/[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 1266 by ph10, Sun Mar 3 11:14:26 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.33 xx-xxxx-201x
5    -------------------------
6    
7    1.  Added 'U' to some constants that are compared to unsigned integers, to
8        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9        variables that are added to signed variables, to ensure the result is
10        signed and can be negated.
11    
12    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
13        CMake config files.
14    
15    3.  Revise the creation of config.h.generic so that all boolean macros are
16        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17        overriding via -D on the command line possible.
18    
19    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20        to unsigned int is reported to make a quite noticeable speed difference in
21        a specific Windows environment. Testing on Linux did also appear to show
22        some benefit (and it is clearly not harmful). Also fixed the definition of
23        Xop which should be unsigned.
24    
25    5.  Related to (4), changing the definition of the intermediate variable cc
26        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27        improvements.
28    
29    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
30        unnecessary spaces.
31    
32    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33        and later.
34    
35    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36    
37    9.  Optimizing fast_forward_start_bits in JIT.
38    
39    10. Adding experimental support for callouts in JIT, and fixing some
40        issues revealed during this work. Namely:
41    
42        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43    
44        (b) Minimum length was not checked before the matching is started.
45    
46    11. The value of capture_last that is passed to callouts was incorrect in some
47        cases when there was a capture on one path that was subsequently abandoned
48        after a backtrack. Also, the capture_last value is now reset after a
49        recursion, since all captures are also reset in this case.
50    
51    12. The interpreter no longer returns the "too many substrings" error in the
52        case when an overflowing capture is in a branch that is subsequently
53        abandoned after a backtrack.
54    
55    13. In the pathological case when an offset vector of size 2 is used, pcretest
56        now prints out the matched string after a yield of 0 or 1.
57    
58    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59        Only supported by the JIT compiler at the moment.
60    
61    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62    
63    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64        offset of the starting point of the matching process, provided the offsets
65        vector is large enough.
66    
67    17. The \A escape now records a lookbehind value of 1, though its execution
68        does not actually inspect the previous character. This is to ensure that,
69        in partial multi-segment matching, at least one character from the old
70        segment is retained when a new segment is processed. Otherwise, if there
71        are no lookbehinds in the pattern, \A might match incorrectly at the start
72        of a new segment.
73    
74    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75    
76    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77        gives some modest performance improvement in 8-bit mode.
78    
79    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80        be expressed in certain programming languages using Universal Character
81        Names.
82    
83    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84        which points out that "non characters" are not "characters that may not
85        appear in Unicode strings" but rather "characters that are reserved for
86        internal use and have only local meaning".
87    
88    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89        there was a conditional group that depended on an assertion, if the
90        assertion was false, the callout that immediately followed the alternation
91        in the condition was skipped when pcre_exec() was used for matching.
92    
93    23. Allow an explicit callout to be inserted before an assertion that is the
94        condition for a conditional group, for compatibility with automatic
95        callouts, which always insert a callout at this point.
96    
97    
98    Version 8.32 30-November-2012
99    -----------------------------
100    
101    1.  Improved JIT compiler optimizations for first character search and single
102        character iterators.
103    
104    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
105        Patch by Daniel Richard G.
106    
107    3.  Single character iterator optimizations in the JIT compiler.
108    
109    4.  Improved JIT compiler optimizations for character ranges.
110    
111    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
112        Reported by Giuseppe D'Angelo.
113    
114    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
115        of a line, was being set incorrectly in cases where .* appeared inside
116        atomic brackets at the start of a pattern, or where there was a subsequent
117        *PRUNE or *SKIP.
118    
119    7.  Improved instruction cache flush for POWER/PowerPC.
120        Patch by Daniel Richard G.
121    
122    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
123        grep:
124    
125        (a) There is now no limit to the number of patterns to be matched.
126    
127        (b) An error is given if a pattern is too long.
128    
129        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
130            are now supported.
131    
132        (d) --exclude-from and --include-from (multiple use) have been added.
133    
134        (e) Exclusions and inclusions now apply to all files and directories, not
135            just to those obtained from scanning a directory recursively.
136    
137        (f) Multiple uses of -f and --file-list are now supported.
138    
139        (g) In a Windows environment, the default for -d has been changed from
140            "read" (the GNU grep default) to "skip", because otherwise the presence
141            of a directory in the file list provokes an error.
142    
143        (h) The documentation has been revised and clarified in places.
144    
145    9.  Improve the matching speed of capturing brackets.
146    
147    10. Changed the meaning of \X so that it now matches a Unicode extended
148        grapheme cluster.
149    
150    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
151        out POSIX threads when JIT support is configured.
152    
153    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
154    
155    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
156        structure could go wrong in environments where size_t is not the same size
157        as int.
158    
159    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
160        set.
161    
162    15. The EBCDIC support had decayed; later updates to the code had included
163        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
164        general tidy up of EBCDIC-related issues, and the documentation was also
165        not quite right. There is now a test that can be run on ASCII systems to
166        check some of the EBCDIC-related things (but is it not a full test).
167    
168    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
169        in a small tidy to the code.
170    
171    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
172    
173    18. If the --only-matching (-o) option in pcregrep is specified multiple
174        times, each one causes appropriate output. For example, -o1 -o2 outputs the
175        substrings matched by the 1st and 2nd capturing parentheses. A separating
176        string can be specified by --om-separator (default empty).
177    
178    19. Improving the first n character searches.
179    
180    20. Turn case lists for horizontal and vertical white space into macros so that
181        they are defined only once.
182    
183    21. This set of changes together give more compatible Unicode case-folding
184        behaviour for characters that have more than one other case when UCP
185        support is available.
186    
187        (a) The Unicode property table now has offsets into a new table of sets of
188            three or more characters that are case-equivalent. The MultiStage2.py
189            script that generates these tables (the pcre_ucd.c file) now scans
190            CaseFolding.txt instead of UnicodeData.txt for character case
191            information.
192    
193        (b) The code for adding characters or ranges of characters to a character
194            class has been abstracted into a generalized function that also handles
195            case-independence. In UTF-mode with UCP support, this uses the new data
196            to handle characters with more than one other case.
197    
198        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
199            whose other case is greater than 256 are now correctly matched
200            caselessly. Previously, the high codepoint matched the low one, but not
201            vice versa.
202    
203        (d) The processing of \h, \H, \v, and \ in character classes now makes use
204            of the new class addition function, using character lists defined as
205            macros alongside the case definitions of 20 above.
206    
207        (e) Caseless back references now work with characters that have more than
208            one other case.
209    
210        (f) General caseless matching of characters with more than one other case
211            is supported.
212    
213    22. Unicode character properties were updated from Unicode 6.2.0
214    
215    23. Improved CMake support under Windows. Patch by Daniel Richard G.
216    
217    24. Add support for 32-bit character strings, and UTF-32
218    
219    25. Major JIT compiler update (code refactoring and bugfixing).
220        Experimental Sparc 32 support is added.
221    
222    26. Applied a modified version of Daniel Richard G's patch to create
223        pcre.h.generic and config.h.generic by "make" instead of in the
224        PrepareRelease script.
225    
226    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
227        pcre_compile.c when checking for a zero character.
228    
229    28. Introducing a native interface for JIT. Through this interface, the compiled
230        machine code can be directly executed. The purpose of this interface is to
231        provide fast pattern matching, so several sanity checks are not performed.
232        However, feature tests are still performed. The new interface provides
233        1.4x speedup compared to the old one.
234    
235    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
236        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
237        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
238    
239    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
240        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
241        the "old" RFC 2279). Instead, it ended up passing a negative length to
242        pcre_exec().
243    
244    31. Add support for GCC's visibility feature to hide internal functions.
245    
246    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
247        "unknown -C option" after outputting 0 or 1.
248    
249    33. There is now support for generating a code coverage report for the test
250        suite in environments where gcc is the compiler and lcov is installed. This
251        is mainly for the benefit of the developers.
252    
253    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
254        unaddressable using valgrind annotations, allowing valgrind to detect
255        invalid memory accesses. This is mainly for the benefit of the developers.
256    
257    25. (*UTF) can now be used to start a pattern in any of the three libraries.
258    
259    26. Give configure error if --enable-cpp but no C++ compiler found.
260    
261    
262    Version 8.31 06-July-2012
263    -------------------------
264    
265    1.  Fixing a wrong JIT test case and some compiler warnings.
266    
267    2.  Removed a bashism from the RunTest script.
268    
269    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
270        to unsigned type, result still unsigned" that was given by an MS compiler
271        on encountering the code "-sizeof(xxx)".
272    
273    4.  Partial matching support is added to the JIT compiler.
274    
275    5.  Fixed several bugs concerned with partial matching of items that consist
276        of more than one character:
277    
278        (a) /^(..)\1/ did not partially match "aba" because checking references was
279            done on an "all or nothing" basis. This also applied to repeated
280            references.
281    
282        (b) \R did not give a hard partial match if \r was found at the end of the
283            subject.
284    
285        (c) \X did not give a hard partial match after matching one or more
286            characters at the end of the subject.
287    
288        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
289            a partial match for the string "\r".
290    
291        (e) When newline was set to CRLF, the metacharacter "." did not recognize
292            a partial match for a CR character at the end of the subject string.
293    
294    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
295        running pcretest, the text "(JIT)" added to the output whenever JIT is
296        actually used to run the match.
297    
298    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
299        or /S+[+] with a digit between 1 and 7.
300    
301    8.  OP_NOT now supports any UTF character not just single-byte ones.
302    
303    9.  (*MARK) control verb is now supported by the JIT compiler.
304    
305    10. The command "./RunTest list" lists the available tests without actually
306        running any of them. (Because I keep forgetting what they all are.)
307    
308    11. Add PCRE_INFO_MAXLOOKBEHIND.
309    
310    12. Applied a (slightly modified) user-supplied patch that improves performance
311        when the heap is used for recursion (compiled with --disable-stack-for-
312        recursion). Instead of malloc and free for each heap frame each time a
313        logical recursion happens, frames are retained on a chain and re-used where
314        possible. This sometimes gives as much as 30% improvement.
315    
316    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
317        call.
318    
319    14. As documented, (*COMMIT) is now confined to within a positive assertion.
320    
321    15. It is now possible to link pcretest with libedit as an alternative to
322        libreadline.
323    
324    16. (*COMMIT) control verb is now supported by the JIT compiler.
325    
326    17. The Unicode data tables have been updated to Unicode 6.1.0.
327    
328    18. Added --file-list option to pcregrep.
329    
330    19. Added binary file support to pcregrep, including the -a, --binary-files,
331        -I, and --text options.
332    
333    20. The madvise function is renamed for posix_madvise for QNX compatibility
334        reasons. Fixed by Giuseppe D'Angelo.
335    
336    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
337    
338    22. Changed the option for creating symbolic links for 16-bit man pages from
339        -s to -sf so that re-installing does not cause issues.
340    
341    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
342    
343    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
344        matches in certain environments (the workspace was not being correctly
345        retained). Also added to pcre_dfa_exec() a simple plausibility check on
346        some of the workspace data at the beginning of a restart.
347    
348    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
349        was not doing so when it should - probably a typo introduced by SVN 528
350        (change 8.10/14).
351    
352    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
353        \w+ when the character tables indicated that \x{c4} was a word character.
354        There were several related cases, all because the tests for doing a table
355        lookup were testing for characters less than 127 instead of 255.
356    
357    27. If a pattern contains capturing parentheses that are not used in a match,
358        their slots in the ovector are set to -1. For those that are higher than
359        any matched groups, this happens at the end of processing. In the case when
360        there were back references that the ovector was too small to contain
361        (causing temporary malloc'd memory to be used during matching), and the
362        highest capturing number was not used, memory off the end of the ovector
363        was incorrectly being set to -1. (It was using the size of the temporary
364        memory instead of the true size.)
365    
366    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
367        ovector size, it uses memory at the end of the block that it has got.
368    
369    29. Check for an overlong MARK name and give an error at compile time. The
370        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
371    
372    30. JIT compiler update.
373    
374    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
375        Rill for the patch.
376    
377    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
378    
379    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
380    
381    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
382        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
383        was enabled.
384    
385    35. Improve JIT code generation for greedy plus quantifier.
386    
387    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
388        1 to "aa" instead of to an empty string. The bug affected repeated groups
389        that could potentially match an empty string.
390    
391    37. Optimizing single character iterators in JIT.
392    
393    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
394        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
395        codepoints that are too big for the mode are faulted, and in a UTF mode,
396        disallowed codepoints are also faulted.
397    
398    39. If PCRE was compiled with UTF support, in three places in the DFA
399        matcher there was code that should only have been obeyed in UTF mode, but
400        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
401        processing when bytes with values greater than 127 were present. In 16-bit
402        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
403        both cases the values are those that cannot be the first data item in a UTF
404        character. The three items that might have provoked this were recursions,
405        possessively repeated groups, and atomic groups.
406    
407    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
408        and pcregrep, because some OS require shared objects to be explicitly
409        passed to ld, causing the link step to fail if they are not.
410    
411    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
412        mode, patterns that started with \h* or \R* might be incorrectly matched.
413    
414    
415    Version 8.30 04-February-2012
416    -----------------------------
417    
418    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
419        name is defined in ctype.h.
420    
421    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
422        only in quite long subpatterns.
423    
424    3.  Removed the function pcre_info(), which has been obsolete and deprecated
425        since it was replaced by pcre_fullinfo() in February 2000.
426    
427    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
428        match a (*MARK), and the match failed at the start of the subject, a
429        reference to memory before the start of the subject could occur. This bug
430        was introduced by fix 17 of release 8.21.
431    
432    5.  A reference to an unset group with zero minimum repetition was giving
433        totally wrong answers (in non-JavaScript-compatibility mode). For example,
434        /(another)?(\1?)test/ matched against "hello world test". This bug was
435        introduced in release 8.13.
436    
437    6.  Add support for 16-bit character strings (a large amount of work involving
438        many changes and refactorings).
439    
440    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
441        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
442        from a file.
443    
444    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
445        rounding is not applied in this particular case).
446    
447    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
448        if they appear, or are escaped, in patterns.
449    
450    10. Get rid of a number of -Wunused-but-set-variable warnings.
451    
452    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
453        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
454        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
455        also returns the mark "x". This bug applied to capturing parentheses,
456        non-capturing parentheses, and atomic parentheses. It also applied to some
457        assertions.
458    
459    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
460        information out of configure.ac instead of relying on pcre.h.generic, which
461        is not stored in the repository.
462    
463    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
464        -lreadline.
465    
466    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
467    
468    15. Applied Graycode's patch to put the top-level frame on the stack rather
469        than the heap when not using the stack for recursion. This gives a
470        performance improvement in many cases when recursion is not deep.
471    
472    16. Experimental code added to "pcretest -C" to output the stack frame size.
473    
474    
475    Version 8.21 12-Dec-2011
476    ------------------------
477    
478    1.  Updating the JIT compiler.
479    
480    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
481        are added as well.
482    
483    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
484        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
485        calling _pcre_jit_exec. Some extra comments are added.
486    
487    4.  (*MARK) settings inside atomic groups that do not contain any capturing
488        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
489        was introduced by change 18 for 8.20.
