class names. For
1032        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1033        more like Perl. (But see 8.20/1 above.)
1034    
1035    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1036        was because it thought it was \N{name}, which is not supported).
1037    
1038    39. Add minix to OS list not supporting the -S option in pcretest.
1039    
1040    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1041        cannot analyze patterns in sufficient detail to catch mutual recursions
1042        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1043        subgroup is called recursively as a subpattern for a second time at the
1044        same position in the subject string. In previous releases this might have
1045        been caught by the recursion limit, or it might have run out of stack.
1046    
1047    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1048        happen only once. PCRE was, however incorrectly giving a compile time error
1049        "recursive call could loop indefinitely" because it cannot analyze the
1050        pattern in sufficient detail. The compile time test no longer happens when
1051        PCRE is compiling a conditional subpattern, but actual runaway loops are
1052        now caught at runtime (see 40 above).
1053    
1054    42. It seems that Perl allows any characters other than a closing parenthesis
1055        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1056        has been changed to be the same.
1057    
1058    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1059        as not to get warnings when autogen.sh is called. Also changed
1060        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1061    
1062    44. To help people who use pcregrep to scan files containing exceedingly long
1063        lines, the following changes have been made:
1064    
1065        (a) The default value of the buffer size parameter has been increased from
1066            8K to 20K. (The actual buffer used is three times this size.)
1067    
1068        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1069            PCRE is built.
1070    
1071        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1072            to be set at run time.
1073    
1074        (d) Numerical values in pcregrep options can be followed by K or M, for
1075            example --buffer-size=50K.
1076    
1077        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1078            given and the return code is set to 2.
1079    
1080    45. Add a pointer to the latest mark to the callout data block.
1081    
1082    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1083        partial match of an empty string instead of no match. This was specific to
1084        the use of ".".
1085    
1086    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1087        complete match instead of a partial match. This bug was dependent on both
1088        the PCRE_UTF8 and PCRE_DOTALL options being set.
1089    
1090    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1091        starting byte set, because \b was not being ignored.
1092    
1093    
1094    Version 8.12 15-Jan-2011
1095    ------------------------
1096    
1097    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1098        checks for such things as part of the documentation building process.
1099    
1100    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1101        --match-limit and --recursion-limit options (added for 8.11). In
1102        particular, this made one of the standard tests fail. (The integer value
1103        went into the wrong half of a long int.)
1104    
1105    3.  If the --colour option was given to pcregrep with -v (invert match), it
1106        did strange things, either producing crazy output, or crashing. It should,
1107        of course, ignore a request for colour when reporting lines that do not
1108        match.
1109    
1110    4.  Another pcregrep bug caused similar problems if --colour was specified with
1111        -M (multiline) and the pattern match finished with a line ending.
1112    
1113    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1114        matched in multiline mode, the following line was shown as part of the
1115        match. This seems wrong, so I have changed it.
1116    
1117    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1118        the check for further matches in the same line (so they could be coloured)
1119        to overrun the end of the current line. If another match was found, it was
1120        incorrectly shown (and then shown again when found in the next line).
1121    
1122    7.  If pcregrep was compiled under Windows, there was a reference to the
1123        function pcregrep_exit() before it was defined. I am assuming this was
1124        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1125        reported by a user. I've moved the definition above the reference.
1126    
1127    
1128    Version 8.11 10-Dec-2010
1129    ------------------------
1130    
1131    1.  (*THEN) was not working properly if there were untried alternatives prior
1132        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1133        backtracked to try for "b" instead of moving to the next alternative branch
1134        at the same level (in this case, to look for "c"). The Perl documentation
1135        is clear that when (*THEN) is backtracked onto, it goes to the "next
1136        alternative in the innermost enclosing group".
1137    
1138    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1139        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1140        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1141        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1142        (*THEN).
1143    
1144    3.  If \s appeared in a character class, it removed the VT character from
1145        the class, even if it had been included by some previous item, for example
1146        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1147        of \s, but is part of the POSIX "space" class.)
1148    
1149    4.  A partial match never returns an empty string (because you can always
1150        match an empty string at the end of the subject); however the checking for
1151        an empty string was starting at the "start of match" point. This has been
1152        changed to the "earliest inspected character" point, because the returned
1153        data for a partial match starts at this character. This means that, for
1154        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1155        (previously it gave "no match").
1156    
1157    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1158        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1159        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1160        has an implication that the given string is incomplete (because a partial
1161        match is preferred over a full match). For this reason, these items now
1162        give a partial match in this situation. [Aside: previously, the one case
1163        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1164        match rather than a full match, which was wrong by the old rules, but is
1165        now correct.]
1166    
1167    6.  There was a bug in the handling of #-introduced comments, recognized when
1168        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1169        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1170        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1171        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1172        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1173        places in pcre_compile().
1174    
1175    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1176        comments when looking ahead for named forward references to subpatterns,
1177        the only newline sequence it recognized was NL. It now handles newlines
1178        according to the set newline convention.
1179    
1180    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1181        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1182        cater for a lack of strerror(). These oversights have been fixed.
1183    
1184    9.  Added --match-limit and --recursion-limit to pcregrep.
1185    
1186    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1187    
1188    11. When the -o option was used, pcregrep was setting a return code of 1, even
1189        when matches were found, and --line-buffered was not being honoured.
1190    
1191    12. Added an optional parentheses number to the -o and --only-matching options
1192        of pcregrep.
1193    
1194    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1195        can match an empty string. The code to do it in pcretest and pcredemo
1196        needed fixing:
1197    
1198        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1199            only one byte after an empty string match just before CRLF (this case
1200            just got forgotten; "any" and "anycrlf" were OK).
1201    
1202        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1203            mode when an empty string match preceded an ASCII character followed by
1204            a non-ASCII character. (The code for advancing by one character rather
1205            than one byte was nonsense.)
1206    
1207        (c) The pcredemo.c sample program did not have any code at all to handle
1208            the cases when CRLF is a valid newline sequence.
1209    
1210    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1211        as a starting offset was within the subject string. There is now a new
1212        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1213        negative or greater than the length of the string. In order to test this,
1214        pcretest is extended to allow the setting of negative starting offsets.
1215    
1216    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1217        starting offset points to the beginning of a UTF-8 character was
1218        unnecessarily clumsy. I tidied it up.
1219    
1220    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1221        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1222    
1223    17. Nobody had reported that the --include_dir option, which was added in
1224        release 7.7 should have been called --include-dir (hyphen, not underscore)
1225        for compatibility with GNU grep. I have changed it to --include-dir, but
1226        left --include_dir as an undocumented synonym, and the same for
1227        --exclude-dir, though that is not available in GNU grep, at least as of
1228        release 2.5.4.
1229    
1230    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1231        characters from a string of bytes) have been redefined so as not to use
1232        loops, in order to improve performance in some environments. At the same
1233        time, I abstracted some of the common code into auxiliary macros to save
1234        repetition (this should not affect the compiled code).
1235    
1236    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1237        compile-time error is now given if \c is not followed by an ASCII
1238        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1239        different, and any byte value is allowed.)
1240    
1241    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1242        START_OPTIMIZE option, which is now allowed at compile time - but just
1243        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1244        to pcregrep and other applications that have no direct access to PCRE
1245        options. The new /Y option in pcretest sets this option when calling
1246        pcre_compile().
1247    
1248    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1249        back references. Groups containing recursive back references were forced to
1250        be atomic by that change, but in the case of named groups, the amount of
1251        memory required was incorrectly computed, leading to "Failed: internal
1252        error: code overflow". This has been fixed.
1253    
1254    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1255        pcretest.c, to avoid build problems in some Borland environments.
1256    
1257    
1258    Version 8.10 25-Jun-2010
1259    ------------------------
1260    
1261    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1262        THEN.
1263    
1264    2.  (*ACCEPT) was not working when inside an atomic group.
1265    
1266    3.  Inside a character class, \B is treated as a literal by default, but
1267        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1268        causes the error). The code is unchanged, but I tidied the documentation.
1269    
1270    4.  Inside a character class, PCRE always treated \R and \X as literals,
1271        whereas Perl faults them if its -w option is set. I have changed PCRE so
1272        that it faults them when PCRE_EXTRA is set.
1273    
1274    5.  Added support for \N, which always matches any character other than
1275        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1276    
1277    6.  When compiling pcregrep with newer versions of gcc which may have
1278        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1279        declared with attribute warn_unused_result" were given. Just casting the
1280        result to (void) does not stop the warnings; a more elaborate fudge is
1281        needed. I've used a macro to implement this.
1282    
1283    7.  Minor change to pcretest.c to avoid a compiler warning.
1284    
1285    8.  Added four artifical Unicode properties to help with an option to make
1286        \s etc use properties (see next item). The new properties are: Xan
1287        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1288    
1289    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1290        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1291        this option. Modified pcretest to add /W to test this facility. Added
1292        REG_UCP to make it available via the POSIX interface.
1293    
1294    10. Added --line-buffered to pcregrep.
1295    
1296    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1297        studied, and the match started with a letter with a code point greater than
1298        127 whose first byte was different to the first byte of the other case of
1299        the letter, the other case of this starting letter was not recognized
1300        (#976).
1301    
1302    12. If a pattern that was studied started with a repeated Unicode property
1303        test, for example, \p{Nd}+, there was the theoretical possibility of
1304        setting up an incorrect bitmap of starting bytes, but fortunately it could
1305        not have actually happened in practice until change 8 above was made (it
1306        added property types that matched character-matching opcodes).
1307    
1308    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1309        possible starting bytes for non-anchored patterns.
1310    
1311    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1312        \R, and also a number of cases that involve Unicode properties, both
1313        explicit and implicit when PCRE_UCP is set.
1314    
1315    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1316        input, it could crash or give wrong results if characters with values
1317        greater than 0xc0 were present in the subject string. (Detail: it assumed
1318        UTF-8 input when processing these items.)
1319    
1320    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1321        size_t is 64-bit (#991).
1322    
1323    17. Added a check for running out of memory when PCRE is compiled with
1324        --disable-stack-for-recursion (#990).
1325    
1326    18. If the last data line in a file for pcretest does not have a newline on
1327        the end, a newline was missing in the output.
1328    
1329    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1330        less than 128) in its various bitmaps. However, there is a facility for
1331        generating tables according to the current locale when PCRE is compiled. It
1332        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1333        characters, are recognized by isspace() and therefore were getting set in
1334        these tables, and indeed these tables seem to approximate to ISO 8859. This
1335        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1336        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1337        which of course cannot start UTF-8 characters. I have changed the code so
1338        that only real ASCII characters (less than 128) and the correct starting
1339        bytes for UTF-8 encodings are set for characters greater than 127 when in
1340        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1341        altogether.)
1342    
1343    20. Added the /T option to pcretest so as to be able to run tests with non-
1344        standard character tables, thus making it possible to include the tests
1345        used for 19 above in the standard set of tests.
1346    
1347    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1348        reference to a subpattern the other side of a comment that contains an
1349        opening parenthesis caused either an internal compiling error, or a
1350        reference to the wrong subpattern.
1351    
1352    
1353    Version 8.02 19-Mar-2010
1354    ------------------------
1355    
1356    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1357    
1358    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1359        configured.
1360    
1361    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1362        original author of that file, following a query about its status.
1363    
1364    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1365        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1366    
1367    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1368        quantifier applied to a forward-referencing subroutine call, could compile
1369        incorrect code or give the error "internal error: previously-checked
1370        referenced subpattern not found".
1371    
1372    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1373        variables to point to external functions. For these systems, therefore,
1374        pcre_malloc etc. are now initialized to local functions that call the
1375        relevant global functions.
1376    
1377    7.  There were two entries missing in the vectors called coptable and poptable
1378        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1379        I've fixed the data, and added a kludgy way of testing at compile time that
1380        the lengths are correct (equal to the number of opcodes).
1381    
1382    8.  Following on from 7, I added a similar kludge to check the length of the
1383        eint vector in pcreposix.c.
1384    
1385    9.  Error texts for pcre_compile() are held as one long string to avoid too
1386        much relocation at load time. To find a text, the string is searched,
1387        counting zeros. There was no check for running off the end of the string,
1388        which could happen if a new error number was added without updating the
1389        string.