490    
491    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
492        ECMA-262 standard.
493    
494    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
495        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
496        This bug was probably introduced by change 9 of 8.13.
497    
498    7.  While fixing 6 above, I noticed that a number of other items were being
499        incorrectly rejected as "not fixed length". This arose partly because newer
500        opcodes had not been added to the fixed-length checking code. I have (a)
501        corrected the bug and added tests for these items, and (b) arranged for an
502        error to occur if an unknown opcode is encountered while checking for fixed
503        length instead of just assuming "not fixed length". The items that were
504        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
505        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
506        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
507    
508    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
509        being incorrectly compiled and would have given unpredicatble results.
510    
511    9.  A possessively repeated subpattern with minimum repeat count greater than
512        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
513        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
514        the first (A) could occur when it should not.
515    
516    10. Add a cast and remove a redundant test from the code.
517    
518    11. JIT should use pcre_malloc/pcre_free for allocation.
519    
520    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
521        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
522        is anything other than /usr, -L is still shown).
523    
524    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
525    
526    14. Perl does not support \N without a following name in a [] class; PCRE now
527        also gives an error.
528    
529    15. If a forward reference was repeated with an upper limit of around 2000,
530        it caused the error "internal error: overran compiling workspace". The
531        maximum number of forward references (including repeats) was limited by the
532        internal workspace, and dependent on the LINK_SIZE. The code has been
533        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
534        the default depends on LINK_SIZE. There is a new upper limit (for safety)
535        of around 200,000 forward references. While doing this, I also speeded up
536        the filling in of repeated forward references.
537    
538    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
539        incorrectly expecting the subject to contain another "a" after the start.
540    
541    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
542        in the match, the SKIP should be ignored. This was not happening; instead
543        the SKIP was being treated as NOMATCH. For patterns such as
544        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
545        tested.
546    
547    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
548        now much more compatible with Perl, in particular in cases where the result
549        is a non-match for a non-anchored pattern. For example, if
550        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
551        "m", where previously it did not return a name. A side effect of this
552        change is that for partial matches, the last encountered mark name is
553        returned, as for non matches. A number of tests that were previously not
554        Perl-compatible have been moved into the Perl-compatible test files. The
555        refactoring has had the pleasing side effect of removing one argument from
556        the match() function, thus reducing its stack requirements.
557    
558    19. If the /S+ option was used in pcretest to study a pattern using JIT,
559        subsequent uses of /S (without +) incorrectly behaved like /S+.
560    
561    21. Retrieve executable code size support for the JIT compiler and fixing
562        some warnings.
563    
564    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
565        not work when the shorter character appeared right at the end of the
566        subject string.
567    
568    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
569        systems.
570    
571    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
572        output it when the /M option is used in pcretest.
573    
574    25. The CheckMan script was not being included in the distribution. Also, added
575        an explicit "perl" to run Perl scripts from the PrepareRelease script
576        because this is reportedly needed in Windows.
577    
578    26. If study data was being save in a file and studying had not found a set of
579        "starts with" bytes for the pattern, the data written to the file (though
580        never used) was taken from uninitialized memory and so caused valgrind to
581        complain.
582    
583    27. Updated RunTest.bat as provided by Sheri Pierce.
584    
585    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
586    
587    29. Computation of memory usage for the table of capturing group names was
588        giving an unnecessarily large value.
589    
590    
591    Version 8.20 21-Oct-2011
592    ------------------------
593    
594    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
595        a POSIX class. After further experiments with Perl, which convinced me that
596        Perl has bugs and confusions, a closing square bracket is no longer allowed
597        in a POSIX name. This bug also affected patterns with classes that started
598        with full stops.
599    
600    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
601        captured substring, but while checking the failing first alternative,
602        substring 1 is temporarily captured. If the output vector supplied to
603        pcre_exec() was not big enough for this capture, the yield of the function
604        was still zero ("insufficient space for captured substrings"). This cannot
605        be totally fixed without adding another stack variable, which seems a lot
606        of expense for a edge case. However, I have improved the situation in cases
607        such as /(a)(b)x|abc/ matched against "abc", where the return code
608        indicates that fewer than the maximum number of slots in the ovector have
609        been set.
610    
611    3.  Related to (2) above: when there are more back references in a pattern than
612        slots in the output vector, pcre_exec() uses temporary memory during
613        matching, and copies in the captures as far as possible afterwards. It was
614        using the entire output vector, but this conflicts with the specification
615        that only 2/3 is used for passing back captured substrings. Now it uses
616        only the first 2/3, for compatibility. This is, of course, another edge
617        case.
618    
619    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
620        main code base, and can be used by building with --enable-jit. When this is
621        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
622        runtime --no-jit option is given.
623    
624    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
625        ovector, the return from the function was zero, implying that there were
626        other matches that did not fit. The correct "exactly full" value is now
627        returned.
628    
629    6.  If a subpattern that was called recursively or as a subroutine contained
630        (*PRUNE) or any other control that caused it to give a non-standard return,
631        invalid errors such as "Error -26 (nested recursion at the same subject
632        position)" or even infinite loops could occur.
633    
634    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
635        computing the minimum length on reaching *ACCEPT, and so ended up with the
636        wrong value of 1 rather than 0. Further investigation indicates that
637        computing a minimum subject length in the presence of *ACCEPT is difficult
638        (think back references, subroutine calls), and so I have changed the code
639        so that no minimum is registered for a pattern that contains *ACCEPT.
640    
641    8.  If (*THEN) was present in the first (true) branch of a conditional group,
642        it was not handled as intended. [But see 16 below.]
643    
644    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
645        Sheri Pierce.
646    
647    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
648        the first byte in a match must be "a".
649    
650    11. Change 17 for 8.13 increased the recursion depth for patterns like
651        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
652        pattern contains any instances of (*THEN). If it does not, the old
653        optimizations are restored. It would be nice to do this on a per-group
654        basis, but at the moment that is not feasible.
655    
656    12. In some environments, the output of pcretest -C is CRLF terminated. This
657        broke RunTest's code that checks for the link size. A single white space
658        character after the value is now allowed for.
659    
660    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
661        For "fr", it uses the Windows-specific input and output files.
662    
663    14. If (*THEN) appeared in a group that was called recursively or as a
664        subroutine, it did not work as intended. [But see next item.]
665    
666    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
667        pattern fragments (but not containing any | characters). If A and B are
668        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
669        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
670        D. In other words, Perl considers parentheses that do not contain any |
671        characters to be part of a surrounding alternative, whereas PCRE was
672        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
673        differently. PCRE now behaves in the same way as Perl, except in the case
674        of subroutine/recursion calls such as (?1) which have in any case always
675        been different (but PCRE had them first :-).
676    
677    16. Related to 15 above: Perl does not treat the | in a conditional group as
678        creating alternatives. Such a group is treated in the same way as an
679        ordinary group without any | characters when processing (*THEN). PCRE has
680        been changed to match Perl's behaviour.
681    
682    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
683        RunGrepTest script failed.
684    
685    18. Change 22 for version 13 caused atomic groups to use more stack. This is
686        inevitable for groups that contain captures, but it can lead to a lot of
687        stack use in large patterns. The old behaviour has been restored for atomic
688        groups that do not contain any capturing parentheses.
689    
690    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
691        suppress the check for a minimum subject length at run time. (If it was
692        given to pcre_exec() or pcre_dfa_exec() it did work.)
693    
694    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
695        fail to work when decoding hex characters in data strings in EBCDIC
696        environments.
697    
698    21. It appears that in at least one Mac OS environment, the isxdigit() function
699        is implemented as a macro that evaluates to its argument more than once,
700        contravening the C 90 Standard (I haven't checked a later standard). There
701        was an instance in pcretest which caused it to go wrong when processing
702        \x{...} escapes in subject strings. The has been rewritten to avoid using
703        things like p++ in the argument of isxdigit().
704    
705    
706    Version 8.13 16-Aug-2011
707    ------------------------
708    
709    1.  The Unicode data tables have been updated to Unicode 6.0.0.
710    
711    2.  Two minor typos in pcre_internal.h have been fixed.
712    
713    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
714        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
715        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
716    
717    4.  There were a number of related bugs in the code for matching backrefences
718        caselessly in UTF-8 mode when codes for the characters concerned were
719        different numbers of bytes. For example, U+023A and U+2C65 are an upper
720        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
721        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
722        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
723        2-byte code at the end of the subject (it thought there wasn't enough data
724        left).
725    
726    5.  Comprehensive information about what went wrong is now returned by
727        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
728        as the output vector has at least 2 elements. The offset of the start of
729        the failing character and a reason code are placed in the vector.
730    
731    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
732        now returned is for the first byte of the failing character, instead of the
733        last byte inspected. This is an incompatible change, but I hope it is small
734        enough not to be a problem. It makes the returned offset consistent with
735        pcre_exec() and pcre_dfa_exec().
736    
737    7.  pcretest now gives a text phrase as well as the error number when
738        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
739        failure, the offset and reason code are output.
740    
741    8.  When \R was used with a maximizing quantifier it failed to skip backwards
742        over a \r\n pair if the subsequent match failed. Instead, it just skipped
743        back over a single character (\n). This seems wrong (because it treated the
744        two characters as a single entity when going forwards), conflicts with the
745        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
746        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
747        has been changed.
748    
749    9.  Some internal refactoring has changed the processing so that the handling
750        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
751        time (the PCRE_DOTALL option was changed this way some time ago: version
752        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
753        which was always a bit of a fudge. It also means that there is one less
754        argument for the match() function, which reduces its stack requirements
755        slightly. This change also fixes an incompatibility with Perl: the pattern
756        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
757    
758    10. More internal refactoring has drastically reduced the number of recursive
759        calls to match() for possessively repeated groups such as (abc)++ when
760        using pcre_exec().
761    
762    11. While implementing 10, a number of bugs in the handling of groups were
763        discovered and fixed:
764    
765        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
766        (a|)*(?1) gave a compile-time internal error.
767        ((a|)+)+  did not notice that the outer group could match an empty string.
768        (^a|^)+   was not marked as anchored.
769        (.*a|.*)+ was not marked as matching at start or after a newline.
770    
771    12. Yet more internal refactoring has removed another argument from the match()
772        function. Special calls to this function are now indicated by setting a
773        value in a variable in the "match data" data block.
774    
775    13. Be more explicit in pcre_study() instead of relying on "default" for
776        opcodes that mean there is no starting character; this means that when new
777        ones are added and accidentally left out of pcre_study(), testing should
778        pick them up.
779    
780    14. The -s option of pcretest has been documented for ages as being an old
781        synonym of -m (show memory usage). I have changed it to mean "force study
782        for every regex", that is, assume /S for every regex. This is similar to -i
783        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
784        using it. It makes it easier to run collections of tests with and without
785        study enabled, and thereby test pcre_study() more easily. All the standard
786        tests are now run with and without -s (but some patterns can be marked as
787        "never study" - see 20 below).
788    
789    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
790        restoration of the capturing data to the outer values was not happening
791        correctly.
792    
793    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
794        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
795        pattern had matched an empty string, and so incorrectly returned a no
796        match.
797    
798    17. There was optimizing code for the last branch of non-capturing parentheses,
799        and also for the obeyed branch of a conditional subexpression, which used
800        tail recursion to cut down on stack usage. Unfortunately, now that there is
801        the possibility of (*THEN) occurring in these branches, tail recursion is
802        no longer possible because the return has to be checked for (*THEN). These
803        two optimizations have therefore been removed. [But see 8.20/11 above.]
804    
805    18. If a pattern containing \R was studied, it was assumed that \R always
806        matched two bytes, thus causing the minimum subject length to be
807        incorrectly computed because \R can also match just one byte.
808    
809    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
810        was incorrectly computed.
811    
812    20. If /S is present twice on a test pattern in pcretest input, it now
813        *disables* studying, thereby overriding the use of -s on the command line
814        (see 14 above). This is necessary for one or two tests to keep the output
815        identical in both cases.
816    
817    21. When (*ACCEPT) was used in an assertion that matched an empty string and
818        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
819    
820    22. When an atomic group that contained a capturing parenthesis was
821        successfully matched, but the branch in which it appeared failed, the
822        capturing was not being forgotten if a higher numbered group was later
823        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
824        group 1 to "a", when in fact it should be unset. This applied to multi-
825        branched capturing and non-capturing groups, repeated or not, and also to
826        positive assertions (capturing in negative assertions does not happen
827        in PCRE) and also to nested atomic groups.
828    
829    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
830        subject after a captured substring, to make it easier to tell which of a
831        number of identical substrings has been captured.
832    
833    24. The way atomic groups are processed by pcre_exec() has been changed so that
834        if they are repeated, backtracking one repetition now resets captured
835        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
836        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
837        "aaa". Previously, it would have been "a". As part of this code
838        refactoring, the way recursive calls are handled has also been changed.
839    
840    25. If an assertion condition captured any substrings, they were not passed
841        back unless some other capturing happened later. For example, if
842        (?(?=(a))a) was matched against "a", no capturing was returned.
843    
844    26. When studying a pattern that contained subroutine calls or assertions,
845        the code for finding the minimum length of a possible match was handling
846        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
847        group 1 called group 2 while simultaneously a separate group 2 called group
848        1). A stack overflow occurred in this case. I have fixed this by limiting
849        the recursion depth to 10.
850    
851    27. Updated RunTest.bat in the distribution to the version supplied by Tom
852        Fortmann. This supports explicit test numbers on the command line, and has
853        argument validation and error reporting.
854    
855    28. An instance of \X with an unlimited repeat could fail if at any point the
856        first character it looked at was a mark character.
857    
858    29. Some minor code refactoring concerning Unicode properties and scripts
859        should reduce the stack requirement of match() slightly.
860    
861    30. Added the '=' option to pcretest to check the setting of unused capturing
862        slots at the end of the pattern, which are documented as being -1, but are
863        not included in the return count.
864    
865    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
866        compiled something random. Now it gives a compile-time error (as does
867        Perl).
868    
869    32. A *MARK encountered during the processing of a positive assertion is now
870        recorded and passed back (compatible with Perl).
871    
872    33. If --only-matching or --colour was set on a pcregrep call whose pattern
873        had alternative anchored branches, the search for a second match in a line
874        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
875        matched the line "0102" twice. The same bug affected patterns that started
876        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
877        twice.
878    
879    34. Previously, PCRE did not allow quantification of assertions. However, Perl
880        does, and because of capturing effects, quantifying parenthesized
881        assertions may at times be useful. Quantifiers are now allowed for
882        parenthesized assertions.
883    
884    35. A minor code tidy in pcre_compile() when checking options for \R usage.
885    
886    36. \g was being checked for fancy things in a character class, when it should
887        just be a literal "g".
888    
889    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
890        appearance of a nested POSIX class supersedes an apparent external class.
891        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
892        unescaped square brackets may also appear as part of class names. For
893        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
894        more like Perl. (But see 8.20/1 above.)
895    
896    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
897        was because it thought it was \N{name}, which is not supported).
898    
899    39. Add minix to OS list not supporting the -S option in pcretest.
900    
901    40. PCRE tries to detect cases of infinite recursion at compile time, but it
902        cannot analyze patterns in sufficient detail to catch mutual recursions
903        such as ((?1))((?2)). There is now a runtime test that gives an error if a
904        subgroup is called recursively as a subpattern for a second time at the
905        same position in the subject string. In previous releases this might have
906        been caught by the recursion limit, or it might have run out of stack.
907    
908    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
909        happen only once. PCRE was, however incorrectly giving a compile time error
910        "recursive call could loop indefinitely" because it cannot analyze the
911        pattern in sufficient detail. The compile time test no longer happens when
912        PCRE is compiling a conditional subpattern, but actual runaway loops are
913        now caught at runtime (see 40 above).