1390    
1391    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1392    
1393    11. \K was not working if it appeared in an atomic group or in a group that
1394        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1395        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1396        the assertion is positive, but not if it is negative.
1397    
1398    12. Change 11 fortuitously reduced the size of the stack frame used in the
1399        "match()" function of pcre_exec.c by one pointer. Forthcoming
1400        implementation of support for (*MARK) will need an extra pointer on the
1401        stack; I have reserved it now, so that the stack frame size does not
1402        decrease.
1403    
1404    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1405        item in branch that calls a recursion is a subroutine call - as in the
1406        second branch in the above example - was incorrectly given the compile-
1407        time error "recursive call could loop indefinitely" because pcre_compile()
1408        was not correctly checking the subroutine for matching a non-empty string.
1409    
1410    14. The checks for overrunning compiling workspace could trigger after an
1411        overrun had occurred. This is a "should never occur" error, but it can be
1412        triggered by pathological patterns such as hundreds of nested parentheses.
1413        The checks now trigger 100 bytes before the end of the workspace.
1414    
1415    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1416    
1417    
1418    Version 8.01 19-Jan-2010
1419    ------------------------
1420    
1421    1.  If a pattern contained a conditional subpattern with only one branch (in
1422        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1423        computed the wrong minimum data length (which is of course zero for such
1424        subpatterns). This could cause incorrect "no match" results.
1425    
1426    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1427        the pattern is reset in the first branch, pcre_compile() failed with
1428        "internal error: code overflow at offset...". This happened only when
1429        the reset was to the original external option setting. (An optimization
1430        abstracts leading options settings into an external setting, which was the
1431        cause of this.)
1432    
1433    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1434        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1435        assertion pattern did not match (meaning that the assertion was true), it
1436        was incorrectly treated as false if the SKIP had been reached during the
1437        matching. This also applied to assertions used as conditions.
1438    
1439    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1440        assertion subpattern, including such a pattern used as a condition,
1441        unpredictable results occurred, instead of the error return
1442        PCRE_ERROR_DFA_UITEM.
1443    
1444    5.  The C++ GlobalReplace function was not working like Perl for the special
1445        situation when an empty string is matched. It now does the fancy magic
1446        stuff that is necessary.
1447    
1448    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1449        removed. (These were left over from very, very early versions of PCRE.)
1450    
1451    7.  Some cosmetic changes to the code to make life easier when compiling it
1452        as part of something else:
1453    
1454        (a) Change DEBUG to PCRE_DEBUG.
1455    
1456        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1457            called "current" as "current_branch", to prevent a collision with the
1458            Linux macro when compiled as a kernel module.
1459    
1460        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1461            prevent a collision with the Linux macro when compiled as a kernel
1462            module.
1463    
1464    8.  In pcre_compile() there are some checks for integer overflows that used to
1465        cast potentially large values to (double). This has been changed to that
1466        when building, a check for int64_t is made, and if it is found, it is used
1467        instead, thus avoiding the use of floating point arithmetic. (There is no
1468        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1469        double.
1470    
1471    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1472        2005 (difference between two addresses compared to an unsigned value).
1473    
1474    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1475        custom one, because of the following reported problem in Windows:
1476    
1477          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1478              under Win32.
1479          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1480              therefore missing the function definition.
1481          - The compiler thus generates a "C" signature for the test function.
1482          - The linker fails to find the "C" function.
1483          - PCRE fails to configure if asked to do so against libbz2.
1484    
1485    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1486        messages were output:
1487    
1488          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1489          rerunning libtoolize, to keep the correct libtool macros in-tree.
1490          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1491    
1492        I have done both of these things.
1493    
1494    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1495        most of the time, it *can* run out if it is given a pattern that contains a
1496        runaway infinite recursion. I updated the discussion in the pcrestack man
1497        page.
1498    
1499    13. Now that we have gone to the x.xx style of version numbers, the minor
1500        version may start with zero. Using 08 or 09 is a bad idea because users
1501        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1502        interpreted as invalid octal numbers. I've updated the previous comment in
1503        configure.ac, and also added a check that gives an error if 08 or 09 are
1504        used.
1505    
1506    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1507        causing partial matching to fail when the end of the subject matched \W
1508        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1509    
1510    15. There were some discrepancies between the declarations in pcre_internal.h
1511        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1512        their definitions. The declarations used "const uschar *" and the
1513        definitions used USPTR. Even though USPTR is normally defined as "const
1514        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1515        reported that: "This difference in casting confuses some C++ compilers, for
1516        example, SunCC recognizes above declarations as different functions and
1517        generates broken code for hbpcre." I have changed the declarations to use
1518        USPTR.
1519    
1520    16. GNU libtool is named differently on some systems. The autogen.sh script now
1521        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1522        (FreeBSD).
1523    
1524    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1525        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1526        comment: "Figure out how to create a longlong from a string: strtoll and
1527        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1528        instance, but it only takes 2 args instead of 3!"
1529    
1530    18. A subtle bug concerned with back references has been fixed by a change of
1531        specification, with a corresponding code fix. A pattern such as
1532        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1533        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1534        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1535        same bug. Such groups have to be quantified to be useful, or contained
1536        inside another quantified group. (If there's no repetition, the reference
1537        can never match.) The problem arises because, having left the group and
1538        moved on to the rest of the pattern, a later failure that backtracks into
1539        the group uses the captured value from the final iteration of the group
1540        rather than the correct earlier one. I have fixed this in PCRE by forcing
1541        any group that contains a reference to itself to be an atomic group; that
1542        is, there cannot be any backtracking into it once it has completed. This is
1543        similar to recursive and subroutine calls.
1544    
1545    
1546    Version 8.00 19-Oct-09
1547    ----------------------
1548    
1549    1.  The table for translating pcre_compile() error codes into POSIX error codes
1550        was out-of-date, and there was no check on the pcre_compile() error code
1551        being within the table. This could lead to an OK return being given in
1552        error.
1553    
1554    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1555        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1556        in a Windows environment.
1557    
1558    3.  The pcregrep --count option prints the count for each file even when it is
1559        zero, as does GNU grep. However, pcregrep was also printing all files when
1560        --files-with-matches was added. Now, when both options are given, it prints
1561        counts only for those files that have at least one match. (GNU grep just
1562        prints the file name in this circumstance, but including the count seems
1563        more useful - otherwise, why use --count?) Also ensured that the
1564        combination -clh just lists non-zero counts, with no names.
1565    
1566    4.  The long form of the pcregrep -F option was incorrectly implemented as
1567        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1568        but it seems right to fix it, and I didn't think it was worth preserving
1569        the old behaviour.
1570    
1571    5.  The command line items --regex=pattern and --regexp=pattern were not
1572        recognized by pcregrep, which required --regex pattern or --regexp pattern
1573        (with a space rather than an '='). The man page documented the '=' forms,
1574        which are compatible with GNU grep; these now work.
1575    
1576    6.  No libpcreposix.pc file was created for pkg-config; there was just
1577        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1578    
1579    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1580        when UCP support is not needed, by modifying the Python script that
1581        generates it from Unicode data files. This should not matter if the module
1582        is correctly used as a library, but I received one complaint about 50K of
1583        unwanted data. My guess is that the person linked everything into his
1584        program rather than using a library. Anyway, it does no harm.
1585    
1586    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1587        was a minimum greater than 1 for a wide character in a possessive
1588        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1589        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1590        character. Chaos in the form of incorrect output or a compiling loop could
1591        result.
1592    
1593    9.  The restrictions on what a pattern can contain when partial matching is
1594        requested for pcre_exec() have been removed. All patterns can now be
1595        partially matched by this function. In addition, if there are at least two
1596        slots in the offset vector, the offset of the earliest inspected character
1597        for the match and the offset of the end of the subject are set in them when
1598        PCRE_ERROR_PARTIAL is returned.
1599    
1600    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1601        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1602        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1603        and may be more useful for multi-segment matching.
1604    
1605    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1606        used to be given if ever the end of the subject was reached; now it is
1607        given only if matching could not proceed because another character was
1608        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1609        string "Z", which now yields "no match" instead of "partial match". In the
1610        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1611        final character ended with (*FAIL).
1612    
1613    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1614        if the pattern had a "must contain" character that was already found in the
1615        earlier partial match, unless partial matching was again requested. For
1616        example, with the pattern /dog.(body)?/, the "must contain" character is
1617        "g". If the first part-match was for the string "dog", restarting with
1618        "sbody" failed. This bug has been fixed.
1619    
1620    13. The string returned by pcre_dfa_exec() after a partial match has been
1621        changed so that it starts at the first inspected character rather than the
1622        first character of the match. This makes a difference only if the pattern
1623        starts with a lookbehind assertion or \b or \B (\K is not supported by
1624        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1625        matching functions compatible, and I think it's the right thing to do.
1626    
1627    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1628        so that the demonstration program is easily available in environments where
1629        PCRE has not been installed from source.
1630    
1631    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1632        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1633        library.
1634    
1635    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1636        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1637        is not the first non-POSIX option to be added. Clearly some people find
1638        these options useful.
1639    
1640    17. If a caller to the POSIX matching function regexec() passes a non-zero
1641        value for nmatch with a NULL value for pmatch, the value of
1642        nmatch is forced to zero.
1643    
1644    18. RunGrepTest did not have a test for the availability of the -u option of
1645        the diff command, as RunTest does. It now checks in the same way as
1646        RunTest, and also checks for the -b option.
1647    
1648    19. If an odd number of negated classes containing just a single character
1649        interposed, within parentheses, between a forward reference to a named
1650        subpattern and the definition of the subpattern, compilation crashed with
1651        an internal error, complaining that it could not find the referenced
1652        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1653        [The bug was that it was starting one character too far in when skipping
1654        over the character class, thus treating the ] as data rather than
1655        terminating the class. This meant it could skip too much.]
1656    
1657    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1658        /g option in pcretest when the pattern contains \K, which makes it possible
1659        to have an empty string match not at the start, even when the pattern is
1660        anchored. Updated pcretest and pcredemo to use this option.
1661    
1662    21. If the maximum number of capturing subpatterns in a recursion was greater
1663        than the maximum at the outer level, the higher number was returned, but
1664        with unset values at the outer level. The correct (outer level) value is
1665        now given.
1666    
1667    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1668        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1669        make it do so. The string so far is captured, making this feature
1670        compatible with Perl.
1671    
1672    23. The tests have been re-organized, adding tests 11 and 12, to make it
1673        possible to check the Perl 5.10 features against Perl 5.10.
1674    
1675    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1676        pattern matches a fixed length string. PCRE did not allow this; now it
1677        does. Neither allows recursion.
1678    
1679    25. I finally figured out how to implement a request to provide the minimum
1680        length of subject string that was needed in order to match a given pattern.
1681        (It was back references and recursion that I had previously got hung up
1682        on.) This code has now been added to pcre_study(); it finds a lower bound
1683        to the length of subject needed. It is not necessarily the greatest lower
1684        bound, but using it to avoid searching strings that are too short does give
1685        some useful speed-ups. The value is available to calling programs via
1686        pcre_fullinfo().
1687    
1688    26. While implementing 25, I discovered to my embarrassment that pcretest had
1689        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1690        study optimizations had never been tested with that matching function.
1691        Oops. What is worse, even when it was passed study data, there was a bug in
1692        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1693        were also very few tests of studied patterns with pcre_dfa_exec().
1694    
1695    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1696        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1697        on the other side of the coin, they are no longer allowed to have different
1698        names, because these cannot be distinguished in PCRE, and this has caused
1699        confusion. (This is a difference from Perl.)
1700    
1701    28. When duplicate subpattern names are present (necessarily with different
1702        numbers, as required by 27 above), and a test is made by name in a
1703        conditional pattern, either for a subpattern having been matched, or for
1704        recursion in such a pattern, all the associated numbered subpatterns are
1705        tested, and the overall condition is true if the condition is true for any
1706        one of them. This is the way Perl works, and is also more like the way
1707        testing by number works.
1708    
1709    
1710    Version 7.9 11-Apr-09
1711    ---------------------
1712    
1713    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1714        (pcretest), all targets were linked against these libraries. This included
1715        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1716        libraries. This caused unwanted dependencies to be created. This problem
1717        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1718        pcretest is linked with readline.