914    
915    42. It seems that Perl allows any characters other than a closing parenthesis
916        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
917        has been changed to be the same.
918    
919    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
920        as not to get warnings when autogen.sh is called. Also changed
921        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
922    
923    44. To help people who use pcregrep to scan files containing exceedingly long
924        lines, the following changes have been made:
925    
926        (a) The default value of the buffer size parameter has been increased from
927            8K to 20K. (The actual buffer used is three times this size.)
928    
929        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
930            PCRE is built.
931    
932        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
933            to be set at run time.
934    
935        (d) Numerical values in pcregrep options can be followed by K or M, for
936            example --buffer-size=50K.
937    
938        (e) If a line being scanned overflows pcregrep's buffer, an error is now
939            given and the return code is set to 2.
940    
941    45. Add a pointer to the latest mark to the callout data block.
942    
943    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
944        partial match of an empty string instead of no match. This was specific to
945        the use of ".".
946    
947    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
948        complete match instead of a partial match. This bug was dependent on both
949        the PCRE_UTF8 and PCRE_DOTALL options being set.
950    
951    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
952        starting byte set, because \b was not being ignored.
953    
954    
955    Version 8.12 15-Jan-2011
956    ------------------------
957    
958    1.  Fixed some typos in the markup of the man pages, and wrote a script that
959        checks for such things as part of the documentation building process.
960    
961    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
962        --match-limit and --recursion-limit options (added for 8.11). In
963        particular, this made one of the standard tests fail. (The integer value
964        went into the wrong half of a long int.)
965    
966    3.  If the --colour option was given to pcregrep with -v (invert match), it
967        did strange things, either producing crazy output, or crashing. It should,
968        of course, ignore a request for colour when reporting lines that do not
969        match.
970    
971    4.  Another pcregrep bug caused similar problems if --colour was specified with
972        -M (multiline) and the pattern match finished with a line ending.
973    
974    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
975        matched in multiline mode, the following line was shown as part of the
976        match. This seems wrong, so I have changed it.
977    
978    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
979        the check for further matches in the same line (so they could be coloured)
980        to overrun the end of the current line. If another match was found, it was
981        incorrectly shown (and then shown again when found in the next line).
982    
983    7.  If pcregrep was compiled under Windows, there was a reference to the
984        function pcregrep_exit() before it was defined. I am assuming this was
985        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
986        reported by a user. I've moved the definition above the reference.
987    
988    
989    Version 8.11 10-Dec-2010
990    ------------------------
991    
992    1.  (*THEN) was not working properly if there were untried alternatives prior
993        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
994        backtracked to try for "b" instead of moving to the next alternative branch
995        at the same level (in this case, to look for "c"). The Perl documentation
996        is clear that when (*THEN) is backtracked onto, it goes to the "next
997        alternative in the innermost enclosing group".
998    
999    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1000        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1001        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1002        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1003        (*THEN).
1004    
1005    3.  If \s appeared in a character class, it removed the VT character from
1006        the class, even if it had been included by some previous item, for example
1007        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1008        of \s, but is part of the POSIX "space" class.)
1009    
1010    4.  A partial match never returns an empty string (because you can always
1011        match an empty string at the end of the subject); however the checking for
1012        an empty string was starting at the "start of match" point. This has been
1013        changed to the "earliest inspected character" point, because the returned
1014        data for a partial match starts at this character. This means that, for
1015        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1016        (previously it gave "no match").
1017    
1018    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1019        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1020        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1021        has an implication that the given string is incomplete (because a partial
1022        match is preferred over a full match). For this reason, these items now
1023        give a partial match in this situation. [Aside: previously, the one case
1024        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1025        match rather than a full match, which was wrong by the old rules, but is
1026        now correct.]
1027    
1028    6.  There was a bug in the handling of #-introduced comments, recognized when
1029        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1030        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1031        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1032        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1033        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1034        places in pcre_compile().
1035    
1036    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1037        comments when looking ahead for named forward references to subpatterns,
1038        the only newline sequence it recognized was NL. It now handles newlines
1039        according to the set newline convention.
1040    
1041    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1042        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1043        cater for a lack of strerror(). These oversights have been fixed.
1044    
1045    9.  Added --match-limit and --recursion-limit to pcregrep.
1046    
1047    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1048    
1049    11. When the -o option was used, pcregrep was setting a return code of 1, even
1050        when matches were found, and --line-buffered was not being honoured.
1051    
1052    12. Added an optional parentheses number to the -o and --only-matching options
1053        of pcregrep.
1054    
1055    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1056        can match an empty string. The code to do it in pcretest and pcredemo
1057        needed fixing:
1058    
1059        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1060            only one byte after an empty string match just before CRLF (this case
1061            just got forgotten; "any" and "anycrlf" were OK).
1062    
1063        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1064            mode when an empty string match preceded an ASCII character followed by
1065            a non-ASCII character. (The code for advancing by one character rather
1066            than one byte was nonsense.)
1067    
1068        (c) The pcredemo.c sample program did not have any code at all to handle
1069            the cases when CRLF is a valid newline sequence.
1070    
1071    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1072        as a starting offset was within the subject string. There is now a new
1073        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1074        negative or greater than the length of the string. In order to test this,
1075        pcretest is extended to allow the setting of negative starting offsets.
1076    
1077    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1078        starting offset points to the beginning of a UTF-8 character was
1079        unnecessarily clumsy. I tidied it up.
1080    
1081    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1082        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1083    
1084    17. Nobody had reported that the --include_dir option, which was added in
1085        release 7.7 should have been called --include-dir (hyphen, not underscore)
1086        for compatibility with GNU grep. I have changed it to --include-dir, but
1087        left --include_dir as an undocumented synonym, and the same for
1088        --exclude-dir, though that is not available in GNU grep, at least as of
1089        release 2.5.4.
1090    
1091    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1092        characters from a string of bytes) have been redefined so as not to use
1093        loops, in order to improve performance in some environments. At the same
1094        time, I abstracted some of the common code into auxiliary macros to save
1095        repetition (this should not affect the compiled code).
1096    
1097    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1098        compile-time error is now given if \c is not followed by an ASCII
1099        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1100        different, and any byte value is allowed.)
1101    
1102    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1103        START_OPTIMIZE option, which is now allowed at compile time - but just
1104        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1105        to pcregrep and other applications that have no direct access to PCRE
1106        options. The new /Y option in pcretest sets this option when calling
1107        pcre_compile().
1108    
1109    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1110        back references. Groups containing recursive back references were forced to
1111        be atomic by that change, but in the case of named groups, the amount of
1112        memory required was incorrectly computed, leading to "Failed: internal
1113        error: code overflow". This has been fixed.
1114    
1115    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1116        pcretest.c, to avoid build problems in some Borland environments.
1117    
1118    
1119    Version 8.10 25-Jun-2010
1120    ------------------------
1121    
1122    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1123        THEN.
1124    
1125    2.  (*ACCEPT) was not working when inside an atomic group.
1126    
1127    3.  Inside a character class, \B is treated as a literal by default, but
1128        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1129        causes the error). The code is unchanged, but I tidied the documentation.
1130    
1131    4.  Inside a character class, PCRE always treated \R and \X as literals,
1132        whereas Perl faults them if its -w option is set. I have changed PCRE so
1133        that it faults them when PCRE_EXTRA is set.
1134    
1135    5.  Added support for \N, which always matches any character other than
1136        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1137    
1138    6.  When compiling pcregrep with newer versions of gcc which may have
1139        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1140        declared with attribute warn_unused_result" were given. Just casting the
1141        result to (void) does not stop the warnings; a more elaborate fudge is
1142        needed. I've used a macro to implement this.
1143    
1144    7.  Minor change to pcretest.c to avoid a compiler warning.
1145    
1146    8.  Added four artifical Unicode properties to help with an option to make
1147        \s etc use properties (see next item). The new properties are: Xan
1148        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1149    
1150    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1151        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1152        this option. Modified pcretest to add /W to test this facility. Added
1153        REG_UCP to make it available via the POSIX interface.
1154    
1155    10. Added --line-buffered to pcregrep.
1156    
1157    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1158        studied, and the match started with a letter with a code point greater than
1159        127 whose first byte was different to the first byte of the other case of
1160        the letter, the other case of this starting letter was not recognized
1161        (#976).
1162    
1163    12. If a pattern that was studied started with a repeated Unicode property
1164        test, for example, \p{Nd}+, there was the theoretical possibility of
1165        setting up an incorrect bitmap of starting bytes, but fortunately it could
1166        not have actually happened in practice until change 8 above was made (it
1167        added property types that matched character-matching opcodes).
1168    
1169    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1170        possible starting bytes for non-anchored patterns.
1171    
1172    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1173        \R, and also a number of cases that involve Unicode properties, both
1174        explicit and implicit when PCRE_UCP is set.
1175    
1176    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1177        input, it could crash or give wrong results if characters with values
1178        greater than 0xc0 were present in the subject string. (Detail: it assumed
1179        UTF-8 input when processing these items.)
1180    
1181    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1182        size_t is 64-bit (#991).
1183    
1184    17. Added a check for running out of memory when PCRE is compiled with
1185        --disable-stack-for-recursion (#990).
1186    
1187    18. If the last data line in a file for pcretest does not have a newline on
1188        the end, a newline was missing in the output.
1189    
1190    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1191        less than 128) in its various bitmaps. However, there is a facility for
1192        generating tables according to the current locale when PCRE is compiled. It
1193        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1194        characters, are recognized by isspace() and therefore were getting set in
1195        these tables, and indeed these tables seem to approximate to ISO 8859. This
1196        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1197        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1198        which of course cannot start UTF-8 characters. I have changed the code so
1199        that only real ASCII characters (less than 128) and the correct starting
1200        bytes for UTF-8 encodings are set for characters greater than 127 when in
1201        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1202        altogether.)
1203    
1204    20. Added the /T option to pcretest so as to be able to run tests with non-
1205        standard character tables, thus making it possible to include the tests
1206        used for 19 above in the standard set of tests.
1207    
1208    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1209        reference to a subpattern the other side of a comment that contains an
1210        opening parenthesis caused either an internal compiling error, or a
1211        reference to the wrong subpattern.
1212    
1213    
1214    Version 8.02 19-Mar-2010
1215    ------------------------
1216    
1217    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1218    
1219    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1220        configured.
1221    
1222    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1223        original author of that file, following a query about its status.
1224    
1225    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1226        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1227    
1228    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1229        quantifier applied to a forward-referencing subroutine call, could compile
1230        incorrect code or give the error "internal error: previously-checked
1231        referenced subpattern not found".
1232    
1233    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1234        variables to point to external functions. For these systems, therefore,
1235        pcre_malloc etc. are now initialized to local functions that call the
1236        relevant global functions.
1237    
1238    7.  There were two entries missing in the vectors called coptable and poptable
1239        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1240        I've fixed the data, and added a kludgy way of testing at compile time that
1241        the lengths are correct (equal to the number of opcodes).
1242    
1243    8.  Following on from 7, I added a similar kludge to check the length of the
1244        eint vector in pcreposix.c.
1245    
1246    9.  Error texts for pcre_compile() are held as one long string to avoid too
1247        much relocation at load time. To find a text, the string is searched,
1248        counting zeros. There was no check for running off the end of the string,
1249        which could happen if a new error number was added without updating the
1250        string.
1251    
1252    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1253    
1254    11. \K was not working if it appeared in an atomic group or in a group that
1255        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1256        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1257        the assertion is positive, but not if it is negative.
1258    
1259    12. Change 11 fortuitously reduced the size of the stack frame used in the
1260        "match()" function of pcre_exec.c by one pointer. Forthcoming
1261        implementation of support for (*MARK) will need an extra pointer on the
1262        stack; I have reserved it now, so that the stack frame size does not
1263        decrease.
1264    
1265    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1266        item in branch that calls a recursion is a subroutine call - as in the
1267        second branch in the above example - was incorrectly given the compile-
1268        time error "recursive call could loop indefinitely" because pcre_compile()
1269        was not correctly checking the subroutine for matching a non-empty string.
1270    
1271    14. The checks for overrunning compiling workspace could trigger after an
1272        overrun had occurred. This is a "should never occur" error, but it can be
1273        triggered by pathological patterns such as hundreds of nested parentheses.
1274        The checks now trigger 100 bytes before the end of the workspace.
1275    
1276    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1277    
1278    
1279    Version 8.01 19-Jan-2010
1280    ------------------------
1281    
1282    1.  If a pattern contained a conditional subpattern with only one branch (in
1283        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1284        computed the wrong minimum data length (which is of course zero for such
1285        subpatterns). This could cause incorrect "no match" results.
1286    
1287    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1288        the pattern is reset in the first branch, pcre_compile() failed with
1289        "internal error: code overflow at offset...". This happened only when
1290        the reset was to the original external option setting. (An optimization
1291        abstracts leading options settings into an external setting, which was the
1292        cause of this.)
1293    
1294    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1295        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1296        assertion pattern did not match (meaning that the assertion was true), it
1297        was incorrectly treated as false if the SKIP had been reached during the
1298        matching. This also applied to assertions used as conditions.
1299    
1300    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1301        assertion subpattern, including such a pattern used as a condition,
1302        unpredictable results occurred, instead of the error return
1303        PCRE_ERROR_DFA_UITEM.
1304    
1305    5.  The C++ GlobalReplace function was not working like Perl for the special
1306        situation when an empty string is matched. It now does the fancy magic
1307        stuff that is necessary.
1308    
1309    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1310        removed. (These were left over from very, very early versions of PCRE.)
1311    
1312    7.  Some cosmetic changes to the code to make life easier when compiling it
1313        as part of something else:
1314    
1315        (a) Change DEBUG to PCRE_DEBUG.
1316    
1317        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1318            called "current" as "current_branch", to prevent a collision with the
1319            Linux macro when compiled as a kernel module.
1320    
1321        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1322            prevent a collision with the Linux macro when compiled as a kernel
1323            module.
1324    
1325    8.  In pcre_compile() there are some checks for integer overflows that used to
1326        cast potentially large values to (double). This has been changed to that
1327        when building, a check for int64_t is made, and if it is found, it is used
1328        instead, thus avoiding the use of floating point arithmetic. (There is no
1329        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1330        double.
1331    
1332    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1333        2005 (difference between two addresses compared to an unsigned value).
1334    
1335    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1336        custom one, because of the following reported problem in Windows:
1337    
1338          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1339              under Win32.
1340          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1341              therefore missing the function definition.
1342          - The compiler thus generates a "C" signature for the test function.
1343          - The linker fails to find the "C" function.
1344          - PCRE fails to configure if asked to do so against libbz2.
1345    
1346    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1347        messages were output:
1348    
1349          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1350          rerunning libtoolize, to keep the correct libtool macros in-tree.
1351          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1352    
1353        I have done both of these things.
1354    
1355    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1356        most of the time, it *can* run out if it is given a pattern that contains a
1357        runaway infinite recursion. I updated the discussion in the pcrestack man
1358        page.
1359    
1360    13. Now that we have gone to the x.xx style of version numbers, the minor
1361        version may start with zero. Using 08 or 09 is a bad idea because users
1362        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1363        interpreted as invalid octal numbers. I've updated the previous comment in
1364        configure.ac, and also added a check that gives an error if 08 or 09 are
1365        used.
1366    
1367    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1368        causing partial matching to fail when the end of the subject matched \W
1369        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1370    
1371    15. There were some discrepancies between the declarations in pcre_internal.h
1372        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1373        their definitions. The declarations used "const uschar *" and the
1374        definitions used USPTR. Even though USPTR is normally defined as "const
1375        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1376        reported that: "This difference in casting confuses some C++ compilers, for
1377        example, SunCC recognizes above declarations as different functions and
1378        generates broken code for hbpcre." I have changed the declarations to use
1379        USPTR.
1380    
1381    16. GNU libtool is named differently on some systems. The autogen.sh script now
1382        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1383        (FreeBSD).
1384    
1385    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1386        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1387        comment: "Figure out how to create a longlong from a string: strtoll and
1388        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1389        instance, but it only takes 2 args instead of 3!"
1390    
1391    18. A subtle bug concerned with back references has been fixed by a change of
1392        specification, with a corresponding code fix. A pattern such as
1393        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1394        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1395        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1396        same bug. Such groups have to be quantified to be useful, or contained
1397        inside another quantified group. (If there's no repetition, the reference
1398        can never match.) The problem arises because, having left the group and
1399        moved on to the rest of the pattern, a later failure that backtracks into
1400        the group uses the captured value from the final iteration of the group
1401        rather than the correct earlier one. I have fixed this in PCRE by forcing
1402        any group that contains a reference to itself to be an atomic group; that
1403        is, there cannot be any backtracking into it once it has completed. This is
1404        similar to recursive and subroutine calls.
1405    
1406    
1407    Version 8.00 19-Oct-09
1408    ----------------------
1409    
1410    1.  The table for translating pcre_compile() error codes into POSIX error codes
1411        was out-of-date, and there was no check on the pcre_compile() error code
1412        being within the table. This could lead to an OK return being given in
1413        error.
1414    
1415    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1416        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1417        in a Windows environment.
1418    
1419    3.  The pcregrep --count option prints the count for each file even when it is
1420        zero, as does GNU grep. However, pcregrep was also printing all files when
1421        --files-with-matches was added. Now, when both options are given, it prints
1422        counts only for those files that have at least one match. (GNU grep just
1423        prints the file name in this circumstance, but including the count seems
1424        more useful - otherwise, why use --count?) Also ensured that the
1425        combination -clh just lists non-zero counts, with no names.
1426    
1427    4.  The long form of the pcregrep -F option was incorrectly implemented as
1428        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1429        but it seems right to fix it, and I didn't think it was worth preserving
1430        the old behaviour.
1431    
1432    5.  The command line items --regex=pattern and --regexp=pattern were not
1433        recognized by pcregrep, which required --regex pattern or --regexp pattern
1434        (with a space rather than an '='). The man page documented the '=' forms,
1435        which are compatible with GNU grep; these now work.
1436    
1437    6.  No libpcreposix.pc file was created for pkg-config; there was just
1438        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1439    
1440    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1441        when UCP support is not needed, by modifying the Python script that
1442        generates it from Unicode data files. This should not matter if the module
1443        is correctly used as a library, but I received one complaint about 50K of
1444        unwanted data. My guess is that the person linked everything into his
1445        program rather than using a library. Anyway, it does no harm.
1446    
1447    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1448        was a minimum greater than 1 for a wide character in a possessive
1449        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1450        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1451        character. Chaos in the form of incorrect output or a compiling loop could
1452        result.
1453    
1454    9.  The restrictions on what a pattern can contain when partial matching is
1455        requested for pcre_exec() have been removed. All patterns can now be
1456        partially matched by this function. In addition, if there are at least two
1457        slots in the offset vector, the offset of the earliest inspected character
1458        for the match and the offset of the end of the subject are set in them when
1459        PCRE_ERROR_PARTIAL is returned.
1460    
1461    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1462        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1463        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1464        and may be more useful for multi-segment matching.
1465    
1466    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1467        used to be given if ever the end of the subject was reached; now it is
1468        given only if matching could not proceed because another character was
1469        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1470        string "Z", which now yields "no match" instead of "partial match". In the
1471        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1472        final character ended with (*FAIL).
1473    
1474    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1475        if the pattern had a "must contain" character that was already found in the
1476        earlier partial match, unless partial matching was again requested. For
1477        example, with the pattern /dog.(body)?/, the "must contain" character is
1478        "g". If the first part-match was for the string "dog", restarting with
1479        "sbody" failed. This bug has been fixed.
1480    
1481    13. The string returned by pcre_dfa_exec() after a partial match has been
1482        changed so that it starts at the first inspected character rather than the
1483        first character of the match. This makes a difference only if the pattern
1484        starts with a lookbehind assertion or \b or \B (\K is not supported by
1485        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1486        matching functions compatible, and I think it's the right thing to do.
1487    
1488    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1489        so that the demonstration program is easily available in environments where
1490        PCRE has not been installed from source.
1491    
1492    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1493        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1494        library.
1495    
1496    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1497        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1498        is not the first non-POSIX option to be added. Clearly some people find
1499        these options useful.
1500    
1501    17. If a caller to the POSIX matching function regexec() passes a non-zero
1502        value for nmatch with a NULL value for pmatch, the value of
1503        nmatch is forced to zero.
1504    
1505    18. RunGrepTest did not have a test for the availability of the -u option of
1506        the diff command, as RunTest does. It now checks in the same way as
1507        RunTest, and also checks for the -b option.
1508    
1509    19. If an odd number of negated classes containing just a single character
1510        interposed, within parentheses, between a forward reference to a named
1511        subpattern and the definition of the subpattern, compilation crashed with
1512        an internal error, complaining that it could not find the referenced
1513        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1514        [The bug was that it was starting one character too far in when skipping
1515        over the character class, thus treating the ] as data rather than
1516        terminating the class. This meant it could skip too much.]
1517    
1518    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1519        /g option in pcretest when the pattern contains \K, which makes it possible
1520        to have an empty string match not at the start, even when the pattern is
1521        anchored. Updated pcretest and pcredemo to use this option.
1522    
1523    21. If the maximum number of capturing subpatterns in a recursion was greater
1524        than the maximum at the outer level, the higher number was returned, but
1525        with unset values at the outer level. The correct (outer level) value is
1526        now given.
1527    
1528    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1529        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1530        make it do so. The string so far is captured, making this feature
1531        compatible with Perl.
1532    
1533    23. The tests have been re-organized, adding tests 11 and 12, to make it
1534        possible to check the Perl 5.10 features against Perl 5.10.
1535    
1536    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1537        pattern matches a fixed length string. PCRE did not allow this; now it
1538        does. Neither allows recursion.
1539    
1540    25. I finally figured out how to implement a request to provide the minimum
1541        length of subject string that was needed in order to match a given pattern.
1542        (It was back references and recursion that I had previously got hung up
1543        on.) This code has now been added to pcre_study(); it finds a lower bound
1544        to the length of subject needed. It is not necessarily the greatest lower
1545        bound, but using it to avoid searching strings that are too short does give
1546        some useful speed-ups. The value is available to calling programs via
1547        pcre_fullinfo().
1548    
1549    26. While implementing 25, I discovered to my embarrassment that pcretest had
1550        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1551        study optimizations had never been tested with that matching function.
1552        Oops. What is worse, even when it was passed study data, there was a bug in
1553        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1554        were also very few tests of studied patterns with pcre_dfa_exec().
1555    
1556    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1557        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1558        on the other side of the coin, they are no longer allowed to have different
1559        names, because these cannot be distinguished in PCRE, and this has caused
1560        confusion. (This is a difference from Perl.)
1561    
1562    28. When duplicate subpattern names are present (necessarily with different
1563        numbers, as required by 27 above), and a test is made by name in a
1564        conditional pattern, either for a subpattern having been matched, or for
1565        recursion in such a pattern, all the associated numbered subpatterns are
1566        tested, and the overall condition is true if the condition is true for any
1567        one of them. This is the way Perl works, and is also more like the way
1568        testing by number works.
1569    
1570    
1571    Version 7.9 11-Apr-09
1572    ---------------------
1573    
1574    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1575        (pcretest), all targets were linked against these libraries. This included
1576        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1577        libraries. This caused unwanted dependencies to be created. This problem
1578        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1579        pcretest is linked with readline.
1580    
1581    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1582        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1583        moved outside it again, because FALSE and TRUE are already defined in AIX,
1584        but BOOL is not.
1585    
1586    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1587        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1588    
1589    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1590        hyphens) following file names and line numbers when outputting matching
1591        lines. This is not true; no spaces are inserted. I have also clarified the
1592        wording for the --colour (or --color) option.
1593    
1594    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1595        was not coloured; this is different to GNU grep, so I have changed it to be
1596        the same.
1597    
1598    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1599        each matching line was coloured. Now it goes on to look for further matches
1600        of any of the test patterns, which is the same behaviour as GNU grep.
1601    
1602    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1603        doesn't make sense to accept an empty string match in pcregrep, so I have
1604        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1605        seems to be how GNU grep behaves.
1606    
1607    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1608        start or after a newline", because the conditional assertion was not being
1609        correctly handled. The rule now is that both the assertion and what follows
1610        in the first alternative must satisfy the test.
1611    
1612    9.  If auto-callout was enabled in a pattern with a conditional group whose
1613        condition was an assertion, PCRE could crash during matching, both with
1614        pcre_exec() and pcre_dfa_exec().
1615    
1616    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1617        used for matching.
1618    
1619    11. Unicode property support in character classes was not working for
1620        characters (bytes) greater than 127 when not in UTF-8 mode.
1621    
1622    12. Added the -M command line option to pcretest.
1623    
1624    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1625    
1626    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1627    
1628    16. Added comments and documentation about mis-use of no_arg in the C++
1629        wrapper.
1630    
1631    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1632        from Martin Jerabek that uses macro names for all relevant character and
1633        string constants.
1634    
1635    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1636        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1637        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1638        these, but not everybody uses configure.
1639    
1640    19. A conditional group that had only one branch was not being correctly
1641        recognized as an item that could match an empty string. This meant that an
1642        enclosing group might also not be so recognized, causing infinite looping
1643        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1644        with the subject "ab", where knowledge that the repeated group can match
1645        nothing is needed in order to break the loop.
1646    
1647    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1648        exec(), but without supplying a callout function, matching went wrong.
1649    
1650    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1651        leak if the size of the offset vector was greater than 30. When the vector
1652        is smaller, the saved offsets during recursion go onto a local stack
1653        vector, but for larger vectors malloc() is used. It was failing to free
1654        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1655        error, in fact).
1656    
1657    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1658        heapframe that is used only when UTF-8 support is enabled. This caused no
1659        problem, but was untidy.
1660    
1661    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1662        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1663        included within another project.
1664    
1665    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1666        slightly modified by me:
1667    
1668          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1669              not building pcregrep.
1670    
1671          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1672              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1673    
1674    25. Forward references, both numeric and by name, in patterns that made use of
1675        duplicate group numbers, could behave incorrectly or give incorrect errors,
1676        because when scanning forward to find the reference group, PCRE was not
1677        taking into account the duplicate group numbers. A pattern such as
1678        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1679    
1680    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1681        the feature of a custom pointer more persuasive (as requested by a user).
1682    
1683    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1684        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1685        pre-defined.
1686    
1687    28. Added support for (*UTF8) at the start of a pattern.
1688    
1689    29. Arrange for flags added by the "release type" setting in CMake to be shown
1690        in the configuration summary.
1691    
1692    
1693    Version 7.8 05-Sep-08
1694    ---------------------
1695    
1696    1.  Replaced UCP searching code with optimized version as implemented for Ad
1697        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1698        stage table and inline lookup instead of a function, giving speed ups of 2
1699        to 5 times on some simple patterns that I tested. Permission was given to
1700        distribute the MultiStage2.py script that generates the tables (it's not in
1701        the tarball, but is in the Subversion repository).
1702    
1703    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1704        scripts.
1705    
1706    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1707        a group with a zero qualifier. The result of the study could be incorrect,
1708        or the function might crash, depending on the pattern.
1709    
1710    4.  Caseless matching was not working for non-ASCII characters in back
1711        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1712        It now works when Unicode Property Support is available.
1713    
1714    5.  In pcretest, an escape such as \x{de} in the data was always generating
1715        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1716        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1717        truncation.
1718    
1719    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1720    
1721    7.  Added two (int) casts to pcregrep when printing the difference of two
1722        pointers, in case they are 64-bit values.
1723    
1724    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1725        test 2 if it fails.
1726    
1727    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1728        and a #define of that name to empty if it is not externally set. This is to
1729        allow users of MSVC to set it if necessary.
1730    
1731    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1732        the convenience functions in the pcre_get.c source file.
1733    
1734    11. An option change at the start of a pattern that had top-level alternatives
1735        could cause overwriting and/or a crash. This command provoked a crash in
1736        some environments:
1737    
1738          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1739    
1740        This potential security problem was recorded as CVE-2008-2371.
1741    
1742    12. For a pattern where the match had to start at the beginning or immediately
1743        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1744        pcre_dfa_exec() could read past the end of the passed subject if there was
1745        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1746        pcretest so that it places the subject at the end of its malloc-ed buffer.
1747    
1748    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1749        exec() might read past the end of the data buffer in UTF-8 mode.
1750    
1751    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1752        the data contained the byte 0x85 as part of a UTF-8 character within its
1753        first line. This applied both to normal and DFA matching.
1754    
1755    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1756        /^[^d]*?$/8 failed to match "abc".
1757    
1758    16. Added a missing copyright notice to pcrecpp_internal.h.
1759    
1760    17. Make it more clear in the documentation that values returned from
1761        pcre_exec() in ovector are byte offsets, not character counts.
1762    
1763    18. Tidied a few places to stop certain compilers from issuing warnings.
1764    
1765    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1766        supplied by Stefan Weber. I made a further small update for 7.8 because
1767        there is a change of source arrangements: the pcre_searchfuncs.c module is
1768        replaced by pcre_ucd.c.
1769    
1770    
1771    Version 7.7 07-May-08
1772    ---------------------
1773    
1774    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1775        a string to a long long, pretend we don't even have a long long." This is
1776        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1777    
1778    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1779        pre-7.6 versions, which defined a global no_arg variable instead of putting
1780        it in the RE class. (See also #8 below.)
1781    
1782    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1783        Lopes.
1784    
1785    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1786    
1787        (1) The include/exclude patterns were being applied to the whole pathnames
1788            of files, instead of just to the final components.
1789    
1790        (2) If there was more than one level of directory, the subdirectories were
1791            skipped unless they satisfied the include/exclude conditions. This is
1792            inconsistent with GNU grep (and could even be seen as contrary to the
1793            pcregrep specification - which I improved to make it absolutely clear).
1794            The action now is always to scan all levels of directory, and just
1795            apply the include/exclude patterns to regular files.
1796    
1797    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1798        --exclude_dir in the tests to avoid scanning .svn directories.
1799    
1800    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1801        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1802        doesn't support NULs in patterns.
1803    
1804    7.  Added some missing "const"s to declarations of static tables in
1805        pcre_compile.c and pcre_dfa_exec.c.
1806    
1807    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1808        caused by fix #2  above. (Subsequently also a second patch to fix the
1809        first patch. And a third patch - this was a messy problem.)
1810    
1811    9.  Applied Craig's patch to remove the use of push_back().
1812    
1813    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1814        matching function regexec().