1719    
1720    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1721        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1722        moved outside it again, because FALSE and TRUE are already defined in AIX,
1723        but BOOL is not.
1724    
1725    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1726        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1727    
1728    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1729        hyphens) following file names and line numbers when outputting matching
1730        lines. This is not true; no spaces are inserted. I have also clarified the
1731        wording for the --colour (or --color) option.
1732    
1733    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1734        was not coloured; this is different to GNU grep, so I have changed it to be
1735        the same.
1736    
1737    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1738        each matching line was coloured. Now it goes on to look for further matches
1739        of any of the test patterns, which is the same behaviour as GNU grep.
1740    
1741    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1742        doesn't make sense to accept an empty string match in pcregrep, so I have
1743        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1744        seems to be how GNU grep behaves. [But see later change 40 for release
1745        8.33.]
1746    
1747    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1748        start or after a newline", because the conditional assertion was not being
1749        correctly handled. The rule now is that both the assertion and what follows
1750        in the first alternative must satisfy the test.
1751    
1752    9.  If auto-callout was enabled in a pattern with a conditional group whose
1753        condition was an assertion, PCRE could crash during matching, both with
1754        pcre_exec() and pcre_dfa_exec().
1755    
1756    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1757        used for matching.
1758    
1759    11. Unicode property support in character classes was not working for
1760        characters (bytes) greater than 127 when not in UTF-8 mode.
1761    
1762    12. Added the -M command line option to pcretest.
1763    
1764    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1765    
1766    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1767    
1768    16. Added comments and documentation about mis-use of no_arg in the C++
1769        wrapper.
1770    
1771    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1772        from Martin Jerabek that uses macro names for all relevant character and
1773        string constants.
1774    
1775    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1776        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1777        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1778        these, but not everybody uses configure.
1779    
1780    19. A conditional group that had only one branch was not being correctly
1781        recognized as an item that could match an empty string. This meant that an
1782        enclosing group might also not be so recognized, causing infinite looping
1783        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1784        with the subject "ab", where knowledge that the repeated group can match
1785        nothing is needed in order to break the loop.
1786    
1787    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1788        exec(), but without supplying a callout function, matching went wrong.
1789    
1790    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1791        leak if the size of the offset vector was greater than 30. When the vector
1792        is smaller, the saved offsets during recursion go onto a local stack
1793        vector, but for larger vectors malloc() is used. It was failing to free
1794        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1795        error, in fact).
1796    
1797    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1798        heapframe that is used only when UTF-8 support is enabled. This caused no
1799        problem, but was untidy.
1800    
1801    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1802        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1803        included within another project.
1804    
1805    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1806        slightly modified by me:
1807    
1808          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1809              not building pcregrep.
1810    
1811          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1812              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1813    
1814    25. Forward references, both numeric and by name, in patterns that made use of
1815        duplicate group numbers, could behave incorrectly or give incorrect errors,
1816        because when scanning forward to find the reference group, PCRE was not
1817        taking into account the duplicate group numbers. A pattern such as
1818        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1819    
1820    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1821        the feature of a custom pointer more persuasive (as requested by a user).
1822    
1823    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1824        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1825        pre-defined.
1826    
1827    28. Added support for (*UTF8) at the start of a pattern.
1828    
1829    29. Arrange for flags added by the "release type" setting in CMake to be shown
1830        in the configuration summary.
1831    
1832    
1833    Version 7.8 05-Sep-08
1834    ---------------------
1835    
1836    1.  Replaced UCP searching code with optimized version as implemented for Ad
1837        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1838        stage table and inline lookup instead of a function, giving speed ups of 2
1839        to 5 times on some simple patterns that I tested. Permission was given to
1840        distribute the MultiStage2.py script that generates the tables (it's not in
1841        the tarball, but is in the Subversion repository).
1842    
1843    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1844        scripts.
1845    
1846    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1847        a group with a zero qualifier. The result of the study could be incorrect,
1848        or the function might crash, depending on the pattern.
1849    
1850    4.  Caseless matching was not working for non-ASCII characters in back
1851        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1852        It now works when Unicode Property Support is available.
1853    
1854    5.  In pcretest, an escape such as \x{de} in the data was always generating
1855        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1856        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1857        truncation.
1858    
1859    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1860    
1861    7.  Added two (int) casts to pcregrep when printing the difference of two
1862        pointers, in case they are 64-bit values.
1863    
1864    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1865        test 2 if it fails.
1866    
1867    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1868        and a #define of that name to empty if it is not externally set. This is to
1869        allow users of MSVC to set it if necessary.
1870    
1871    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1872        the convenience functions in the pcre_get.c source file.
1873    
1874    11. An option change at the start of a pattern that had top-level alternatives
1875        could cause overwriting and/or a crash. This command provoked a crash in
1876        some environments:
1877    
1878          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1879    
1880        This potential security problem was recorded as CVE-2008-2371.
1881    
1882    12. For a pattern where the match had to start at the beginning or immediately
1883        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1884        pcre_dfa_exec() could read past the end of the passed subject if there was
1885        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1886        pcretest so that it places the subject at the end of its malloc-ed buffer.
1887    
1888    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1889        exec() might read past the end of the data buffer in UTF-8 mode.
1890    
1891    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1892        the data contained the byte 0x85 as part of a UTF-8 character within its
1893        first line. This applied both to normal and DFA matching.
1894    
1895    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1896        /^[^d]*?$/8 failed to match "abc".
1897    
1898    16. Added a missing copyright notice to pcrecpp_internal.h.
1899    
1900    17. Make it more clear in the documentation that values returned from
1901        pcre_exec() in ovector are byte offsets, not character counts.
1902    
1903    18. Tidied a few places to stop certain compilers from issuing warnings.
1904    
1905    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1906        supplied by Stefan Weber. I made a further small update for 7.8 because
1907        there is a change of source arrangements: the pcre_searchfuncs.c module is
1908        replaced by pcre_ucd.c.
1909    
1910    
1911    Version 7.7 07-May-08
1912    ---------------------
1913    
1914    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1915        a string to a long long, pretend we don't even have a long long." This is
1916        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1917    
1918    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1919        pre-7.6 versions, which defined a global no_arg variable instead of putting
1920        it in the RE class. (See also #8 below.)
1921    
1922    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1923        Lopes.
1924    
1925    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1926    
1927        (1) The include/exclude patterns were being applied to the whole pathnames
1928            of files, instead of just to the final components.
1929    
1930        (2) If there was more than one level of directory, the subdirectories were
1931            skipped unless they satisfied the include/exclude conditions. This is
1932            inconsistent with GNU grep (and could even be seen as contrary to the
1933            pcregrep specification - which I improved to make it absolutely clear).
1934            The action now is always to scan all levels of directory, and just
1935            apply the include/exclude patterns to regular files.
1936    
1937    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1938        --exclude_dir in the tests to avoid scanning .svn directories.
1939    
1940    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1941        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1942        doesn't support NULs in patterns.
1943    
1944    7.  Added some missing "const"s to declarations of static tables in
1945        pcre_compile.c and pcre_dfa_exec.c.
1946    
1947    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1948        caused by fix #2  above. (Subsequently also a second patch to fix the
1949        first patch. And a third patch - this was a messy problem.)
1950    
1951    9.  Applied Craig's patch to remove the use of push_back().
1952    
1953    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1954        matching function regexec().
1955    
1956    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1957        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1958        references. PCRE supports relative numbers with this syntax (I don't think
1959        Oniguruma does).
1960    
1961    12. Previously, a group with a zero repeat such as (...){0} was completely
1962        omitted from the compiled regex. However, this means that if the group
1963        was called as a subroutine from elsewhere in the pattern, things went wrong
1964        (an internal error was given). Such groups are now left in the compiled
1965        pattern, with a new opcode that causes them to be skipped at execution
1966        time.
1967    
1968    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1969        to the way PCRE behaves:
1970    
1971        (a) A lone ] character is dis-allowed (Perl treats it as data).
1972    
1973        (b) A back reference to an unmatched subpattern matches an empty string
1974            (Perl fails the current match path).
1975    
1976        (c) A data ] in a character class must be notated as \] because if the
1977            first data character in a class is ], it defines an empty class. (In
1978            Perl it is not possible to have an empty class.) The empty class []
1979            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1980            The negative empty class [^] matches any one character, independently
1981            of the DOTALL setting.
1982    
1983    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1984        non-existent subpattern following a character class starting with ']' and
1985        containing () gave an internal compiling error instead of "reference to
1986        non-existent subpattern". Fortunately, when the pattern did exist, the
1987        compiled code was correct. (When scanning forwards to check for the
1988        existence of the subpattern, it was treating the data ']' as terminating
1989        the class, so got the count wrong. When actually compiling, the reference
1990        was subsequently set up correctly.)
1991    
1992    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1993        it was being rejected as not supported by pcre_dfa_exec(), even though
1994        other assertions are supported. I have made pcre_dfa_exec() support
1995        (*FAIL).
1996    
1997    16. The implementation of 13c above involved the invention of a new opcode,
1998        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1999        cannot be changed at match time, I realized I could make a small
2000        improvement to matching performance by compiling OP_ALLANY instead of
2001        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2002        on the OP_ANY path.
2003    
2004    17. Compiling pcretest on Windows with readline support failed without the
2005        following two fixes: (1) Make the unistd.h include conditional on
2006        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2007    
2008    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2009        ncurses library to be included for pcretest when ReadLine support is
2010        requested, but also to allow for it to be overridden. This patch came from
2011        Daniel Bergström.
2012    
2013    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2014        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2015        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2016        spotting this.
2017    
2018    
2019    Version 7.6 28-Jan-08
2020    ---------------------
2021    
2022    1.  A character class containing a very large number of characters with
2023        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2024        overflow.
2025    
2026    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2027        HAVE_LONG_LONG is not defined.
2028    
2029    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2030        bring them up to date and include new features. This patch includes:
2031    
2032        - Fixed PH's badly added libz and libbz2 support.
2033        - Fixed a problem with static linking.
2034        - Added pcredemo. [But later removed - see 7 below.]
2035        - Fixed dftables problem and added an option.
2036        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2037            HAVE_LONG_LONG.
2038        - Added readline support for pcretest.
2039        - Added an listing of the option settings after cmake has run.
2040    
2041    4.  A user submitted a patch to Makefile that makes it easy to create
2042        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2043        Makefile.am that cause it to include this special target, without
2044        affecting anything else. Note that the same mingw target plus all
2045        the other distribution libraries and programs are now supported
2046        when configuring with CMake (see 6 below) instead of with
2047        Configure/Make.
2048    
2049    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2050        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2051        exported in the Windows port". It has not yet been confirmed that the patch
2052        solves the problem, but it does no harm.
2053    
2054    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2055        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2056        with CMake, and also correct the comment about stack recursion.
2057    
2058    7.  Remove the automatic building of pcredemo from the ./configure system and
2059        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2060        of a program that users should build themselves after PCRE is installed, so
2061        building it automatically is not really right. What is more, it gave
2062        trouble in some build environments.
2063    
2064    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2065    
2066    
2067    Version 7.5 10-Jan-08
2068    ---------------------
2069    
2070    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2071        values in parens when parsing an RE using the C++ wrapper."
2072    
2073    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2074        Characters greater than 255 were excluded from the class instead of being
2075        included.
2076    
2077    3.  The same bug as (2) above applied to negated POSIX classes such as
2078        [:^space:].
2079    
2080    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2081        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2082        I have changed it.
2083    
2084    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2085        first named subpattern) and a construct such as (?&a) would reference the
2086        first named subpattern whose name started with "a" (in other words, the
2087        length check was missing). Both these problems are fixed. "Subpattern name
2088        expected" is now given for (?&) (a zero-length name), and this patch also
2089        makes it give the same error for \k'' (previously it complained that that
2090        was a reference to a non-existent subpattern).
2091    
2092    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2093        this is right because (?- can be followed by option settings as well as by
2094        digits. I have, however, made the messages clearer.
2095    
2096    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2097        than the number used in the conditional) now cause a compile-time error.
2098        This is actually not compatible with Perl, which accepts such patterns, but
2099        treats the conditional as always being FALSE (as PCRE used to), but it
2100        seems to me that giving a diagnostic is better.
2101    
2102    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2103        and messages.
2104    
2105    9.  Fix two occurrences of "backslash" in comments that should have been
2106        "backspace".
2107    
2108    10. Remove two redundant lines of code that can never be obeyed (their function
2109        was moved elsewhere).
2110    
2111    11. The program that makes PCRE's Unicode character property table had a bug
2112        which caused it to generate incorrect table entries for sequences of
2113        characters that have the same character type, but are in different scripts.
2114        It amalgamated them into a single range, with the script of the first of
2115        them. In other words, some characters were in the wrong script. There were
2116        thirteen such cases, affecting characters in the following ranges:
2117    
2118          U+002b0 - U+002c1
2119          U+0060c - U+0060d
2120          U+0061e - U+00612
2121          U+0064b - U+0065e
2122          U+0074d - U+0076d
2123          U+01800 - U+01805
2124          U+01d00 - U+01d77
2125          U+01d9b - U+01dbf
2126          U+0200b - U+0200f
2127          U+030fc - U+030fe
2128          U+03260 - U+0327f
2129          U+0fb46 - U+0fbb1
2130          U+10450 - U+1049d
2131    
2132    12. The -o option (show only the matching part of a line) for pcregrep was not
2133        compatible with GNU grep in that, if there was more than one match in a
2134        line, it showed only the first of them. It now behaves in the same way as
2135        GNU grep.
2136    
2137    13. If the -o and -v options were combined for pcregrep, it printed a blank
2138        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2139        does the same. The return code can be used to tell if there were any
2140        non-matching lines.
2141    
2142    14. Added --file-offsets and --line-offsets to pcregrep.
2143    
2144    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2145        infinitely looping recursion. The bug was that positive lookaheads were not
2146        being skipped when checking for a possible empty match (negative lookaheads
2147        and both kinds of lookbehind were skipped).
2148    
2149    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2150        inclusion of <windows.h> to before rather than after the definition of
2151        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2152    
2153    17. Specifying a possessive quantifier with a specific limit for a Unicode
2154        character property caused pcre_compile() to compile bad code, which led at
2155        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2156        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2157        caused the error; without that there was no problem.
2158    
2159    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2160    
2161    19. Added --enable-pcretest-libreadline.
2162    
2163    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2164        RE::GlobalReplace(). As a result, the number of replacements returned was
2165        double what it should be. I removed one of the increments, but Craig sent a
2166        later patch that removed the other one (the right fix) and added unit tests
2167        that check the return values (which was not done before).
2168    
2169    21. Several CMake things:
2170    
2171        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2172            the names libpcre and libpcreposix, not just pcre and pcreposix.
2173    
2174        (2) The above change means that pcretest and pcregrep are now correctly
2175            linked with the newly-built libraries, not previously installed ones.
2176    
2177        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2178    
2179    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2180        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2181        UTF-8 newline character). The key issue is that the pattern starts .*;
2182        this means that the match must be either at the beginning, or after a
2183        newline. The bug was in the code for advancing after a failed match and
2184        checking that the new position followed a newline. It was not taking
2185        account of UTF-8 characters correctly.
2186    
2187    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2188        character classes. PCRE was not treating the sequence [:...:] as a
2189        character class unless the ... were all letters. Perl, however, seems to
2190        allow any characters between [: and :], though of course it rejects as
2191        unknown any "names" that contain non-letters, because all the known class
2192        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2193        for example, whereas PCRE did not - it did not recognize a POSIX character
2194        class. This seemed a bit dangerous, so the code has been changed to be
2195        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2196        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2197        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2198        Perl does, and where it didn't before.
2199    
2200    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2201        Windows environments %n is disabled by default.
2202    
2203    
2204    Version 7.4 21-Sep-07
2205    ---------------------
2206    
2207    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2208        means that a class such as [\s] counted as "explicit reference to CR or
2209        LF". That isn't really right - the whole point of the change was to try to
2210        help when there was an actual mention of one of the two characters. So now
2211        the change happens only if \r or \n (or a literal CR or LF) character is
2212        encountered.
2213    
2214    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2215        of both had grown to the point where there were only 3 bits left.
2216        Fortunately, there was spare space in the data structure, and so I have
2217        moved the internal flags into a new 16-bit field to free up more option
2218        bits.
2219    
2220    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2221        but did not set the internal JCHANGED flag - either of these is enough to
2222        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2223        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2224        start sets both bits.
2225    
2226    4.  Added options (at build time, compile time, exec time) to change \R from
2227        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2228    
2229    5.  doc/pcresyntax.html was missing from the distribution.
2230    
2231    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2232        compatibility, even though it is no longer used.
2233    
2234    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2235        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2236        windows.h file is present (where different names are used). [This was
2237        reversed later after testing - see 16 below.]
2238    
2239    8.  Changed all #include <config.h> to #include "config.h". There were also
2240        some further <pcre.h> cases that I changed to "pcre.h".
2241    
2242    9.  When pcregrep was used with the --colour option, it missed the line ending
2243        sequence off the lines that it output.
2244    
2245    10. It was pointed out to me that arrays of string pointers cause lots of
2246        relocations when a shared library is dynamically loaded. A technique of
2247        using a single long string with a table of offsets can drastically reduce
2248        these. I have refactored PCRE in four places to do this. The result is
2249        dramatic:
2250    
2251          Originally:                          290
2252          After changing UCP table:            187
2253          After changing error message table:   43
2254          After changing table of "verbs"       36
2255          After changing table of Posix names   22
2256    
2257        Thanks to the folks working on Gregex for glib for this insight.
2258    
2259    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2260        unicode-properties was also set.
2261    
2262    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2263    
2264    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2265        checked only for CRLF.
2266    
2267    14. Added casts to pcretest.c to avoid compiler warnings.
2268    
2269    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2270    
2271    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2272        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2273        entirely. This removes changes made in 7 above.
2274    
2275    17. The CMake files have been updated, and there is now more information about
2276        building with CMake in the NON-UNIX-USE document.
2277    
2278    
2279    Version 7.3 28-Aug-07
2280    ---------------------
2281    
2282     1. In the rejigging of the build system that eventually resulted in 7.1, the
2283        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2284        brackets there is not right, since it causes compilers to look for an
2285        installed pcre.h, not the version that is in the source that is being
2286        compiled (which of course may be different). I have changed it back to:
2287    
2288          #include "pcre.h"
2289    
2290        I have a vague recollection that the change was concerned with compiling in
2291        different directories, but in the new build system, that is taken care of
2292        by the VPATH setting the Makefile.
2293    
2294     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2295        when the subject happened to end in the byte 0x85 (e.g. if the last
2296        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2297        characters but of course it shouldn't be taken as a newline when it is part
2298        of another character. The bug was that, for an unlimited repeat of . in
2299        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2300        characters when looking for a newline.
2301    
2302     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2303    
2304     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2305        in debug output.
2306    
2307     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2308        long printing in the pcrecpp unittest when running under MinGW.
2309    
2310     6. ESC_K was left out of the EBCDIC table.
2311    
2312     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2313        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2314        limit also applies to "virtual nesting" when a pattern is recursive, and in
2315        this case 1000 isn't so big. I have been able to remove this limit at the
2316        expense of backing off one optimization in certain circumstances. Normally,
2317        when pcre_exec() would call its internal match() function recursively and
2318        immediately return the result unconditionally, it uses a "tail recursion"
2319        feature to save stack. However, when a subpattern that can match an empty
2320        string has an unlimited repetition quantifier, it no longer makes this
2321        optimization. That gives it a stack frame in which to save the data for
2322        checking that an empty string has been matched. Previously this was taken
2323        from the 1000-entry workspace that had been reserved. So now there is no
2324        explicit limit, but more stack is used.
2325    
2326     8. Applied Daniel's patches to solve problems with the import/export magic
2327        syntax that is required for Windows, and which was going wrong for the
2328        pcreposix and pcrecpp parts of the library. These were overlooked when this
2329        problem was solved for the main library.
2330    
2331     9. There were some crude static tests to avoid integer overflow when computing
2332        the size of patterns that contain repeated groups with explicit upper
2333        limits. As the maximum quantifier is 65535, the maximum group length was
2334        set at 30,000 so that the product of these two numbers did not overflow a
2335        32-bit integer. However, it turns out that people want to use groups that
2336        are longer than 30,000 bytes (though not repeat them that many times).
2337        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2338        made it possible to implement the integer overflow checks in a much more
2339        dynamic way, which I have now done. The artificial limitation on group
2340        length has been removed - we now have only the limit on the total length of
2341        the compiled pattern, which depends on the LINK_SIZE setting.
2342    
2343    10. Fixed a bug in the documentation for get/copy named substring when
2344        duplicate names are permitted. If none of the named substrings are set, the
2345        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2346        empty string.
2347    
2348    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2349        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2350        because the ] is interpreted as the first data character and the
2351        terminating ] is not found. PCRE has been made compatible with Perl in this
2352        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2353        cause memory overwriting.
2354    
2355    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2356        string has been matched (to stop an infinite loop). It was not recognizing
2357        a conditional subpattern that could match an empty string if that
2358        subpattern was within another subpattern. For example, it looped when
2359        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2360        condition was not nested. This bug has been fixed.
2361    
2362    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2363        past the start of the subject in the presence of bytes with the top bit
2364        set, for example "\x8aBCD".
2365    
2366    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2367        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2368    
2369    14. Optimized (?!) to (*FAIL).
2370    
2371    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2372        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2373        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2374        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2375        does: it's just the validity check that is more restrictive.
2376    
2377    16. Inserted checks for integer overflows during escape sequence (backslash)
2378        processing, and also fixed erroneous offset values for syntax errors during
2379        backslash processing.
2380    
2381    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2382        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2383    
2384    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2385        caused an overrun.
2386    
2387    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2388        something other than just ASCII characters) inside a group that had an
2389        unlimited repeat caused a loop at compile time (while checking to see
2390        whether the group could match an empty string).
2391    
2392    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2393        [\P{Any}] did so. (Error in the code for printing property names.)
2394    
2395    21. An orphan \E inside a character class could cause a crash.
2396    
2397    22. A repeated capturing bracket such as (A)? could cause a wild memory
2398        reference during compilation.
2399    
2400    23. There are several functions in pcre_compile() that scan along a compiled
2401        expression for various reasons (e.g. to see if it's fixed length for look
2402        behind). There were bugs in these functions when a repeated \p or \P was
2403        present in the pattern. These operators have additional parameters compared
2404        with \d, etc, and these were not being taken into account when moving along
2405        the compiled data. Specifically:
2406    
2407        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2408            length.
2409    
2410        (b) An item such as \pL+ within a repeated group could cause crashes or
2411            loops.
2412    
2413        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2414            "reference to non-existent subpattern" error.
2415    
2416        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2417    
2418    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2419        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2420    
2421    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2422    
2423    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2424        character were causing crashes (broken optimization).
2425    
2426    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2427        \p or \P) caused a compile-time loop.
2428    
2429    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2430        break. For example, the unstudied pattern [\r\n]A does not match the string
2431        "\r\nA" because change 7.0/46 below moves the current point on by two
2432        characters after failing to match at the start. However, the pattern \nA
2433        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2434        the same is true. There doesn't seem any very clean way out of this, but
2435        what I have chosen to do makes the common cases work: PCRE now takes note
2436        of whether there can be an explicit match for \r or \n anywhere in the
2437        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2438        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2439        pattern has explicit CR or LF references.
2440    
2441    29. Added (*CR) etc for changing newline setting at start of pattern.
2442    
2443    
2444    Version 7.2 19-Jun-07
2445    ---------------------
2446    
2447     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2448        which is apparently normally available under Windows.
2449    
2450     2. Re-jig the pcregrep tests with different newline settings in an attempt
2451        to make them independent of the local environment's newline setting.
2452    
2453     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2454    
2455     4. Some of the "internals" tests were previously cut out when the link size
2456        was not 2, because the output contained actual offsets. The recent new
2457        "Z" feature of pcretest means that these can be cut out, making the tests
2458        usable with all link sizes.
2459    
2460     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2461        stack recursion. This gives a massive performance boost under BSD, but just
2462        a small improvement under Linux. However, it saves one field in the frame
2463        in all cases.
2464    
2465     6. Added more features from the forthcoming Perl 5.10:
2466    
2467        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2468            recursion call. It refers to the nth most recently opened parentheses.