1815    
1816    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1817        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1818        references. PCRE supports relative numbers with this syntax (I don't think
1819        Oniguruma does).
1820    
1821    12. Previously, a group with a zero repeat such as (...){0} was completely
1822        omitted from the compiled regex. However, this means that if the group
1823        was called as a subroutine from elsewhere in the pattern, things went wrong
1824        (an internal error was given). Such groups are now left in the compiled
1825        pattern, with a new opcode that causes them to be skipped at execution
1826        time.
1827    
1828    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1829        to the way PCRE behaves:
1830    
1831        (a) A lone ] character is dis-allowed (Perl treats it as data).
1832    
1833        (b) A back reference to an unmatched subpattern matches an empty string
1834            (Perl fails the current match path).
1835    
1836        (c) A data ] in a character class must be notated as \] because if the
1837            first data character in a class is ], it defines an empty class. (In
1838            Perl it is not possible to have an empty class.) The empty class []
1839            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1840            The negative empty class [^] matches any one character, independently
1841            of the DOTALL setting.
1842    
1843    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1844        non-existent subpattern following a character class starting with ']' and
1845        containing () gave an internal compiling error instead of "reference to
1846        non-existent subpattern". Fortunately, when the pattern did exist, the
1847        compiled code was correct. (When scanning forwards to check for the
1848        existencd of the subpattern, it was treating the data ']' as terminating
1849        the class, so got the count wrong. When actually compiling, the reference
1850        was subsequently set up correctly.)
1851    
1852    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1853        it was being rejected as not supported by pcre_dfa_exec(), even though
1854        other assertions are supported. I have made pcre_dfa_exec() support
1855        (*FAIL).
1856    
1857    16. The implementation of 13c above involved the invention of a new opcode,
1858        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1859        cannot be changed at match time, I realized I could make a small
1860        improvement to matching performance by compiling OP_ALLANY instead of
1861        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1862        on the OP_ANY path.
1863    
1864    17. Compiling pcretest on Windows with readline support failed without the
1865        following two fixes: (1) Make the unistd.h include conditional on
1866        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1867    
1868    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1869        ncurses library to be included for pcretest when ReadLine support is
1870        requested, but also to allow for it to be overridden. This patch came from
1871        Daniel Bergström.
1872    
1873    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1874        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1875        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1876        spotting this.
1877    
1878    
1879    Version 7.6 28-Jan-08
1880    ---------------------
1881    
1882    1.  A character class containing a very large number of characters with
1883        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1884        overflow.
1885    
1886    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1887        HAVE_LONG_LONG is not defined.
1888    
1889    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1890        bring them up to date and include new features. This patch includes:
1891    
1892        - Fixed PH's badly added libz and libbz2 support.
1893        - Fixed a problem with static linking.
1894        - Added pcredemo. [But later removed - see 7 below.]
1895        - Fixed dftables problem and added an option.
1896        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1897            HAVE_LONG_LONG.
1898        - Added readline support for pcretest.
1899        - Added an listing of the option settings after cmake has run.
1900    
1901    4.  A user submitted a patch to Makefile that makes it easy to create
1902        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1903        Makefile.am that cause it to include this special target, without
1904        affecting anything else. Note that the same mingw target plus all
1905        the other distribution libraries and programs are now supported
1906        when configuring with CMake (see 6 below) instead of with
1907        Configure/Make.
1908    
1909    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1910        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1911        exported in the Windows port". It has not yet been confirmed that the patch
1912        solves the problem, but it does no harm.
1913    
1914    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1915        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1916        with CMake, and also correct the comment about stack recursion.
1917    
1918    7.  Remove the automatic building of pcredemo from the ./configure system and
1919        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1920        of a program that users should build themselves after PCRE is installed, so
1921        building it automatically is not really right. What is more, it gave
1922        trouble in some build environments.
1923    
1924    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1925    
1926    
1927    Version 7.5 10-Jan-08
1928    ---------------------
1929    
1930    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1931        values in parens when parsing an RE using the C++ wrapper."
1932    
1933    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1934        Characters greater than 255 were excluded from the class instead of being
1935        included.
1936    
1937    3.  The same bug as (2) above applied to negated POSIX classes such as
1938        [:^space:].
1939    
1940    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1941        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1942        I have changed it.
1943    
1944    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1945        first named subpattern) and a construct such as (?&a) would reference the
1946        first named subpattern whose name started with "a" (in other words, the
1947        length check was missing). Both these problems are fixed. "Subpattern name
1948        expected" is now given for (?&) (a zero-length name), and this patch also
1949        makes it give the same error for \k'' (previously it complained that that
1950        was a reference to a non-existent subpattern).
1951    
1952    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1953        this is right because (?- can be followed by option settings as well as by
1954        digits. I have, however, made the messages clearer.
1955    
1956    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1957        than the number used in the conditional) now cause a compile-time error.
1958        This is actually not compatible with Perl, which accepts such patterns, but
1959        treats the conditional as always being FALSE (as PCRE used to), but it
1960        seems to me that giving a diagnostic is better.
1961    
1962    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1963        and messages.
1964    
1965    9.  Fix two occurrences of "backslash" in comments that should have been
1966        "backspace".
1967    
1968    10. Remove two redundant lines of code that can never be obeyed (their function
1969        was moved elsewhere).
1970    
1971    11. The program that makes PCRE's Unicode character property table had a bug
1972        which caused it to generate incorrect table entries for sequences of
1973        characters that have the same character type, but are in different scripts.
1974        It amalgamated them into a single range, with the script of the first of
1975        them. In other words, some characters were in the wrong script. There were
1976        thirteen such cases, affecting characters in the following ranges:
1977    
1978          U+002b0 - U+002c1
1979          U+0060c - U+0060d
1980          U+0061e - U+00612
1981          U+0064b - U+0065e
1982          U+0074d - U+0076d
1983          U+01800 - U+01805
1984          U+01d00 - U+01d77
1985          U+01d9b - U+01dbf
1986          U+0200b - U+0200f
1987          U+030fc - U+030fe
1988          U+03260 - U+0327f
1989          U+0fb46 - U+0fbb1
1990          U+10450 - U+1049d
1991    
1992    12. The -o option (show only the matching part of a line) for pcregrep was not
1993        compatible with GNU grep in that, if there was more than one match in a
1994        line, it showed only the first of them. It now behaves in the same way as
1995        GNU grep.
1996    
1997    13. If the -o and -v options were combined for pcregrep, it printed a blank
1998        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1999        does the same. The return code can be used to tell if there were any
2000        non-matching lines.
2001    
2002    14. Added --file-offsets and --line-offsets to pcregrep.
2003    
2004    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2005        infinitely looping recursion. The bug was that positive lookaheads were not
2006        being skipped when checking for a possible empty match (negative lookaheads
2007        and both kinds of lookbehind were skipped).
2008    
2009    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2010        inclusion of <windows.h> to before rather than after the definition of
2011        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2012    
2013    17. Specifying a possessive quantifier with a specific limit for a Unicode
2014        character property caused pcre_compile() to compile bad code, which led at
2015        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2016        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2017        caused the error; without that there was no problem.
2018    
2019    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2020    
2021    19. Added --enable-pcretest-libreadline.
2022    
2023    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2024        RE::GlobalReplace(). As a result, the number of replacements returned was
2025        double what it should be. I removed one of the increments, but Craig sent a
2026        later patch that removed the other one (the right fix) and added unit tests
2027        that check the return values (which was not done before).
2028    
2029    21. Several CMake things:
2030    
2031        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2032            the names libpcre and libpcreposix, not just pcre and pcreposix.
2033    
2034        (2) The above change means that pcretest and pcregrep are now correctly
2035            linked with the newly-built libraries, not previously installed ones.
2036    
2037        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2038    
2039    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2040        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2041        UTF-8 newline character). The key issue is that the pattern starts .*;
2042        this means that the match must be either at the beginning, or after a
2043        newline. The bug was in the code for advancing after a failed match and
2044        checking that the new position followed a newline. It was not taking
2045        account of UTF-8 characters correctly.
2046    
2047    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2048        character classes. PCRE was not treating the sequence [:...:] as a
2049        character class unless the ... were all letters. Perl, however, seems to
2050        allow any characters between [: and :], though of course it rejects as
2051        unknown any "names" that contain non-letters, because all the known class
2052        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2053        for example, whereas PCRE did not - it did not recognize a POSIX character
2054        class. This seemed a bit dangerous, so the code has been changed to be
2055        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2056        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2057        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2058        Perl does, and where it didn't before.
2059    
2060    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2061        Windows environments %n is disabled by default.
2062    
2063    
2064    Version 7.4 21-Sep-07
2065    ---------------------
2066    
2067    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2068        means that a class such as [\s] counted as "explicit reference to CR or
2069        LF". That isn't really right - the whole point of the change was to try to
2070        help when there was an actual mention of one of the two characters. So now
2071        the change happens only if \r or \n (or a literal CR or LF) character is
2072        encountered.
2073    
2074    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2075        of both had grown to the point where there were only 3 bits left.
2076        Fortunately, there was spare space in the data structure, and so I have
2077        moved the internal flags into a new 16-bit field to free up more option
2078        bits.
2079    
2080    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2081        but did not set the internal JCHANGED flag - either of these is enough to
2082        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2083        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2084        start sets both bits.
2085    
2086    4.  Added options (at build time, compile time, exec time) to change \R from
2087        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2088    
2089    5.  doc/pcresyntax.html was missing from the distribution.
2090    
2091    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2092        compatibility, even though it is no longer used.
2093    
2094    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2095        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2096        windows.h file is present (where different names are used). [This was
2097        reversed later after testing - see 16 below.]
2098    
2099    8.  Changed all #include <config.h> to #include "config.h". There were also
2100        some further <pcre.h> cases that I changed to "pcre.h".
2101    
2102    9.  When pcregrep was used with the --colour option, it missed the line ending
2103        sequence off the lines that it output.
2104    
2105    10. It was pointed out to me that arrays of string pointers cause lots of
2106        relocations when a shared library is dynamically loaded. A technique of
2107        using a single long string with a table of offsets can drastically reduce
2108        these. I have refactored PCRE in four places to do this. The result is
2109        dramatic:
2110    
2111          Originally:                          290
2112          After changing UCP table:            187
2113          After changing error message table:   43
2114          After changing table of "verbs"       36
2115          After changing table of Posix names   22
2116    
2117        Thanks to the folks working on Gregex for glib for this insight.
2118    
2119    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2120        unicode-properties was also set.
2121    
2122    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2123    
2124    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2125        checked only for CRLF.
2126    
2127    14. Added casts to pcretest.c to avoid compiler warnings.
2128    
2129    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2130    
2131    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2132        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2133        entirely. This removes changes made in 7 above.
2134    
2135    17. The CMake files have been updated, and there is now more information about
2136        building with CMake in the NON-UNIX-USE document.
2137    
2138    
2139    Version 7.3 28-Aug-07
2140    ---------------------
2141    
2142     1. In the rejigging of the build system that eventually resulted in 7.1, the
2143        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2144        brackets there is not right, since it causes compilers to look for an
2145        installed pcre.h, not the version that is in the source that is being
2146        compiled (which of course may be different). I have changed it back to:
2147    
2148          #include "pcre.h"
2149    
2150        I have a vague recollection that the change was concerned with compiling in
2151        different directories, but in the new build system, that is taken care of
2152        by the VPATH setting the Makefile.
2153    
2154     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2155        when the subject happened to end in the byte 0x85 (e.g. if the last
2156        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2157        characters but of course it shouldn't be taken as a newline when it is part
2158        of another character. The bug was that, for an unlimited repeat of . in
2159        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2160        characters when looking for a newline.
2161    
2162     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2163    
2164     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2165        in debug output.
2166    
2167     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2168        long printing in the pcrecpp unittest when running under MinGW.
2169    
2170     6. ESC_K was left out of the EBCDIC table.
2171    
2172     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2173        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2174        limit also applies to "virtual nesting" when a pattern is recursive, and in
2175        this case 1000 isn't so big. I have been able to remove this limit at the
2176        expense of backing off one optimization in certain circumstances. Normally,
2177        when pcre_exec() would call its internal match() function recursively and
2178        immediately return the result unconditionally, it uses a "tail recursion"
2179        feature to save stack. However, when a subpattern that can match an empty
2180        string has an unlimited repetition quantifier, it no longer makes this
2181        optimization. That gives it a stack frame in which to save the data for
2182        checking that an empty string has been matched. Previously this was taken
2183        from the 1000-entry workspace that had been reserved. So now there is no
2184        explicit limit, but more stack is used.
2185    
2186     8. Applied Daniel's patches to solve problems with the import/export magic
2187        syntax that is required for Windows, and which was going wrong for the
2188        pcreposix and pcrecpp parts of the library. These were overlooked when this
2189        problem was solved for the main library.
2190    
2191     9. There were some crude static tests to avoid integer overflow when computing
2192        the size of patterns that contain repeated groups with explicit upper
2193        limits. As the maximum quantifier is 65535, the maximum group length was
2194        set at 30,000 so that the product of these two numbers did not overflow a
2195        32-bit integer. However, it turns out that people want to use groups that
2196        are longer than 30,000 bytes (though not repeat them that many times).
2197        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2198        made it possible to implement the integer overflow checks in a much more
2199        dynamic way, which I have now done. The artificial limitation on group
2200        length has been removed - we now have only the limit on the total length of
2201        the compiled pattern, which depends on the LINK_SIZE setting.
2202    
2203    10. Fixed a bug in the documentation for get/copy named substring when
2204        duplicate names are permitted. If none of the named substrings are set, the
2205        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2206        empty string.
2207    
2208    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2209        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2210        because the ] is interpreted as the first data character and the
2211        terminating ] is not found. PCRE has been made compatible with Perl in this
2212        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2213        cause memory overwriting.
2214    
2215    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2216        string has been matched (to stop an infinite loop). It was not recognizing
2217        a conditional subpattern that could match an empty string if that
2218        subpattern was within another subpattern. For example, it looped when
2219        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2220        condition was not nested. This bug has been fixed.
2221    
2222    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2223        past the start of the subject in the presence of bytes with the top bit
2224        set, for example "\x8aBCD".
2225    
2226    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2227        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2228    
2229    14. Optimized (?!) to (*FAIL).
2230    
2231    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2232        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2233        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2234        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2235        does: it's just the validity check that is more restrictive.
2236    
2237    16. Inserted checks for integer overflows during escape sequence (backslash)
2238        processing, and also fixed erroneous offset values for syntax errors during
2239        backslash processing.
2240    
2241    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2242        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2243    
2244    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2245        caused an overrun.
2246    
2247    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2248        something other than just ASCII characters) inside a group that had an
2249        unlimited repeat caused a loop at compile time (while checking to see
2250        whether the group could match an empty string).
2251    
2252    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2253        [\P{Any}] did so. (Error in the code for printing property names.)
2254    
2255    21. An orphan \E inside a character class could cause a crash.
2256    
2257    22. A repeated capturing bracket such as (A)? could cause a wild memory
2258        reference during compilation.
2259    
2260    23. There are several functions in pcre_compile() that scan along a compiled
2261        expression for various reasons (e.g. to see if it's fixed length for look
2262        behind). There were bugs in these functions when a repeated \p or \P was
2263        present in the pattern. These operators have additional parameters compared
2264        with \d, etc, and these were not being taken into account when moving along
2265        the compiled data. Specifically:
2266    
2267        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2268            length.