2469    
2470        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2471            to be opened parentheses.
2472    
2473        (c) Conditions that refer to capturing parentheses can be specified
2474            relatively, for example, (?(-2)... or (?(+3)...
2475    
2476        (d) \K resets the start of the current match so that everything before
2477            is not part of it.
2478    
2479        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2480    
2481        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2482            reference syntax.
2483    
2484        (g) (?| introduces a group in which the numbering of parentheses in each
2485            alternative starts with the same number.
2486    
2487        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2488    
2489     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2490        PCRE_INFO_JCHANGED.
2491    
2492     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2493        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2494        for detecting groups that can match an empty string.
2495    
2496     9. A pattern with a very large number of alternatives (more than several
2497        hundred) was running out of internal workspace during the pre-compile
2498        phase, where pcre_compile() figures out how much memory will be needed. A
2499        bit of new cunning has reduced the workspace needed for groups with
2500        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2501        workspace instead of running out of the 4096 that are available.
2502    
2503    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2504    
2505    11. Applied patch from Google to remove an optimization that didn't quite work.
2506        The report of the bug said:
2507    
2508          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2509          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2510          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2511    
2512    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2513        it matched the wrong number of bytes.
2514    
2515    
2516    Version 7.1 24-Apr-07
2517    ---------------------
2518    
2519     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2520        that is more "standard", making use of automake and other Autotools. There
2521        is some re-arrangement of the files and adjustment of comments consequent
2522        on this.
2523    
2524     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2525        for recursive directory scanning broke on some systems because the files
2526        are not scanned in any specific order and on different systems the order
2527        was different. A call to "sort" has been inserted into RunGrepTest for the
2528        approprate test as a short-term fix. In the longer term there may be an
2529        alternative.
2530    
2531     3. I had an email from Eric Raymond about problems translating some of PCRE's
2532        man pages to HTML (despite the fact that I distribute HTML pages, some
2533        people do their own conversions for various reasons). The problems
2534        concerned the use of low-level troff macros .br and .in. I have therefore
2535        removed all such uses from the man pages (some were redundant, some could
2536        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2537        HTML has been updated to handle .nf/.fi and to complain if it encounters
2538        .br or .in.
2539    
2540     4. Updated comments in configure.ac that get placed in config.h.in and also
2541        arranged for config.h to be included in the distribution, with the name
2542        config.h.generic, for the benefit of those who have to compile without
2543        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2544    
2545     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2546        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2547        makevp.bat for the current PCRE, using the additional files
2548        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2549    
2550     6. A Windows user reported a minor discrepancy with test 2, which turned out
2551        to be caused by a trailing space on an input line that had got lost in his
2552        copy. The trailing space was an accident, so I've just removed it.
2553    
2554     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2555        that is needed.
2556    
2557     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2558        as "const" (a) because they are and (b) because it helps the PHP
2559        maintainers who have recently made a script to detect big data structures
2560        in the php code that should be moved to the .rodata section. I remembered
2561        to update Builducptable as well, so it won't revert if ucptable.h is ever
2562        re-created.
2563    
2564     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2565        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2566        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2567        support is not required. This saves 1.5-2K of code, which is important in
2568        some applications.
2569    
2570        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2571        so as not to refer to the tables, even though these functions will never be
2572        called when UTF-8 support is disabled. Otherwise there are problems with a
2573        shared library.
2574    
2575    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2576    
2577        (a) It was defining its arguments as char * instead of void *.
2578    
2579        (b) It was assuming that all moves were upwards in memory; this was true
2580            a long time ago when I wrote it, but is no longer the case.
2581    
2582        The emulated memove() is provided for those environments that have neither
2583        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2584        is clearly not the case, as these two bugs were recently reported.
2585    
2586    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2587        and Detrail to create the HTML documentation, the .txt form of the man
2588        pages, and it removes trailing spaces from listed files. It also creates
2589        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2590        case, it wraps all the #defines with #ifndefs. This script should be run
2591        before "make dist".
2592    
2593    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2594        with Unicode property support.
2595    
2596        (a) For a maximizing quantifier, if the two different cases of the
2597            character were of different lengths in their UTF-8 codings (there are
2598            some cases like this - I found 11), and the matching function had to
2599            back up over a mixture of the two cases, it incorrectly assumed they
2600            were both the same length.
2601    
2602        (b) When PCRE was configured to use the heap rather than the stack for
2603            recursion during matching, it was not correctly preserving the data for
2604            the other case of a UTF-8 character when checking ahead for a match
2605            while processing a minimizing repeat. If the check also involved
2606            matching a wide character, but failed, corruption could cause an
2607            erroneous result when trying to check for a repeat of the original
2608            character.
2609    
2610    13. Some tidying changes to the testing mechanism:
2611    
2612        (a) The RunTest script now detects the internal link size and whether there
2613            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2614            values substituted by "configure". (The RunGrepTest script already did
2615            this for UTF-8.) The configure.ac script no longer substitutes the
2616            relevant variables.
2617    
2618        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2619            with length and offset values. This means that the output is different
2620            for different internal link sizes. Test 2 is skipped for link sizes
2621            other than 2 because of this, bypassing the problem. Unfortunately,
2622            there was also a test in test 3 (the locale tests) that used /B and
2623            failed for link sizes other than 2. Rather than cut the whole test out,
2624            I have added a new /Z option to pcretest that replaces the length and
2625            offset values with spaces. This is now used to make test 3 independent
2626            of link size. (Test 2 will be tidied up later.)
2627    
2628    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2629        segmentation fault instead of returning the appropriate error message.
2630    
2631    15. In multiline mode when the newline sequence was set to "any", the pattern
2632        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2633        This doesn't seem right; it now treats the CRLF combination as the line
2634        ending, and so does not match in that case. It's only a pattern such as ^$
2635        that would hit this one: something like ^ABC$ would have failed after \r
2636        and then tried again after \r\n.
2637    
2638    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2639        in an attempt to make files that differ only in their line terminators
2640        compare equal. This works on Linux.
2641    
2642    17. Under certain error circumstances pcregrep might try to free random memory
2643        as it exited. This is now fixed, thanks to valgrind.
2644    
2645    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2646        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2647        was because its rules for how to advance for /g after matching an empty
2648        string at the end of a line did not allow for this case. They now check for
2649        it specially.
2650    
2651    20. pcretest is supposed to handle patterns and data of any length, by
2652        extending its buffers when necessary. It was getting this wrong when the
2653        buffer for a data line had to be extended.
2654    
2655    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2656        CRLF as a newline sequence.
2657    
2658    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2659        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2660        I have nevertheless tidied it up.
2661    
2662    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2663    
2664    24. Added a man page for pcre-config.
2665    
2666    
2667    Version 7.0 19-Dec-06
2668    ---------------------
2669    
2670     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2671        moving to gcc 4.1.1.
2672    
2673     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2674        sys/time.h, which is documented as needed for this function. It doesn't
2675        seem to matter on Linux, but it showed up on some releases of OS X.
2676    
2677     3. It seems that there are systems where bytes whose values are greater than
2678        127 match isprint() in the "C" locale. The "C" locale should be the
2679        default when a C program starts up. In most systems, only ASCII printing
2680        characters match isprint(). This difference caused the output from pcretest
2681        to vary, making some of the tests fail. I have changed pcretest so that:
2682    
2683        (a) When it is outputting text in the compiled version of a pattern, bytes
2684            other than 32-126 are always shown as hex escapes.
2685    
2686        (b) When it is outputting text that is a matched part of a subject string,
2687            it does the same, unless a different locale has been set for the match
2688            (using the /L modifier). In this case, it uses isprint() to decide.
2689    
2690     4. Fixed a major bug that caused incorrect computation of the amount of memory
2691        required for a compiled pattern when options that changed within the
2692        pattern affected the logic of the preliminary scan that determines the
2693        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2694        that the computed length was too small. The symptoms of this bug were
2695        either the PCRE error "internal error: code overflow" from pcre_compile(),
2696        or a glibc crash with a message such as "pcretest: free(): invalid next
2697        size (fast)". Examples of patterns that provoked this bug (shown in
2698        pcretest format) are:
2699    
2700          /(?-x: )/x
2701          /(?x)(?-x: \s*#\s*)/
2702          /((?i)[\x{c0}])/8
2703          /(?i:[\x{c0}])/8
2704    
2705        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2706        is now done differently.
2707    
2708     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2709        wrapper classes; (b) implement a new function in the C++ scanner that is
2710        more efficient than the old way of doing things because it avoids levels of
2711        recursion in the regex matching; (c) add a paragraph to the documentation
2712        for the FullMatch() function.
2713    
2714     6. The escape sequence \n was being treated as whatever was defined as
2715        "newline". Not only was this contrary to the documentation, which states
2716        that \n is character 10 (hex 0A), but it also went horribly wrong when
2717        "newline" was defined as CRLF. This has been fixed.
2718    
2719     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2720        was being set to -1 for the "end of line" case (supposedly a value that no
2721        character can have). Though this value is never used (the check for end of
2722        line is "zero bytes in current character"), it caused compiler complaints.
2723        I've changed it to 0xffffffff.
2724    
2725     8. In pcre_version.c, the version string was being built by a sequence of
2726        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2727        string (as it is for production releases) called a macro with an empty
2728        argument. The C standard says the result of this is undefined. The gcc
2729        compiler treats it as an empty string (which was what was wanted) but it is
2730        reported that Visual C gives an error. The source has been hacked around to
2731        avoid this problem.
2732    
2733     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2734        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2735        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2736        of them did).
2737    
2738    10. Originally, pcretest opened its input and output without "b"; then I was
2739        told that "b" was needed in some environments, so it was added for release
2740        5.0 to both the input and output. (It makes no difference on Unix-like
2741        systems.) Later I was told that it is wrong for the input on Windows. I've
2742        now abstracted the modes into two macros, to make it easier to fiddle with
2743        them, and removed "b" from the input mode under Windows.
2744    
2745    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2746    
2747    12. Added -help and --help to pcretest as an official way of being reminded
2748        of the options.
2749    
2750    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2751        and pcrecpp.cc because they annoy compilers at high warning levels.
2752    
2753    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2754    
2755    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2756        scripts are not C programs :-) and which was not noticed because it works
2757        on Linux.
2758    
2759    16. pcretest is supposed to handle any length of pattern and data line (as one
2760        line or as a continued sequence of lines) by extending its input buffer if
2761        necessary. This feature was broken for very long pattern lines, leading to
2762        a string of junk being passed to pcre_compile() if the pattern was longer
2763        than about 50K.
2764    
2765    17. I have done a major re-factoring of the way pcre_compile() computes the
2766        amount of memory needed for a compiled pattern. Previously, there was code
2767        that made a preliminary scan of the pattern in order to do this. That was
2768        OK when PCRE was new, but as the facilities have expanded, it has become
2769        harder and harder to keep it in step with the real compile phase, and there
2770        have been a number of bugs (see for example, 4 above). I have now found a
2771        cunning way of running the real compile function in a "fake" mode that
2772        enables it to compute how much memory it would need, while actually only
2773        ever using a few hundred bytes of working memory and without too many
2774        tests of the mode. This should make future maintenance and development
2775        easier. A side effect of this work is that the limit of 200 on the nesting
2776        depth of parentheses has been removed (though this was never a serious
2777        limitation, I suspect). However, there is a downside: pcre_compile() now
2778        runs more slowly than before (30% or more, depending on the pattern). I
2779        hope this isn't a big issue. There is no effect on runtime performance.
2780    
2781    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2782        newline (only possible for the last line of a file) and it was a
2783        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2784    
2785    19. Added additional timing features to pcretest. (1) The -tm option now times
2786        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2787        separate command line item, by a number that specifies the number of
2788        repeats to use when timing. The default is 50000; this gives better
2789        precision, but takes uncomfortably long for very large patterns.
2790    
2791    20. Extended pcre_study() to be more clever in cases where a branch of a
2792        subpattern has no definite first character. For example, (a*|b*)[cd] would
2793        previously give no result from pcre_study(). Now it recognizes that the
2794        first character must be a, b, c, or d.
2795    
2796    21. There was an incorrect error "recursive call could loop indefinitely" if
2797        a subpattern (or the entire pattern) that was being tested for matching an
2798        empty string contained only one non-empty item after a nested subpattern.
2799        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2800        incorrectly, because the \d was being skipped in the check.
2801    
2802    22. The pcretest program now has a new pattern option /B and a command line
2803        option -b, which is equivalent to adding /B to every pattern. This causes
2804        it to show the compiled bytecode, without the additional information that
2805        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2806        is the same as /B/I).
2807    
2808    23. A new optimization is now able automatically to treat some sequences such
2809        as a*b as a*+b. More specifically, if something simple (such as a character
2810        or a simple class like \d) has an unlimited quantifier, and is followed by
2811        something that cannot possibly match the quantified thing, the quantifier
2812        is automatically "possessified".
2813    
2814    24. A recursive reference to a subpattern whose number was greater than 39
2815        went wrong under certain circumstances in UTF-8 mode. This bug could also
2816        have affected the operation of pcre_study().
2817    
2818    25. Realized that a little bit of performance could be had by replacing
2819        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2820    
2821    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2822    
2823    27. Possessive quantifiers such as a++ were previously implemented by turning
2824        them into atomic groups such as ($>a+). Now they have their own opcodes,
2825        which improves performance. This includes the automatically created ones
2826        from 23 above.
2827    
2828    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2829        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2830        the first matched character to be a colon. This applied both to named and
2831        numbered groups.
2832    
2833    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2834    
2835    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2836        building PCRE on Windows easier, so I have included it in the distribution.
2837    
2838    31. There is now a check in pcretest against a ridiculously large number being
2839        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2840        loop, the loop is abandoned.
2841    
2842    32. Forward references to subpatterns in conditions such as (?(2)...) where
2843        subpattern 2 is defined later cause pcre_compile() to search forwards in
2844        the pattern for the relevant set of parentheses. This search went wrong
2845        when there were unescaped parentheses in a character class, parentheses
2846        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2847    
2848    33. "Subroutine" calls and backreferences were previously restricted to
2849        referencing subpatterns earlier in the regex. This restriction has now
2850        been removed.
2851    
2852    34. Added a number of extra features that are going to be in Perl 5.10. On the
2853        whole, these are just syntactic alternatives for features that PCRE had
2854        previously implemented using the Python syntax or my own invention. The
2855        other formats are all retained for compatibility.
2856    
2857        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2858            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2859            also .NET compatible.
2860    
2861        (b) A recursion or subroutine call to a named group can now be defined as
2862            (?&name) as well as (?P>name).
2863    
2864        (c) A backreference to a named group can now be defined as \k<name> or
2865            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2866            5.10, are also .NET compatible.
2867    
2868        (d) A conditional reference to a named group can now use the syntax
2869            (?(<name>) or (?('name') as well as (?(name).
2870    
2871        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2872            groups (named and numbered) that are never evaluated inline, but can be
2873            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2874            is always false. There may be only one alternative in such a group.
2875    
2876        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2877            as the simple (?(R). The condition is true only if the most recent
2878            recursion is that of the given number or name. It does not search out
2879            through the entire recursion stack.
2880    
2881        (g) The escape \gN or \g{N} has been added, where N is a positive or
2882            negative number, specifying an absolute or relative reference.
2883    
2884    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2885        some "unreachable code" warnings.
2886    
2887    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2888        things, this adds five new scripts.
2889    
2890    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2891        There were also incompatibilities regarding the handling of \Q..\E inside
2892        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2893        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2894    
2895    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2896        matches an empty string, and forcibly breaks the loop. There were bugs in
2897        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2898        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2899        separate and independent bugs (that affected different cases) have been
2900        fixed.
2901    
2902    39. Refactored the code to abolish the use of different opcodes for small
2903        capturing bracket numbers. This is a tidy that I avoided doing when I
2904        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2905        The new approach is not only tidier, it makes it possible to reduce the
2906        memory needed to fix the previous bug (38).
2907    
2908    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2909        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2910        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2911        mode.
2912    
2913    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2914        report.
2915    
2916    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2917        copy construction and assignment in the C++ wrapper.
2918    
2919    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2920        couple of bugs that could have given wrong results in the "--newline=crlf"
2921        case.
2922    
2923    44. Added a number of casts and did some reorganization of signed/unsigned int
2924        variables following suggestions from Dair Grant. Also renamed the variable
2925        "this" as "item" because it is a C++ keyword.
2926    
2927    45. Arranged for dftables to add
2928    
2929          #include "pcre_internal.h"
2930    
2931        to pcre_chartables.c because without it, gcc 4.x may remove the array
2932        definition from the final binary if PCRE is built into a static library and
2933        dead code stripping is activated.
2934    
2935    46. For an unanchored pattern, if a match attempt fails at the start of a
2936        newline sequence, and the newline setting is CRLF or ANY, and the next two
2937        characters are CRLF, advance by two characters instead of one.
2938    
2939    
2940    Version 6.7 04-Jul-06
2941    ---------------------
2942    
2943     1. In order to handle tests when input lines are enormously long, pcretest has
2944        been re-factored so that it automatically extends its buffers when
2945        necessary. The code is crude, but this _is_ just a test program. The
2946        default size has been increased from 32K to 50K.
2947    
2948     2. The code in pcre_study() was using the value of the re argument before
2949        testing it for NULL. (Of course, in any sensible call of the function, it
2950        won't be NULL.)
2951    
2952     3. The memmove() emulation function in pcre_internal.h, which is used on
2953        systems that lack both memmove() and bcopy() - that is, hardly ever -
2954        was missing a "static" storage class specifier.
2955    
2956     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2957        containing an extended class (one that cannot be represented by a bitmap
2958        because it contains high-valued characters or Unicode property items, e.g.
2959        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2960        pattern, but PCRE should not loop if you do not (it no longer does).
2961        [Detail: two cases were found: (a) a repeated subpattern containing an
2962        extended class; (b) a recursive reference to a subpattern that followed a
2963        previous extended class. It wasn't skipping over the extended class
2964        correctly when UTF-8 mode was not set.]
2965    
2966     5. A negated single-character class was not being recognized as fixed-length
2967        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2968        compile error "lookbehind assertion is not fixed length".
2969    
2970     6. The RunPerlTest auxiliary script was showing an unexpected difference
2971        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2972        write a Perl script that can interpret lines of an input file either as
2973        byte characters or as UTF-8, which is what "perltest" was being required to
2974        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2975        can't do is switch easily at run time between having the "use utf8;" pragma
2976        or not. In the end, I fudged it by using the RunPerlTest script to insert
2977        "use utf8;" explicitly for the UTF-8 tests.
2978    
2979     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2980        the end of the subject string, contrary to the documentation and to what
2981        Perl does. This was true of both matching functions. Now it matches only at
2982        the start of the subject and immediately after *internal* newlines.
2983    
2984     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2985        a pointer to an int instead of a pointer to an unsigned long int. This
2986        caused problems on 64-bit systems.
2987    
2988     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2989        instance of the 'standard' template library not being so standard".
2990    
2991    10. There was no check on the number of named subpatterns nor the maximum
2992        length of a subpattern name. The product of these values is used to compute
2993        the size of the memory block for a compiled pattern. By supplying a very
2994        long subpattern name and a large number of named subpatterns, the size
2995        computation could be caused to overflow. This is now prevented by limiting
2996        the length of names to 32 characters, and the number of named subpatterns
2997        to 10,000.
2998    
2999    11. Subpatterns that are repeated with specific counts have to be replicated in
3000        the compiled pattern. The size of memory for this was computed from the
3001        length of the subpattern and the repeat count. The latter is limited to
3002        65535, but there was no limit on the former, meaning that integer overflow
3003        could in principle occur. The compiled length of a repeated subpattern is
3004        now limited to 30,000 bytes in order to prevent this.
3005    
3006    12. Added the optional facility to have named substrings with the same name.
3007    
3008    13. Added the ability to use a named substring as a condition, using the
3009        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3010        are numbers (not recommended). Forward references are permitted.
3011    
3012    14. Added forward references in named backreferences (if you see what I mean).
3013    
3014    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3015        pattern could run off the end of the subject. For example, the pattern
3016        "(?s)(.{1,5})"8 did this with the subject "ab".
3017    
3018    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3019        PCRE_CASELESS was set when matching characters that were quantified with ?
3020        or *.
3021    
3022    17. A character class other than a single negated character that had a minimum
3023        but no maximum quantifier - for example [ab]{6,} - was not handled
3024        correctly by pce_dfa_exec(). It would match only one character.
3025    
3026    18. A valid (though odd) pattern that looked like a POSIX character
3027        class but used an invalid character after [ (for example [[,abc,]]) caused
3028        pcre_compile() to give the error "Failed: internal error: code overflow" or
3029        in some cases to crash with a glibc free() error. This could even happen if
3030        the pattern terminated after [[ but there just happened to be a sequence of
3031        letters, a binary zero, and a closing ] in the memory that followed.
3032    
3033    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3034        over the years. Originally (before any Unicode support), just the bottom 8
3035        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3036        output from "man perlunicode" includes this:
3037    
3038          The regular expression compiler produces polymorphic opcodes.  That
3039          is, the pattern adapts to the data and automatically switches to
3040          the Unicode character scheme when presented with Unicode data--or
3041          instead uses a traditional byte scheme when presented with byte
3042          data.
3043    
3044        Sadly, a wide octal escape does not cause a switch, and in a string with
3045        no other multibyte characters, these octal escapes are treated as before.
3046        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3047        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3048        Unicode string.
3049    
3050        I have not perpetrated such confusion in PCRE. Up till now, it took just
3051        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3052        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3053        translate to the appropriate multibyte character.
3054    
3055    29. Applied some refactoring to reduce the number of warnings from Microsoft
3056        and Borland compilers. This has included removing the fudge introduced
3057        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3058        a warning about an unused variable.
3059    
3060    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3061        characters since release 4.0, because Perl (from release 5.004) does not.
3062        [Or at least, is documented not to: some releases seem to be in conflict
3063        with the documentation.] However, when a pattern was studied with
3064        pcre_study() and all its branches started with \s, PCRE still included VT
3065        as a possible starting character. Of course, this did no harm; it just
3066        caused an unnecessary match attempt.
3067    
3068    22. Removed a now-redundant internal flag bit that recorded the fact that case
3069        dependency changed within the pattern. This was once needed for "required
3070        byte" processing, but is no longer used. This recovers a now-scarce options
3071        bit. Also moved the least significant internal flag bit to the most-
3072        significant bit of the word, which was not previously used (hangover from
3073        the days when it was an int rather than a uint) to free up another bit for
3074        the future.
3075    
3076    23. Added support for CRLF line endings as well as CR and LF. As well as the
3077        default being selectable at build time, it can now be changed at runtime
3078        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3079        specify that it is scanning data with non-default line endings.
3080    
3081    24. Changed the definition of CXXLINK to make it agree with the definition of
3082        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3083    
3084    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3085        recursions. This makes a big different to stack usage for some patterns.
3086    
3087    26. If a subpattern containing a named recursion or subroutine reference such
3088        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3089        the space required for the compiled pattern went wrong and gave too small a
3090        value. Depending on the environment, this could lead to "Failed: internal
3091        error: code overflow at offset 49" or "glibc detected double free or
3092        corruption" errors.
3093    
3094    27. Applied patches from Google (a) to support the new newline modes and (b) to
3095        advance over multibyte UTF-8 characters in GlobalReplace.
3096    
3097    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3098        difference for some implementation of PCRE in some Windows version.
3099    
3100    29. Added some extra testing facilities to pcretest:
3101    
3102        \q<number>   in a data line sets the "match limit" value
3103        \Q<number>   in a data line sets the "match recursion limt" value
3104        -S <number>  sets the stack size, where <number> is in megabytes
3105    
3106        The -S option isn't available for Windows.
3107    
3108    
3109    Version 6.6 06-Feb-06
3110    ---------------------
3111    
3112     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3113        in pcreposix.h. I have copied the definition from pcre.h.
3114    
3115     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3116        because pcre.h is no longer a built file.
3117    
3118     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3119        not normally included in the compiled code.
3120    
3121    
3122    Version 6.5 01-Feb-06
3123    ---------------------
3124    
3125     1. When using the partial match feature with pcre_dfa_exec(), it was not
3126        anchoring the second and subsequent partial matches at the new starting
3127        point. This could lead to incorrect results. For example, with the pattern
3128        /1234/, partially matching against "123" and then "a4" gave a match.