2269    
2270        (b) An item such as \pL+ within a repeated group could cause crashes or
2271            loops.
2272    
2273        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2274            "reference to non-existent subpattern" error.
2275    
2276        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2277    
2278    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2279        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2280    
2281    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2282    
2283    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2284        character were causing crashes (broken optimization).
2285    
2286    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2287        \p or \P) caused a compile-time loop.
2288    
2289    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2290        break. For example, the unstudied pattern [\r\n]A does not match the string
2291        "\r\nA" because change 7.0/46 below moves the current point on by two
2292        characters after failing to match at the start. However, the pattern \nA
2293        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2294        the same is true. There doesn't seem any very clean way out of this, but
2295        what I have chosen to do makes the common cases work: PCRE now takes note
2296        of whether there can be an explicit match for \r or \n anywhere in the
2297        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2298        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2299        pattern has explicit CR or LF references.
2300    
2301    29. Added (*CR) etc for changing newline setting at start of pattern.
2302    
2303    
2304    Version 7.2 19-Jun-07
2305    ---------------------
2306    
2307     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2308        which is apparently normally available under Windows.
2309    
2310     2. Re-jig the pcregrep tests with different newline settings in an attempt
2311        to make them independent of the local environment's newline setting.
2312    
2313     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2314    
2315     4. Some of the "internals" tests were previously cut out when the link size
2316        was not 2, because the output contained actual offsets. The recent new
2317        "Z" feature of pcretest means that these can be cut out, making the tests
2318        usable with all link sizes.
2319    
2320     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2321        stack recursion. This gives a massive performance boost under BSD, but just
2322        a small improvement under Linux. However, it saves one field in the frame
2323        in all cases.
2324    
2325     6. Added more features from the forthcoming Perl 5.10:
2326    
2327        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2328            recursion call. It refers to the nth most recently opened parentheses.
2329    
2330        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2331            to be opened parentheses.
2332    
2333        (c) Conditions that refer to capturing parentheses can be specified
2334            relatively, for example, (?(-2)... or (?(+3)...
2335    
2336        (d) \K resets the start of the current match so that everything before
2337            is not part of it.
2338    
2339        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2340    
2341        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2342            reference syntax.
2343    
2344        (g) (?| introduces a group in which the numbering of parentheses in each
2345            alternative starts with the same number.
2346    
2347        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2348    
2349     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2350        PCRE_INFO_JCHANGED.
2351    
2352     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2353        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2354        for detecting groups that can match an empty string.
2355    
2356     9. A pattern with a very large number of alternatives (more than several
2357        hundred) was running out of internal workspace during the pre-compile
2358        phase, where pcre_compile() figures out how much memory will be needed. A
2359        bit of new cunning has reduced the workspace needed for groups with
2360        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2361        workspace instead of running out of the 4096 that are available.
2362    
2363    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2364    
2365    11. Applied patch from Google to remove an optimization that didn't quite work.
2366        The report of the bug said:
2367    
2368          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2369          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2370          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2371    
2372    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2373        it matched the wrong number of bytes.
2374    
2375    
2376    Version 7.1 24-Apr-07
2377    ---------------------
2378    
2379     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2380        that is more "standard", making use of automake and other Autotools. There
2381        is some re-arrangement of the files and adjustment of comments consequent
2382        on this.
2383    
2384     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2385        for recursive directory scanning broke on some systems because the files
2386        are not scanned in any specific order and on different systems the order
2387        was different. A call to "sort" has been inserted into RunGrepTest for the
2388        approprate test as a short-term fix. In the longer term there may be an
2389        alternative.
2390    
2391     3. I had an email from Eric Raymond about problems translating some of PCRE's
2392        man pages to HTML (despite the fact that I distribute HTML pages, some
2393        people do their own conversions for various reasons). The problems
2394        concerned the use of low-level troff macros .br and .in. I have therefore
2395        removed all such uses from the man pages (some were redundant, some could
2396        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2397        HTML has been updated to handle .nf/.fi and to complain if it encounters
2398        .br or .in.
2399    
2400     4. Updated comments in configure.ac that get placed in config.h.in and also
2401        arranged for config.h to be included in the distribution, with the name
2402        config.h.generic, for the benefit of those who have to compile without
2403        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2404    
2405     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2406        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2407        makevp.bat for the current PCRE, using the additional files
2408        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2409    
2410     6. A Windows user reported a minor discrepancy with test 2, which turned out
2411        to be caused by a trailing space on an input line that had got lost in his
2412        copy. The trailing space was an accident, so I've just removed it.
2413    
2414     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2415        that is needed.
2416    
2417     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2418        as "const" (a) because they are and (b) because it helps the PHP
2419        maintainers who have recently made a script to detect big data structures
2420        in the php code that should be moved to the .rodata section. I remembered
2421        to update Builducptable as well, so it won't revert if ucptable.h is ever
2422        re-created.
2423    
2424     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2425        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2426        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2427        support is not required. This saves 1.5-2K of code, which is important in
2428        some applications.
2429    
2430        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2431        so as not to refer to the tables, even though these functions will never be
2432        called when UTF-8 support is disabled. Otherwise there are problems with a
2433        shared library.
2434    
2435    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2436    
2437        (a) It was defining its arguments as char * instead of void *.
2438    
2439        (b) It was assuming that all moves were upwards in memory; this was true
2440            a long time ago when I wrote it, but is no longer the case.
2441    
2442        The emulated memove() is provided for those environments that have neither
2443        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2444        is clearly not the case, as these two bugs were recently reported.
2445    
2446    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2447        and Detrail to create the HTML documentation, the .txt form of the man
2448        pages, and it removes trailing spaces from listed files. It also creates
2449        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2450        case, it wraps all the #defines with #ifndefs. This script should be run
2451        before "make dist".
2452    
2453    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2454        with Unicode property support.
2455    
2456        (a) For a maximizing quantifier, if the two different cases of the
2457            character were of different lengths in their UTF-8 codings (there are
2458            some cases like this - I found 11), and the matching function had to
2459            back up over a mixture of the two cases, it incorrectly assumed they
2460            were both the same length.
2461    
2462        (b) When PCRE was configured to use the heap rather than the stack for
2463            recursion during matching, it was not correctly preserving the data for
2464            the other case of a UTF-8 character when checking ahead for a match
2465            while processing a minimizing repeat. If the check also involved
2466            matching a wide character, but failed, corruption could cause an
2467            erroneous result when trying to check for a repeat of the original
2468            character.
2469    
2470    13. Some tidying changes to the testing mechanism:
2471    
2472        (a) The RunTest script now detects the internal link size and whether there
2473            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2474            values substituted by "configure". (The RunGrepTest script already did
2475            this for UTF-8.) The configure.ac script no longer substitutes the
2476            relevant variables.
2477    
2478        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2479            with length and offset values. This means that the output is different
2480            for different internal link sizes. Test 2 is skipped for link sizes
2481            other than 2 because of this, bypassing the problem. Unfortunately,
2482            there was also a test in test 3 (the locale tests) that used /B and
2483            failed for link sizes other than 2. Rather than cut the whole test out,
2484            I have added a new /Z option to pcretest that replaces the length and
2485            offset values with spaces. This is now used to make test 3 independent
2486            of link size. (Test 2 will be tidied up later.)
2487    
2488    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2489        segmentation fault instead of returning the appropriate error message.
2490    
2491    15. In multiline mode when the newline sequence was set to "any", the pattern
2492        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2493        This doesn't seem right; it now treats the CRLF combination as the line
2494        ending, and so does not match in that case. It's only a pattern such as ^$
2495        that would hit this one: something like ^ABC$ would have failed after \r
2496        and then tried again after \r\n.
2497    
2498    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2499        in an attempt to make files that differ only in their line terminators
2500        compare equal. This works on Linux.
2501    
2502    17. Under certain error circumstances pcregrep might try to free random memory
2503        as it exited. This is now fixed, thanks to valgrind.
2504    
2505    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2506        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2507        was because its rules for how to advance for /g after matching an empty
2508        string at the end of a line did not allow for this case. They now check for
2509        it specially.
2510    
2511    20. pcretest is supposed to handle patterns and data of any length, by
2512        extending its buffers when necessary. It was getting this wrong when the
2513        buffer for a data line had to be extended.
2514    
2515    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2516        CRLF as a newline sequence.
2517    
2518    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2519        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2520        I have nevertheless tidied it up.
2521    
2522    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2523    
2524    24. Added a man page for pcre-config.
2525    
2526    
2527    Version 7.0 19-Dec-06
2528    ---------------------
2529    
2530     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2531        moving to gcc 4.1.1.
2532    
2533     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2534        sys/time.h, which is documented as needed for this function. It doesn't
2535        seem to matter on Linux, but it showed up on some releases of OS X.
2536    
2537     3. It seems that there are systems where bytes whose values are greater than
2538        127 match isprint() in the "C" locale. The "C" locale should be the
2539        default when a C program starts up. In most systems, only ASCII printing
2540        characters match isprint(). This difference caused the output from pcretest
2541        to vary, making some of the tests fail. I have changed pcretest so that:
2542    
2543        (a) When it is outputting text in the compiled version of a pattern, bytes
2544            other than 32-126 are always shown as hex escapes.
2545    
2546        (b) When it is outputting text that is a matched part of a subject string,
2547            it does the same, unless a different locale has been set for the match
2548            (using the /L modifier). In this case, it uses isprint() to decide.
2549    
2550     4. Fixed a major bug that caused incorrect computation of the amount of memory
2551        required for a compiled pattern when options that changed within the
2552        pattern affected the logic of the preliminary scan that determines the
2553        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2554        that the computed length was too small. The symptoms of this bug were
2555        either the PCRE error "internal error: code overflow" from pcre_compile(),
2556        or a glibc crash with a message such as "pcretest: free(): invalid next
2557        size (fast)". Examples of patterns that provoked this bug (shown in
2558        pcretest format) are:
2559    
2560          /(?-x: )/x
2561          /(?x)(?-x: \s*#\s*)/
2562          /((?i)[\x{c0}])/8
2563          /(?i:[\x{c0}])/8
2564    
2565        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2566        is now done differently.
2567    
2568     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2569        wrapper classes; (b) implement a new function in the C++ scanner that is
2570        more efficient than the old way of doing things because it avoids levels of
2571        recursion in the regex matching; (c) add a paragraph to the documentation
2572        for the FullMatch() function.
2573    
2574     6. The escape sequence \n was being treated as whatever was defined as
2575        "newline". Not only was this contrary to the documentation, which states
2576        that \n is character 10 (hex 0A), but it also went horribly wrong when
2577        "newline" was defined as CRLF. This has been fixed.
2578    
2579     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2580        was being set to -1 for the "end of line" case (supposedly a value that no
2581        character can have). Though this value is never used (the check for end of
2582        line is "zero bytes in current character"), it caused compiler complaints.
2583        I've changed it to 0xffffffff.
2584    
2585     8. In pcre_version.c, the version string was being built by a sequence of
2586        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2587        string (as it is for production releases) called a macro with an empty
2588        argument. The C standard says the result of this is undefined. The gcc
2589        compiler treats it as an empty string (which was what was wanted) but it is
2590        reported that Visual C gives an error. The source has been hacked around to
2591        avoid this problem.
2592    
2593     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2594        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2595        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2596        of them did).
2597    
2598    10. Originally, pcretest opened its input and output without "b"; then I was
2599        told that "b" was needed in some environments, so it was added for release
2600        5.0 to both the input and output. (It makes no difference on Unix-like
2601        systems.) Later I was told that it is wrong for the input on Windows. I've
2602        now abstracted the modes into two macros, to make it easier to fiddle with
2603        them, and removed "b" from the input mode under Windows.
2604    
2605    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2606    
2607    12. Added -help and --help to pcretest as an official way of being reminded
2608        of the options.
2609    
2610    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2611        and pcrecpp.cc because they annoy compilers at high warning levels.
2612    
2613    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2614    
2615    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2616        scripts are not C programs :-) and which was not noticed because it works
2617        on Linux.
2618    
2619    16. pcretest is supposed to handle any length of pattern and data line (as one
2620        line or as a continued sequence of lines) by extending its input buffer if
2621        necessary. This feature was broken for very long pattern lines, leading to
2622        a string of junk being passed to pcre_compile() if the pattern was longer
2623        than about 50K.
2624    
2625    17. I have done a major re-factoring of the way pcre_compile() computes the
2626        amount of memory needed for a compiled pattern. Previously, there was code
2627        that made a preliminary scan of the pattern in order to do this. That was
2628        OK when PCRE was new, but as the facilities have expanded, it has become
2629        harder and harder to keep it in step with the real compile phase, and there
2630        have been a number of bugs (see for example, 4 above). I have now found a
2631        cunning way of running the real compile function in a "fake" mode that
2632        enables it to compute how much memory it would need, while actually only
2633        ever using a few hundred bytes of working memory and without too many
2634        tests of the mode. This should make future maintenance and development
2635        easier. A side effect of this work is that the limit of 200 on the nesting
2636        depth of parentheses has been removed (though this was never a serious
2637        limitation, I suspect). However, there is a downside: pcre_compile() now
2638        runs more slowly than before (30% or more, depending on the pattern). I
2639        hope this isn't a big issue. There is no effect on runtime performance.
2640    
2641    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2642        newline (only possible for the last line of a file) and it was a
2643        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2644    
2645    19. Added additional timing features to pcretest. (1) The -tm option now times
2646        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2647        separate command line item, by a number that specifies the number of
2648        repeats to use when timing. The default is 50000; this gives better
2649        precision, but takes uncomfortably long for very large patterns.
2650    
2651    20. Extended pcre_study() to be more clever in cases where a branch of a
2652        subpattern has no definite first character. For example, (a*|b*)[cd] would
2653        previously give no result from pcre_study(). Now it recognizes that the
2654        first character must be a, b, c, or d.
2655    
2656    21. There was an incorrect error "recursive call could loop indefinitely" if
2657        a subpattern (or the entire pattern) that was being tested for matching an
2658        empty string contained only one non-empty item after a nested subpattern.
2659        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2660        incorrectly, because the \d was being skipped in the check.
2661    
2662    22. The pcretest program now has a new pattern option /B and a command line
2663        option -b, which is equivalent to adding /B to every pattern. This causes
2664        it to show the compiled bytecode, without the additional information that
2665        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2666        is the same as /B/I).
2667    
2668    23. A new optimization is now able automatically to treat some sequences such
2669        as a*b as a*+b. More specifically, if something simple (such as a character
2670        or a simple class like \d) has an unlimited quantifier, and is followed by
2671        something that cannot possibly match the quantified thing, the quantifier
2672        is automatically "possessified".
2673    
2674    24. A recursive reference to a subpattern whose number was greater than 39
2675        went wrong under certain circumstances in UTF-8 mode. This bug could also
2676        have affected the operation of pcre_study().
2677    
2678    25. Realized that a little bit of performance could be had by replacing
2679        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2680    
2681    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2682    
2683    27. Possessive quantifiers such as a++ were previously implemented by turning
2684        them into atomic groups such as ($>a+). Now they have their own opcodes,
2685        which improves performance. This includes the automatically created ones
2686        from 23 above.
2687    
2688    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2689        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2690        the first matched character to be a colon. This applied both to named and
2691        numbered groups.
2692    
2693    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2694    
2695    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2696        building PCRE on Windows easier, so I have included it in the distribution.
2697    
2698    31. There is now a check in pcretest against a ridiculously large number being
2699        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2700        loop, the loop is abandoned.
2701    
2702    32. Forward references to subpatterns in conditions such as (?(2)...) where
2703        subpattern 2 is defined later cause pcre_compile() to search forwards in
2704        the pattern for the relevant set of parentheses. This search went wrong
2705        when there were unescaped parentheses in a character class, parentheses
2706        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2707    
2708    33. "Subroutine" calls and backreferences were previously restricted to
2709        referencing subpatterns earlier in the regex. This restriction has now
2710        been removed.
2711    
2712    34. Added a number of extra features that are going to be in Perl 5.10. On the
2713        whole, these are just syntactic alternatives for features that PCRE had
2714        previously implemented using the Python syntax or my own invention. The
2715        other formats are all retained for compatibility.
2716    
2717        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2718            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2719            also .NET compatible.
2720    
2721        (b) A recursion or subroutine call to a named group can now be defined as
2722            (?&name) as well as (?P>name).
2723    
2724        (c) A backreference to a named group can now be defined as \k<name> or
2725            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2726            5.10, are also .NET compatible.
2727    
2728        (d) A conditional reference to a named group can now use the syntax
2729            (?(<name>) or (?('name') as well as (?(name).
2730    
2731        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2732            groups (named and numbered) that are never evaluated inline, but can be
2733            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2734            is always false. There may be only one alternative in such a group.
2735    
2736        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2737            as the simple (?(R). The condition is true only if the most recent
2738            recursion is that of the given number or name. It does not search out
2739            through the entire recursion stack.
2740    
2741        (g) The escape \gN or \g{N} has been added, where N is a positive or
2742            negative number, specifying an absolute or relative reference.
2743    
2744    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2745        some "unreachable code" warnings.
2746    
2747    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2748        things, this adds five new scripts.
2749    
2750    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2751        There were also incompatibilities regarding the handling of \Q..\E inside
2752        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2753        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2754    
2755    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2756        matches an empty string, and forcibly breaks the loop. There were bugs in
2757        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2758        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2759        separate and independent bugs (that affected different cases) have been
2760        fixed.
2761    
2762    39. Refactored the code to abolish the use of different opcodes for small
2763        capturing bracket numbers. This is a tidy that I avoided doing when I
2764        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2765        The new approach is not only tidier, it makes it possible to reduce the
2766        memory needed to fix the previous bug (38).
2767    
2768    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2769        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2770        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2771        mode.
2772    
2773    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2774        report.
2775    
2776    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2777        copy construction and assignment in the C++ wrapper.
2778    
2779    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2780        couple of bugs that could have given wrong results in the "--newline=crlf"
2781        case.
2782    
2783    44. Added a number of casts and did some reorganization of signed/unsigned int
2784        variables following suggestions from Dair Grant. Also renamed the variable
2785        "this" as "item" because it is a C++ keyword.
2786    
2787    45. Arranged for dftables to add
2788    
2789          #include "pcre_internal.h"
2790    
2791        to pcre_chartables.c because without it, gcc 4.x may remove the array
2792        definition from the final binary if PCRE is built into a static library and
2793        dead code stripping is activated.
2794    
2795    46. For an unanchored pattern, if a match attempt fails at the start of a
2796        newline sequence, and the newline setting is CRLF or ANY, and the next two
2797        characters are CRLF, advance by two characters instead of one.
2798    
2799    
2800    Version 6.7 04-Jul-06
2801    ---------------------
2802    
2803     1. In order to handle tests when input lines are enormously long, pcretest has
2804        been re-factored so that it automatically extends its buffers when
2805        necessary. The code is crude, but this _is_ just a test program. The
2806        default size has been increased from 32K to 50K.
2807    
2808     2. The code in pcre_study() was using the value of the re argument before
2809        testing it for NULL. (Of course, in any sensible call of the function, it
2810        won't be NULL.)
2811    
2812     3. The memmove() emulation function in pcre_internal.h, which is used on
2813        systems that lack both memmove() and bcopy() - that is, hardly ever -
2814        was missing a "static" storage class specifier.
2815    
2816     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2817        containing an extended class (one that cannot be represented by a bitmap
2818        because it contains high-valued characters or Unicode property items, e.g.
2819        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2820        pattern, but PCRE should not loop if you do not (it no longer does).
2821        [Detail: two cases were found: (a) a repeated subpattern containing an
2822        extended class; (b) a recursive reference to a subpattern that followed a
2823        previous extended class. It wasn't skipping over the extended class
2824        correctly when UTF-8 mode was not set.]
2825    
2826     5. A negated single-character class was not being recognized as fixed-length
2827        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2828        compile error "lookbehind assertion is not fixed length".
2829    
2830     6. The RunPerlTest auxiliary script was showing an unexpected difference
2831        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2832        write a Perl script that can interpret lines of an input file either as
2833        byte characters or as UTF-8, which is what "perltest" was being required to
2834        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2835        can't do is switch easily at run time between having the "use utf8;" pragma
2836        or not. In the end, I fudged it by using the RunPerlTest script to insert
2837        "use utf8;" explicitly for the UTF-8 tests.
2838    
2839     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2840        the end of the subject string, contrary to the documentation and to what
2841        Perl does. This was true of both matching functions. Now it matches only at
2842        the start of the subject and immediately after *internal* newlines.
2843    
2844     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2845        a pointer to an int instead of a pointer to an unsigned long int. This
2846        caused problems on 64-bit systems.
2847    
2848     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2849        instance of the 'standard' template library not being so standard".
2850    
2851    10. There was no check on the number of named subpatterns nor the maximum
2852        length of a subpattern name. The product of these values is used to compute
2853        the size of the memory block for a compiled pattern. By supplying a very
2854        long subpattern name and a large number of named subpatterns, the size
2855        computation could be caused to overflow. This is now prevented by limiting
2856        the length of names to 32 characters, and the number of named subpatterns
2857        to 10,000.
2858    
2859    11. Subpatterns that are repeated with specific counts have to be replicated in
2860        the compiled pattern. The size of memory for this was computed from the
2861        length of the subpattern and the repeat count. The latter is limited to
2862        65535, but there was no limit on the former, meaning that integer overflow
2863        could in principle occur. The compiled length of a repeated subpattern is
2864        now limited to 30,000 bytes in order to prevent this.
2865    
2866    12. Added the optional facility to have named substrings with the same name.
2867    
2868    13. Added the ability to use a named substring as a condition, using the
2869        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2870        are numbers (not recommended). Forward references are permitted.
2871    
2872    14. Added forward references in named backreferences (if you see what I mean).
2873    
2874    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2875        pattern could run off the end of the subject. For example, the pattern
2876        "(?s)(.{1,5})"8 did this with the subject "ab".
2877    
2878    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2879        PCRE_CASELESS was set when matching characters that were quantified with ?
2880        or *.
2881    
2882    17. A character class other than a single negated character that had a minimum
2883        but no maximum quantifier - for example [ab]{6,} - was not handled
2884        correctly by pce_dfa_exec(). It would match only one character.
2885    
2886    18. A valid (though odd) pattern that looked like a POSIX character
2887        class but used an invalid character after [ (for example [[,abc,]]) caused
2888        pcre_compile() to give the error "Failed: internal error: code overflow" or
2889        in some cases to crash with a glibc free() error. This could even happen if
2890        the pattern terminated after [[ but there just happened to be a sequence of
2891        letters, a binary zero, and a closing ] in the memory that followed.
2892    
2893    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2894        over the years. Originally (before any Unicode support), just the bottom 8
2895        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2896        output from "man perlunicode" includes this:
2897    
2898          The regular expression compiler produces polymorphic opcodes.  That
2899          is, the pattern adapts to the data and automatically switches to
2900          the Unicode character scheme when presented with Unicode data--or
2901          instead uses a traditional byte scheme when presented with byte
2902          data.
2903    
2904        Sadly, a wide octal escape does not cause a switch, and in a string with
2905        no other multibyte characters, these octal escapes are treated as before.
2906        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2907        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2908        Unicode string.
2909    
2910        I have not perpetrated such confusion in PCRE. Up till now, it took just
2911        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2912        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2913        translate to the appropriate multibyte character.
2914    
2915    29. Applied some refactoring to reduce the number of warnings from Microsoft
2916        and Borland compilers. This has included removing the fudge introduced
2917        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2918        a warning about an unused variable.
2919    
2920    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2921        characters since release 4.0, because Perl (from release 5.004) does not.
2922        [Or at least, is documented not to: some releases seem to be in conflict
2923        with the documentation.] However, when a pattern was studied with
2924        pcre_study() and all its branches started with \s, PCRE still included VT
2925        as a possible starting character. Of course, this did no harm; it just
2926        caused an unnecessary match attempt.
2927    
2928    22. Removed a now-redundant internal flag bit that recorded the fact that case
2929        dependency changed within the pattern. This was once needed for "required
2930        byte" processing, but is no longer used. This recovers a now-scarce options
2931        bit. Also moved the least significant internal flag bit to the most-
2932        significant bit of the word, which was not previously used (hangover from
2933        the days when it was an int rather than a uint) to free up another bit for
2934        the future.
2935    
2936    23. Added support for CRLF line endings as well as CR and LF. As well as the
2937        default being selectable at build time, it can now be changed at runtime
2938        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2939        specify that it is scanning data with non-default line endings.
2940    
2941    24. Changed the definition of CXXLINK to make it agree with the definition of
2942        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2943    
2944    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2945        recursions. This makes a big different to stack usage for some patterns.
2946    
2947    26. If a subpattern containing a named recursion or subroutine reference such
2948        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2949        the space required for the compiled pattern went wrong and gave too small a
2950        value. Depending on the environment, this could lead to "Failed: internal
2951        error: code overflow at offset 49" or "glibc detected double free or
2952        corruption" errors.
2953    
2954    27. Applied patches from Google (a) to support the new newline modes and (b) to
2955        advance over multibyte UTF-8 characters in GlobalReplace.
2956    
2957    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2958        difference for some implementation of PCRE in some Windows version.
2959    
2960    29. Added some extra testing facilities to pcretest:
2961    
2962        \q<number>   in a data line sets the "match limit" value
2963        \Q<number>   in a data line sets the "match recursion limt" value
2964        -S <number>  sets the stack size, where <number> is in megabytes
2965    
2966        The -S option isn't available for Windows.
2967    
2968    
2969    Version 6.6 06-Feb-06
2970    ---------------------
2971    
2972     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2973        in pcreposix.h. I have copied the definition from pcre.h.
2974    
2975     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2976        because pcre.h is no longer a built file.
2977    
2978     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2979        not normally included in the compiled code.
2980    
2981    
2982    Version 6.5 01-Feb-06
2983    ---------------------
2984    
2985     1. When using the partial match feature with pcre_dfa_exec(), it was not
2986        anchoring the second and subsequent partial matches at the new starting
2987        point. This could lead to incorrect results. For example, with the pattern
2988        /1234/, partially matching against "123" and then "a4" gave a match.
2989    
2990     2. Changes to pcregrep:
2991    
2992        (a) All non-match returns from pcre_exec() were being treated as failures
2993            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2994            error message is output. Some extra information is given for the
2995            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2996            probably the only errors that are likely to be caused by users (by
2997            specifying a regex that has nested indefinite repeats, for instance).
2998            If there are more than 20 of these errors, pcregrep is abandoned.
2999    
3000        (b) A binary zero was treated as data while matching, but terminated the
3001            output line if it was written out. This has been fixed: binary zeroes
3002            are now no different to any other data bytes.
3003    
3004        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3005            used to set a locale for matching. The --locale=xxxx long option has
3006            been added (no short equivalent) to specify a locale explicitly on the
3007            pcregrep command, overriding the environment variables.
3008    
3009        (d) When -B was used with -n, some line numbers in the output were one less
3010            than they should have been.
3011    
3012        (e) Added the -o (--only-matching) option.
3013    
3014        (f) If -A or -C was used with -c (count only), some lines of context were
3015            accidentally printed for the final match.
3016    
3017        (g) Added the -H (--with-filename) option.
3018    
3019        (h) The combination of options -rh failed to suppress file names for files
3020            that were found from directory arguments.
3021    
3022        (i) Added the -D (--devices) and -d (--directories) options.
3023    
3024        (j) Added the -F (--fixed-strings) option.
3025    
3026        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3027    
3028        (l) Added the --colo(u)r option.
3029    
3030        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3031            is not present by default.
3032    
3033     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3034        items such as (?R) or (?1), when the recursion could match a number of
3035        alternatives. If it matched one of the alternatives, but subsequently,
3036        outside the recursion, there was a failure, the code tried to back up into
3037        the recursion. However, because of the way PCRE is implemented, this is not
3038        possible, and the result was an incorrect result from the match.
3039    
3040        In order to prevent this happening, the specification of recursion has
3041        been changed so that all such subpatterns are automatically treated as
3042        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3043    
3044     4. I had overlooked the fact that, in some locales, there are characters for
3045        which isalpha() is true but neither isupper() nor islower() are true. In
3046        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3047        and ordfeminine) are like this. This affected the treatment of \w and \W
3048        when they appeared in character classes, but not when they appeared outside
3049        a character class. The bit map for "word" characters is now created
3050        separately from the results of isalnum() instead of just taking it from the
3051        upper, lower, and digit maps. (Plus the underscore character, of course.)
3052    
3053     5. The above bug also affected the handling of POSIX character classes such as
3054        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3055        permanent tables. Instead, the bit maps for such a class were previously
3056        created as the appropriate unions of the upper, lower, and digit bitmaps.
3057        Now they are created by subtraction from the [[:word:]] class, which has
3058        its own bitmap.
3059    
3060     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3061        It is created by subtracting the vertical space characters (\x09, \x0a,
3062        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3063        subtraction was done in the overall bitmap for a character class, meaning
3064        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3065        be recognized. This bug has been fixed.
3066    
3067     7. Patches from the folks at Google:
3068    
3069          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3070          real life, but is still worth protecting against".
3071    
3072          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3073          regular expressions".
3074    
3075          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3076          have it.
3077    
3078          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3079          "configure" and the latter not, in order to fix a problem somebody had
3080          with compiling the Arg class on HP-UX.
3081    
3082          (e) Improve the error-handling of the C++ wrapper a little bit.
3083    
3084          (f) New tests for checking recursion limiting.
3085    
3086     8. The pcre_memmove() function, which is used only if the environment does not
3087        have a standard memmove() function (and is therefore rarely compiled),
3088        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3089        returning a result (though PCRE never actually uses the result).
3090    
3091     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3092        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3093        returned instead of calling malloc() with an overflowing number that would
3094        most likely cause subsequent chaos.
3095    
3096    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3097    
3098    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3099        with this option is matched, the nmatch and pmatch options of regexec() are
3100        ignored.
3101    
3102    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3103        provided in case anyone wants to the the POSIX interface with UTF-8
3104        strings.
3105    
3106    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3107        C++ linking (needed for some HP-UX environments).
3108    
3109    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3110        (unused parameter) and in the pcre_printint() function (omitted "default"
3111        switch label when the default is to do nothing).
3112    
3113    15. Added some code to make it possible, when PCRE is compiled as a C++
3114        library, to replace subject pointers for pcre_exec() with a smart pointer
3115        class, thus making it possible to process discontinuous strings.