3129    
3130     2. Changes to pcregrep:
3131    
3132        (a) All non-match returns from pcre_exec() were being treated as failures
3133            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3134            error message is output. Some extra information is given for the
3135            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3136            probably the only errors that are likely to be caused by users (by
3137            specifying a regex that has nested indefinite repeats, for instance).
3138            If there are more than 20 of these errors, pcregrep is abandoned.
3139    
3140        (b) A binary zero was treated as data while matching, but terminated the
3141            output line if it was written out. This has been fixed: binary zeroes
3142            are now no different to any other data bytes.
3143    
3144        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3145            used to set a locale for matching. The --locale=xxxx long option has
3146            been added (no short equivalent) to specify a locale explicitly on the
3147            pcregrep command, overriding the environment variables.
3148    
3149        (d) When -B was used with -n, some line numbers in the output were one less
3150            than they should have been.
3151    
3152        (e) Added the -o (--only-matching) option.
3153    
3154        (f) If -A or -C was used with -c (count only), some lines of context were
3155            accidentally printed for the final match.
3156    
3157        (g) Added the -H (--with-filename) option.
3158    
3159        (h) The combination of options -rh failed to suppress file names for files
3160            that were found from directory arguments.
3161    
3162        (i) Added the -D (--devices) and -d (--directories) options.
3163    
3164        (j) Added the -F (--fixed-strings) option.
3165    
3166        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3167    
3168        (l) Added the --colo(u)r option.
3169    
3170        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3171            is not present by default.
3172    
3173     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3174        items such as (?R) or (?1), when the recursion could match a number of
3175        alternatives. If it matched one of the alternatives, but subsequently,
3176        outside the recursion, there was a failure, the code tried to back up into
3177        the recursion. However, because of the way PCRE is implemented, this is not
3178        possible, and the result was an incorrect result from the match.
3179    
3180        In order to prevent this happening, the specification of recursion has
3181        been changed so that all such subpatterns are automatically treated as
3182        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3183    
3184     4. I had overlooked the fact that, in some locales, there are characters for
3185        which isalpha() is true but neither isupper() nor islower() are true. In
3186        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3187        and ordfeminine) are like this. This affected the treatment of \w and \W
3188        when they appeared in character classes, but not when they appeared outside
3189        a character class. The bit map for "word" characters is now created
3190        separately from the results of isalnum() instead of just taking it from the
3191        upper, lower, and digit maps. (Plus the underscore character, of course.)
3192    
3193     5. The above bug also affected the handling of POSIX character classes such as
3194        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3195        permanent tables. Instead, the bit maps for such a class were previously
3196        created as the appropriate unions of the upper, lower, and digit bitmaps.
3197        Now they are created by subtraction from the [[:word:]] class, which has
3198        its own bitmap.
3199    
3200     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3201        It is created by subtracting the vertical space characters (\x09, \x0a,
3202        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3203        subtraction was done in the overall bitmap for a character class, meaning
3204        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3205        be recognized. This bug has been fixed.
3206    
3207     7. Patches from the folks at Google:
3208    
3209          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3210          real life, but is still worth protecting against".
3211    
3212          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3213          regular expressions".
3214    
3215          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3216          have it.
3217    
3218          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3219          "configure" and the latter not, in order to fix a problem somebody had
3220          with compiling the Arg class on HP-UX.
3221    
3222          (e) Improve the error-handling of the C++ wrapper a little bit.
3223    
3224          (f) New tests for checking recursion limiting.
3225    
3226     8. The pcre_memmove() function, which is used only if the environment does not
3227        have a standard memmove() function (and is therefore rarely compiled),
3228        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3229        returning a result (though PCRE never actually uses the result).
3230    
3231     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3232        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3233        returned instead of calling malloc() with an overflowing number that would
3234        most likely cause subsequent chaos.
3235    
3236    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3237    
3238    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3239        with this option is matched, the nmatch and pmatch options of regexec() are
3240        ignored.
3241    
3242    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3243        provided in case anyone wants to the the POSIX interface with UTF-8
3244        strings.
3245    
3246    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3247        C++ linking (needed for some HP-UX environments).
3248    
3249    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3250        (unused parameter) and in the pcre_printint() function (omitted "default"
3251        switch label when the default is to do nothing).
3252    
3253    15. Added some code to make it possible, when PCRE is compiled as a C++
3254        library, to replace subject pointers for pcre_exec() with a smart pointer
3255        class, thus making it possible to process discontinuous strings.
3256    
3257    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3258        much the same function. They were added by different people who were trying
3259        to make PCRE easy to compile on non-Unix systems. It has been suggested
3260        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3261        for compiling on Windows systems. I have therefore replaced it with
3262        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3263        defaults to "extern" for C or "extern C" for C++, which works fine on
3264        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3265        SCOPE with something explicit in config.h. In addition:
3266    
3267        (a) pcreposix.h still had just "extern" instead of either of these macros;
3268            I have replaced it with PCRE_DATA_SCOPE.
3269    
3270        (b) Functions such as _pcre_xclass(), which are internal to the library,
3271            but external in the C sense, all had PCRE_EXPORT in their definitions.
3272            This is apparently wrong for the Windows case, so I have removed it.
3273            (It makes no difference on Unix-like systems.)
3274    
3275    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3276        of recursive calls to match(). This is different to MATCH_LIMIT because
3277        that limits the total number of calls to match(), not all of which increase
3278        the depth of recursion. Limiting the recursion depth limits the amount of
3279        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3280        when PCRE is compiled, and changed at run time. A patch from Google adds
3281        this functionality to the C++ interface.
3282    
3283    18. Changes to the handling of Unicode character properties:
3284    
3285        (a) Updated the table to Unicode 4.1.0.
3286    
3287        (b) Recognize characters that are not in the table as "Cn" (undefined).
3288    
3289        (c) I revised the way the table is implemented to a much improved format
3290            which includes recognition of ranges. It now supports the ranges that
3291            are defined in UnicodeData.txt, and it also amalgamates other
3292            characters into ranges. This has reduced the number of entries in the
3293            table from around 16,000 to around 3,000, thus reducing its size
3294            considerably. I realized I did not need to use a tree structure after
3295            all - a binary chop search is just as efficient. Having reduced the
3296            number of entries, I extended their size from 6 bytes to 8 bytes to
3297            allow for more data.
3298    
3299        (d) Added support for Unicode script names via properties such as \p{Han}.
3300    
3301    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3302        matching that character.
3303    
3304    20. When matching a repeated Unicode property with a minimum greater than zero,
3305        (for example \pL{2,}), PCRE could look past the end of the subject if it
3306        reached it while seeking the minimum number of characters. This could
3307        happen only if some of the characters were more than one byte long, because
3308        there is a check for at least the minimum number of bytes.
3309    
3310    21. Refactored the implementation of \p and \P so as to be more general, to
3311        allow for more different types of property in future. This has changed the
3312        compiled form incompatibly. Anybody with saved compiled patterns that use
3313        \p or \P will have to recompile them.
3314    
3315    22. Added "Any" and "L&" to the supported property types.
3316    
3317    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3318        but give a compile time error if the value is greater than 0xff.
3319    
3320    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3321        accidentally not being installed or uninstalled.
3322    
3323    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3324        made were to insert the current release number. This seemed silly, because
3325        it made things harder for people building PCRE on systems that don't run
3326        "configure". I have turned pcre.h into a distributed file, no longer built
3327        by "configure", with the version identification directly included. There is
3328        no longer a pcre.h.in file.
3329    
3330        However, this change necessitated a change to the pcre-config script as
3331        well. It is built from pcre-config.in, and one of the substitutions was the
3332        release number. I have updated configure.ac so that ./configure now finds
3333        the release number by grepping pcre.h.
3334    
3335    26. Added the ability to run the tests under valgrind.
3336    
3337    
3338    Version 6.4 05-Sep-05
3339    ---------------------
3340    
3341     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3342        "--" to be printed when multiple files were scanned, even when none of the
3343        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3344        consider it to be a bug, and have restored the previous behaviour.
3345    
3346     2. A couple of code tidies to get rid of compiler warnings.
3347    
3348     3. The pcretest program used to cheat by referring to symbols in the library
3349        whose names begin with _pcre_. These are internal symbols that are not
3350        really supposed to be visible externally, and in some environments it is
3351        possible to suppress them. The cheating is now confined to including
3352        certain files from the library's source, which is a bit cleaner.
3353    
3354     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3355        file's purpose clearer.
3356    
3357     5. Reorganized pcre_ucp_findchar().
3358    
3359    
3360    Version 6.3 15-Aug-05
3361    ---------------------
3362    
3363     1. The file libpcre.pc.in did not have general read permission in the tarball.
3364    
3365     2. There were some problems when building without C++ support:
3366    
3367        (a) If C++ support was not built, "make install" and "make test" still
3368            tried to test it.
3369    
3370        (b) There were problems when the value of CXX was explicitly set. Some
3371            changes have been made to try to fix these, and ...
3372    
3373        (c) --disable-cpp can now be used to explicitly disable C++ support.
3374    
3375        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3376            backslash in a target when C++ was disabled. This confuses some
3377            versions of "make", apparently. Using an intermediate variable solves
3378            this. (Same for CPP_LOBJ.)
3379    
3380     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3381        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3382        necessary on certain architectures.
3383    
3384     4. Added a setting of -export-symbols-regex to the link command to remove
3385        those symbols that are exported in the C sense, but actually are local
3386        within the library, and not documented. Their names all begin with
3387        "_pcre_". This is not a perfect job, because (a) we have to except some
3388        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3389        available (and never for static libraries). I have made a note to try to
3390        find a way round (a) in the future.
3391    
3392    
3393    Version 6.2 01-Aug-05
3394    ---------------------
3395    
3396     1. There was no test for integer overflow of quantifier values. A construction
3397        such as {1111111111111111} would give undefined results. What is worse, if
3398        a minimum quantifier for a parenthesized subpattern overflowed and became
3399        negative, the calculation of the memory size went wrong. This could have
3400        led to memory overwriting.
3401    
3402     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3403    
3404     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3405        operating environments where this matters.
3406    
3407     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3408        PCRE options from within the C++ wrapper.
3409    
3410     5. Named capturing subpatterns were not being correctly counted when a pattern
3411        was compiled. This caused two problems: (a) If there were more than 100
3412        such subpatterns, the calculation of the memory needed for the whole
3413        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3414        back references of the form \12, where the number was greater than 9, were
3415        not recognized as back references, even though there were sufficient
3416        previous subpatterns.
3417    
3418     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3419        versions of gcc, e.g. 2.95.4.
3420    
3421    
3422    Version 6.1 21-Jun-05
3423    ---------------------
3424    
3425     1. There was one reference to the variable "posix" in pcretest.c that was not
3426        surrounded by "#if !defined NOPOSIX".
3427    
3428     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3429        the cross-check on the old pcre_info() function, for the benefit of the
3430        cut-down version of PCRE that is currently imported into Exim.
3431    
3432     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3433        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3434        patterns that start (?i) but I don't think that matters. The use of (?i) is
3435        just an example; this all applies to the other options as well.
3436    
3437     4. Since libtool seems to echo the compile commands it is issuing, the output
3438        from "make" can be reduced a bit by putting "@" in front of each libtool
3439        compile command.
3440    
3441     5. Patch from the folks at Google for configure.in to be a bit more thorough
3442        in checking for a suitable C++ installation before trying to compile the
3443        C++ stuff. This should fix a reported problem when a compiler was present,
3444        but no suitable headers.
3445    
3446     6. The man pages all had just "PCRE" as their title. I have changed them to
3447        be the relevant file name. I have also arranged that these names are
3448        retained in the file doc/pcre.txt, which is a concatenation in text format
3449        of all the man pages except the little individual ones for each function.
3450    
3451     7. The NON-UNIX-USE file had not been updated for the different set of source
3452        files that come with release 6. I also added a few comments about the C++
3453        wrapper.
3454    
3455    
3456    Version 6.0 07-Jun-05
3457    ---------------------
3458    
3459     1. Some minor internal re-organization to help with my DFA experiments.
3460    
3461     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3462        didn't matter for the library itself when fully configured, but did matter
3463        when compiling without UCP support, or within Exim, where the ucp files are
3464        not imported.
3465    
3466     3. Refactoring of the library code to split up the various functions into
3467        different source modules. The addition of the new DFA matching code (see
3468        below) to a single monolithic source would have made it really too
3469        unwieldy, quite apart from causing all the code to be include in a
3470        statically linked application, when only some functions are used. This is
3471        relevant even without the DFA addition now that patterns can be compiled in
3472        one application and matched in another.
3473    
3474        The downside of splitting up is that there have to be some external
3475        functions and data tables that are used internally in different modules of
3476        the library but which are not part of the API. These have all had their
3477        names changed to start with "_pcre_" so that they are unlikely to clash
3478        with other external names.
3479    
3480     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3481        a different (DFA) algorithm. Although it is slower than the original
3482        function, it does have some advantages for certain types of matching
3483        problem.
3484    
3485     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3486        including restarting after a partial match.
3487    
3488     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3489        defined when compiling for Windows was sent to me. I have put it into the
3490        code, though I have no means of testing or verifying it.
3491    
3492     7. Added the pcre_refcount() auxiliary function.
3493    
3494     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3495        match before or at the first newline in the subject string. In pcretest,
3496        the /f option on a pattern can be used to set this.
3497    
3498     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3499        would behave wrongly. This has been present in PCRE since release 4.0.
3500    
3501    10. A number of changes to the pcregrep command:
3502    
3503        (a) Refactored how -x works; insert ^(...)$ instead of setting
3504            PCRE_ANCHORED and checking the length, in preparation for adding
3505            something similar for -w.
3506    
3507        (b) Added the -w (match as a word) option.
3508    
3509        (c) Refactored the way lines are read and buffered so as to have more
3510            than one at a time available.
3511    
3512        (d) Implemented a pcregrep test script.
3513    
3514        (e) Added the -M (multiline match) option. This allows patterns to match
3515            over several lines of the subject. The buffering ensures that at least
3516            8K, or the rest of the document (whichever is the shorter) is available
3517            for matching (and similarly the previous 8K for lookbehind assertions).
3518    
3519        (f) Changed the --help output so that it now says
3520    
3521              -w, --word-regex(p)
3522    
3523            instead of two lines, one with "regex" and the other with "regexp"
3524            because that confused at least one person since the short forms are the
3525            same. (This required a bit of code, as the output is generated
3526            automatically from a table. It wasn't just a text change.)
3527    
3528        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3529            option but starts with a hyphen. Could be a pattern or a path name
3530            starting with a hyphen, for instance.
3531    
3532        (h) "-" can be given as a file name to represent stdin.
3533    
3534        (i) When file names are being printed, "(standard input)" is used for
3535            the standard input, for compatibility with GNU grep. Previously
3536            "<stdin>" was used.
3537    
3538        (j) The option --label=xxx can be used to supply a name to be used for
3539            stdin when file names are being printed. There is no short form.
3540    
3541        (k) Re-factored the options decoding logic because we are going to add
3542            two more options that take data. Such options can now be given in four
3543            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3544    
3545        (l) Added the -A, -B, and -C options for requesting that lines of context
3546            around matches be printed.
3547    
3548        (m) Added the -L option to print the names of files that do not contain
3549            any matching lines, that is, the complement of -l.
3550    
3551        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3552            continue to scan other files.
3553    
3554        (o) The -s option was incorrectly implemented. For compatibility with other
3555            greps, it now suppresses the error message for a non-existent or non-
3556            accessible file (but not the return code). There is a new option called
3557            -q that suppresses the output of matching lines, which was what -s was
3558            previously doing.
3559    
3560        (p) Added --include and --exclude options to specify files for inclusion
3561            and exclusion when recursing.
3562    
3563    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3564        Hopefully, it now does.
3565    
3566    12. Missing cast in pcre_study().
3567    
3568    13. Added an "uninstall" target to the makefile.
3569    
3570    14. Replaced "extern" in the function prototypes in Makefile.in with
3571        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3572        world, but is set differently for Windows.
3573    
3574    15. Added a second compiling function called pcre_compile2(). The only
3575        difference is that it has an extra argument, which is a pointer to an
3576        integer error code. When there is a compile-time failure, this is set
3577        non-zero, in addition to the error test pointer being set to point to an
3578        error message. The new argument may be NULL if no error number is required
3579        (but then you may as well call pcre_compile(), which is now just a
3580        wrapper). This facility is provided because some applications need a
3581        numeric error indication, but it has also enabled me to tidy up the way
3582        compile-time errors are handled in the POSIX wrapper.
3583    
3584    16. Added VPATH=.libs to the makefile; this should help when building with one
3585        prefix path and installing with another. (Or so I'm told by someone who
3586        knows more about this stuff than I do.)
3587    
3588    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3589        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3590        match everything, including newlines. This is not POSIX-compatible, but
3591        somebody wanted the feature. From pcretest it can be activated by using
3592        both the P and the s flags.
3593    
3594    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3595    
3596    19. libpcre.pc was being incorrectly installed as executable.
3597    
3598    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3599        it now also looks for '\r' so that it will work unmodified on Windows.
3600    
3601    21. Added Google's contributed C++ wrapper to the distribution.
3602    
3603    22. Added some untidy missing memory free() calls in pcretest, to keep
3604        Electric Fence happy when testing.
3605    
3606    
3607    
3608    Version 5.0 13-Sep-04
3609    ---------------------
3610    
3611     1. Internal change: literal characters are no longer packed up into items
3612        containing multiple characters in a single byte-string. Each character
3613        is now matched using a separate opcode. However, there may be more than one
3614        byte in the character in UTF-8 mode.
3615    
3616     2. The pcre_callout_block structure has two new fields: pattern_position and
3617        next_item_length. These contain the offset in the pattern to the next match
3618        item, and its length, respectively.
3619    
3620     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3621        insertion of callouts before each pattern item. Added the /C option to
3622        pcretest to make use of this.
3623    
3624     4. On the advice of a Windows user, the lines
3625    
3626          #if defined(_WIN32) || defined(WIN32)
3627          _setmode( _fileno( stdout ), 0x8000 );
3628          #endif  /* defined(_WIN32) || defined(WIN32) */
3629    
3630        have been added to the source of pcretest. This apparently does useful
3631        magic in relation to line terminators.
3632    
3633     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3634        for the benefit of those environments where the "b" makes a difference.
3635    
3636     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3637        to know about it. I have put a hack into configure.in that adds in code
3638        to set GCC=yes if CC=icc. This seems to end up at a point in the
3639        generated configure script that is early enough to affect the setting of
3640        compiler options, which is what is needed, but I have no means of testing
3641        whether it really works. (The user who reported this had patched the
3642        generated configure script, which of course I cannot do.)
3643    
3644        LATER: After change 22 below (new libtool files), the configure script
3645        seems to know about icc (and also ecc). Therefore, I have commented out
3646        this hack in configure.in.
3647    
3648     7. Added support for pkg-config (2 patches were sent in).
3649    
3650     8. Negated POSIX character classes that used a combination of internal tables
3651        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3652        [[:^ascii]]. Typically, they would match almost any characters. The other
3653        POSIX classes were not broken in this way.
3654    
3655     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3656        to find the match, as PCRE was deluded into thinking that the match had to
3657        start at the start point or following a newline. The same bug applied to
3658        patterns with negative forward assertions or any backward assertions
3659        preceding ".*" at the start, unless the pattern required a fixed first
3660        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3661    
3662    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3663        starting at the last subject character, bytes beyond the end of the subject
3664        string were read.
3665    
3666    11. Renamed the variable "class" as "classbits" to make life easier for C++
3667        users. (Previously there was a macro definition, but it apparently wasn't
3668        enough.)
3669    
3670    12. Added the new field "tables" to the extra data so that tables can be passed
3671        in at exec time, or the internal tables can be re-selected. This allows
3672        a compiled regex to be saved and re-used at a later time by a different
3673        program that might have everything at different addresses.
3674    
3675    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3676        -R library as well as a -L library.
3677    
3678    14. The debugging options of pcretest (-d on the command line or D on a
3679        pattern) showed incorrect output for anything following an extended class
3680        that contained multibyte characters and which was followed by a quantifier.
3681    
3682    15. Added optional support for general category Unicode character properties
3683        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3684        support. It adds about 90K to the size of the library. The meanings of the
3685        inbuilt class escapes such as \d and \s have NOT been changed.
3686    
3687    16. Updated pcredemo.c to include calls to free() to release the memory for the
3688        compiled pattern.
3689    
3690    17. The generated file chartables.c was being created in the source directory
3691        instead of in the building directory. This caused the build to fail if the
3692        source directory was different from the building directory, and was
3693        read-only.
3694    
3695    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3696        file. No doubt somebody will tell me if they don't make sense... Also added
3697        Dan Mooney's comments about building on OpenVMS.
3698    
3699    19. Added support for partial matching via the PCRE_PARTIAL option for
3700        pcre_exec() and the \P data escape in pcretest.
3701    
3702    20. Extended pcretest with 3 new pattern features:
3703    
3704        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3705              write the compiled pattern to the file whose name is "rest-of-line".
3706              This is a straight binary dump of the data, with the saved pointer to
3707              the character tables forced to be NULL. The study data, if any, is
3708              written too. After writing, pcretest reads a new pattern.
3709    
3710        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3711              compiled pattern from the given file. There must not be any
3712              occurrences of "<" in the file name (pretty unlikely); if there are,
3713              pcretest will instead treat the initial "<" as a pattern delimiter.
3714              After reading in the pattern, pcretest goes on to read data lines as
3715              usual.
3716    
3717        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3718              and 16-bit fields in a compiled pattern, to simulate a pattern that
3719              was compiled on a host of opposite endianness.
3720    
3721    21. The pcre-exec() function can now cope with patterns that were compiled on
3722        hosts of opposite endianness, with this restriction:
3723    
3724          As for any compiled expression that is saved and used later, the tables
3725          pointer field cannot be preserved; the extra_data field in the arguments
3726          to pcre_exec() should be used to pass in a tables address if a value
3727          other than the default internal tables were used at compile time.
3728    
3729    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3730        now diagnosed as an error. Previously, most of the time, a negative number
3731        would have been treated as zero, but if in addition "ovector" was passed as
3732        NULL, a crash could occur.
3733    
3734    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3735        new versions from the libtool 1.5 distribution (the last one is a copy of
3736        a file called libtool.m4). This seems to have fixed the need to patch
3737        "configure" to support Darwin 1.3 (which I used to do). However, I still
3738        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3739        workstation).
3740    
3741    24. Changed the PCRE licence to be the more standard "BSD" licence.
3742    
3743    
3744    Version 4.5 01-Dec-03
3745    ---------------------
3746    
3747     1. There has been some re-arrangement of the code for the match() function so
3748        that it can be compiled in a version that does not call itself recursively.
3749        Instead, it keeps those local variables that need separate instances for
3750        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3751        needs to "recurse". Keeping track of where control must go is done by means
3752        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3753        hide most of the details from the main code, and operates only if
3754        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3755        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3756        operating.
3757    
3758        To make it easier for callers to provide specially tailored get/free
3759        functions for this usage, two new functions, pcre_stack_malloc, and
3760        pcre_stack_free, are used. They are always called in strict stacking order,
3761        and the size of block requested is always the same.
3762    
3763        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3764        PCRE has been compiled to use the stack or the heap for recursion. The
3765        -C option of pcretest uses this to show which version is compiled.
3766    
3767        A new data escape \S, is added to pcretest; it causes the amounts of store
3768        obtained and freed by both kinds of malloc/free at match time to be added
3769        to the output.
3770    
3771     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3772        what's available on my current Linux desktop machine.
3773    
3774     3. When matching a UTF-8 string, the test for a valid string at the start has
3775        been extended. If start_offset is not zero, PCRE now checks that it points
3776        to a byte that is the start of a UTF-8 character. If not, it returns
3777        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3778        this is necessary because there may be backward assertions in the pattern.
3779        When matching the same subject several times, it may save resources to use
3780        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3781    
3782     4. The code for checking the validity of UTF-8 strings has been tightened so
3783        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3784        containing "overlong sequences".
3785    
3786     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3787        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3788        should have been "&", but it just so happened that all the cases this let