3116    
3117    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3118        much the same function. They were added by different people who were trying
3119        to make PCRE easy to compile on non-Unix systems. It has been suggested
3120        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3121        for compiling on Windows systems. I have therefore replaced it with
3122        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3123        defaults to "extern" for C or "extern C" for C++, which works fine on
3124        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3125        SCOPE with something explicit in config.h. In addition:
3126    
3127        (a) pcreposix.h still had just "extern" instead of either of these macros;
3128            I have replaced it with PCRE_DATA_SCOPE.
3129    
3130        (b) Functions such as _pcre_xclass(), which are internal to the library,
3131            but external in the C sense, all had PCRE_EXPORT in their definitions.
3132            This is apparently wrong for the Windows case, so I have removed it.
3133            (It makes no difference on Unix-like systems.)
3134    
3135    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3136        of recursive calls to match(). This is different to MATCH_LIMIT because
3137        that limits the total number of calls to match(), not all of which increase
3138        the depth of recursion. Limiting the recursion depth limits the amount of
3139        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3140        when PCRE is compiled, and changed at run time. A patch from Google adds
3141        this functionality to the C++ interface.
3142    
3143    18. Changes to the handling of Unicode character properties:
3144    
3145        (a) Updated the table to Unicode 4.1.0.
3146    
3147        (b) Recognize characters that are not in the table as "Cn" (undefined).
3148    
3149        (c) I revised the way the table is implemented to a much improved format
3150            which includes recognition of ranges. It now supports the ranges that
3151            are defined in UnicodeData.txt, and it also amalgamates other
3152            characters into ranges. This has reduced the number of entries in the
3153            table from around 16,000 to around 3,000, thus reducing its size
3154            considerably. I realized I did not need to use a tree structure after
3155            all - a binary chop search is just as efficient. Having reduced the
3156            number of entries, I extended their size from 6 bytes to 8 bytes to
3157            allow for more data.
3158    
3159        (d) Added support for Unicode script names via properties such as \p{Han}.
3160    
3161    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3162        matching that character.
3163    
3164    20. When matching a repeated Unicode property with a minimum greater than zero,
3165        (for example \pL{2,}), PCRE could look past the end of the subject if it
3166        reached it while seeking the minimum number of characters. This could
3167        happen only if some of the characters were more than one byte long, because
3168        there is a check for at least the minimum number of bytes.
3169    
3170    21. Refactored the implementation of \p and \P so as to be more general, to
3171        allow for more different types of property in future. This has changed the
3172        compiled form incompatibly. Anybody with saved compiled patterns that use
3173        \p or \P will have to recompile them.
3174    
3175    22. Added "Any" and "L&" to the supported property types.
3176    
3177    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3178        but give a compile time error if the value is greater than 0xff.
3179    
3180    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3181        accidentally not being installed or uninstalled.
3182    
3183    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3184        made were to insert the current release number. This seemed silly, because
3185        it made things harder for people building PCRE on systems that don't run
3186        "configure". I have turned pcre.h into a distributed file, no longer built
3187        by "configure", with the version identification directly included. There is
3188        no longer a pcre.h.in file.
3189    
3190        However, this change necessitated a change to the pcre-config script as
3191        well. It is built from pcre-config.in, and one of the substitutions was the
3192        release number. I have updated configure.ac so that ./configure now finds
3193        the release number by grepping pcre.h.
3194    
3195    26. Added the ability to run the tests under valgrind.
3196    
3197    
3198    Version 6.4 05-Sep-05
3199    ---------------------
3200    
3201     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3202        "--" to be printed when multiple files were scanned, even when none of the
3203        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3204        consider it to be a bug, and have restored the previous behaviour.
3205    
3206     2. A couple of code tidies to get rid of compiler warnings.
3207    
3208     3. The pcretest program used to cheat by referring to symbols in the library
3209        whose names begin with _pcre_. These are internal symbols that are not
3210        really supposed to be visible externally, and in some environments it is
3211        possible to suppress them. The cheating is now confined to including
3212        certain files from the library's source, which is a bit cleaner.
3213    
3214     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3215        file's purpose clearer.
3216    
3217     5. Reorganized pcre_ucp_findchar().
3218    
3219    
3220    Version 6.3 15-Aug-05
3221    ---------------------
3222    
3223     1. The file libpcre.pc.in did not have general read permission in the tarball.
3224    
3225     2. There were some problems when building without C++ support:
3226    
3227        (a) If C++ support was not built, "make install" and "make test" still
3228            tried to test it.
3229    
3230        (b) There were problems when the value of CXX was explicitly set. Some
3231            changes have been made to try to fix these, and ...
3232    
3233        (c) --disable-cpp can now be used to explicitly disable C++ support.
3234    
3235        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3236            backslash in a target when C++ was disabled. This confuses some
3237            versions of "make", apparently. Using an intermediate variable solves
3238            this. (Same for CPP_LOBJ.)
3239    
3240     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3241        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3242        necessary on certain architectures.
3243    
3244     4. Added a setting of -export-symbols-regex to the link command to remove
3245        those symbols that are exported in the C sense, but actually are local
3246        within the library, and not documented. Their names all begin with
3247        "_pcre_". This is not a perfect job, because (a) we have to except some
3248        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3249        available (and never for static libraries). I have made a note to try to
3250        find a way round (a) in the future.
3251    
3252    
3253    Version 6.2 01-Aug-05
3254    ---------------------
3255    
3256     1. There was no test for integer overflow of quantifier values. A construction
3257        such as {1111111111111111} would give undefined results. What is worse, if
3258        a minimum quantifier for a parenthesized subpattern overflowed and became
3259        negative, the calculation of the memory size went wrong. This could have
3260        led to memory overwriting.
3261    
3262     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3263    
3264     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3265        operating environments where this matters.
3266    
3267     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3268        PCRE options from within the C++ wrapper.
3269    
3270     5. Named capturing subpatterns were not being correctly counted when a pattern
3271        was compiled. This caused two problems: (a) If there were more than 100
3272        such subpatterns, the calculation of the memory needed for the whole
3273        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3274        back references of the form \12, where the number was greater than 9, were
3275        not recognized as back references, even though there were sufficient
3276        previous subpatterns.
3277    
3278     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3279        versions of gcc, e.g. 2.95.4.
3280    
3281    
3282    Version 6.1 21-Jun-05
3283    ---------------------
3284    
3285     1. There was one reference to the variable "posix" in pcretest.c that was not
3286        surrounded by "#if !defined NOPOSIX".
3287    
3288     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3289        the cross-check on the old pcre_info() function, for the benefit of the
3290        cut-down version of PCRE that is currently imported into Exim.
3291    
3292     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3293        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3294        patterns that start (?i) but I don't think that matters. The use of (?i) is
3295        just an example; this all applies to the other options as well.
3296    
3297     4. Since libtool seems to echo the compile commands it is issuing, the output
3298        from "make" can be reduced a bit by putting "@" in front of each libtool
3299        compile command.
3300    
3301     5. Patch from the folks at Google for configure.in to be a bit more thorough
3302        in checking for a suitable C++ installation before trying to compile the
3303        C++ stuff. This should fix a reported problem when a compiler was present,
3304        but no suitable headers.
3305    
3306     6. The man pages all had just "PCRE" as their title. I have changed them to
3307        be the relevant file name. I have also arranged that these names are
3308        retained in the file doc/pcre.txt, which is a concatenation in text format
3309        of all the man pages except the little individual ones for each function.
3310    
3311     7. The NON-UNIX-USE file had not been updated for the different set of source
3312        files that come with release 6. I also added a few comments about the C++
3313        wrapper.
3314    
3315    
3316    Version 6.0 07-Jun-05
3317    ---------------------
3318    
3319     1. Some minor internal re-organization to help with my DFA experiments.
3320    
3321     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3322        didn't matter for the library itself when fully configured, but did matter
3323        when compiling without UCP support, or within Exim, where the ucp files are
3324        not imported.
3325    
3326     3. Refactoring of the library code to split up the various functions into
3327        different source modules. The addition of the new DFA matching code (see
3328        below) to a single monolithic source would have made it really too
3329        unwieldy, quite apart from causing all the code to be include in a
3330        statically linked application, when only some functions are used. This is
3331        relevant even without the DFA addition now that patterns can be compiled in
3332        one application and matched in another.
3333    
3334        The downside of splitting up is that there have to be some external
3335        functions and data tables that are used internally in different modules of
3336        the library but which are not part of the API. These have all had their
3337        names changed to start with "_pcre_" so that they are unlikely to clash
3338        with other external names.
3339    
3340     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3341        a different (DFA) algorithm. Although it is slower than the original
3342        function, it does have some advantages for certain types of matching
3343        problem.
3344    
3345     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3346        including restarting after a partial match.
3347    
3348     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3349        defined when compiling for Windows was sent to me. I have put it into the
3350        code, though I have no means of testing or verifying it.
3351    
3352     7. Added the pcre_refcount() auxiliary function.
3353    
3354     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3355        match before or at the first newline in the subject string. In pcretest,
3356        the /f option on a pattern can be used to set this.
3357    
3358     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3359        would behave wrongly. This has been present in PCRE since release 4.0.
3360    
3361    10. A number of changes to the pcregrep command:
3362    
3363        (a) Refactored how -x works; insert ^(...)$ instead of setting
3364            PCRE_ANCHORED and checking the length, in preparation for adding
3365            something similar for -w.
3366    
3367        (b) Added the -w (match as a word) option.
3368    
3369        (c) Refactored the way lines are read and buffered so as to have more
3370            than one at a time available.
3371    
3372        (d) Implemented a pcregrep test script.
3373    
3374        (e) Added the -M (multiline match) option. This allows patterns to match
3375            over several lines of the subject. The buffering ensures that at least
3376            8K, or the rest of the document (whichever is the shorter) is available
3377            for matching (and similarly the previous 8K for lookbehind assertions).
3378    
3379        (f) Changed the --help output so that it now says
3380    
3381              -w, --word-regex(p)
3382    
3383            instead of two lines, one with "regex" and the other with "regexp"
3384            because that confused at least one person since the short forms are the
3385            same. (This required a bit of code, as the output is generated
3386            automatically from a table. It wasn't just a text change.)
3387    
3388        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3389            option but starts with a hyphen. Could be a pattern or a path name
3390            starting with a hyphen, for instance.
3391    
3392        (h) "-" can be given as a file name to represent stdin.
3393    
3394        (i) When file names are being printed, "(standard input)" is used for
3395            the standard input, for compatibility with GNU grep. Previously
3396            "<stdin>" was used.
3397    
3398        (j) The option --label=xxx can be used to supply a name to be used for
3399            stdin when file names are being printed. There is no short form.
3400    
3401        (k) Re-factored the options decoding logic because we are going to add
3402            two more options that take data. Such options can now be given in four
3403            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3404    
3405        (l) Added the -A, -B, and -C options for requesting that lines of context
3406            around matches be printed.
3407    
3408        (m) Added the -L option to print the names of files that do not contain
3409            any matching lines, that is, the complement of -l.
3410    
3411        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3412            continue to scan other files.
3413    
3414        (o) The -s option was incorrectly implemented. For compatibility with other
3415            greps, it now suppresses the error message for a non-existent or non-
3416            accessible file (but not the return code). There is a new option called
3417            -q that suppresses the output of matching lines, which was what -s was
3418            previously doing.
3419    
3420        (p) Added --include and --exclude options to specify files for inclusion
3421            and exclusion when recursing.
3422    
3423    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3424        Hopefully, it now does.
3425    
3426    12. Missing cast in pcre_study().
3427    
3428    13. Added an "uninstall" target to the makefile.
3429    
3430    14. Replaced "extern" in the function prototypes in Makefile.in with
3431        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3432        world, but is set differently for Windows.
3433    
3434    15. Added a second compiling function called pcre_compile2(). The only
3435        difference is that it has an extra argument, which is a pointer to an
3436        integer error code. When there is a compile-time failure, this is set
3437        non-zero, in addition to the error test pointer being set to point to an
3438        error message. The new argument may be NULL if no error number is required
3439        (but then you may as well call pcre_compile(), which is now just a
3440        wrapper). This facility is provided because some applications need a
3441        numeric error indication, but it has also enabled me to tidy up the way
3442        compile-time errors are handled in the POSIX wrapper.
3443    
3444    16. Added VPATH=.libs to the makefile; this should help when building with one
3445        prefix path and installing with another. (Or so I'm told by someone who
3446        knows more about this stuff than I do.)
3447    
3448    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3449        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3450        match everything, including newlines. This is not POSIX-compatible, but
3451        somebody wanted the feature. From pcretest it can be activated by using
3452        both the P and the s flags.
3453    
3454    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3455    
3456    19. libpcre.pc was being incorrectly installed as executable.
3457    
3458    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3459        it now also looks for '\r' so that it will work unmodified on Windows.
3460    
3461    21. Added Google's contributed C++ wrapper to the distribution.
3462    
3463    22. Added some untidy missing memory free() calls in pcretest, to keep
3464        Electric Fence happy when testing.
3465    
3466    
3467    
3468    Version 5.0 13-Sep-04
3469    ---------------------
3470    
3471     1. Internal change: literal characters are no longer packed up into items
3472        containing multiple characters in a single byte-string. Each character
3473        is now matched using a separate opcode. However, there may be more than one
3474        byte in the character in UTF-8 mode.
3475    
3476     2. The pcre_callout_block structure has two new fields: pattern_position and
3477        next_item_length. These contain the offset in the pattern to the next match
3478        item, and its length, respectively.
3479    
3480     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3481        insertion of callouts before each pattern item. Added the /C option to
3482        pcretest to make use of this.
3483    
3484     4. On the advice of a Windows user, the lines
3485    
3486          #if defined(_WIN32) || defined(WIN32)
3487          _setmode( _fileno( stdout ), 0x8000 );
3488          #endif  /* defined(_WIN32) || defined(WIN32) */
3489    
3490        have been added to the source of pcretest. This apparently does useful
3491        magic in relation to line terminators.
3492    
3493     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3494        for the benefit of those environments where the "b" makes a difference.
3495    
3496     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3497        to know about it. I have put a hack into configure.in that adds in code
3498        to set GCC=yes if CC=icc. This seems to end up at a point in the
3499        generated configure script that is early enough to affect the setting of
3500        compiler options, which is what is needed, but I have no means of testing
3501        whether it really works. (The user who reported this had patched the
3502        generated configure script, which of course I cannot do.)
3503    
3504        LATER: After change 22 below (new libtool files), the configure script
3505        seems to know about icc (and also ecc). Therefore, I have commented out
3506        this hack in configure.in.
3507    
3508     7. Added support for pkg-config (2 patches were sent in).
3509    
3510     8. Negated POSIX character classes that used a combination of internal tables
3511        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3512        [[:^ascii]]. Typically, they would match almost any characters. The other
3513        POSIX classes were not broken in this way.
3514    
3515     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3516        to find the match, as PCRE was deluded into thinking that the match had to
3517        start at the start point or following a newline. The same bug applied to
3518        patterns with negative forward assertions or any backward assertions
3519        preceding ".*" at the start, unless the pattern required a fixed first
3520        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3521    
3522    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3523        starting at the last subject character, bytes beyond the end of the subject
3524        string were read.
3525    
3526    11. Renamed the variable "class" as "classbits" to make life easier for C++
3527        users. (Previously there was a macro definition, but it apparently wasn't
3528        enough.)
3529    
3530    12. Added the new field "tables" to the extra data so that tables can be passed
3531        in at exec time, or the internal tables can be re-selected. This allows
3532        a compiled regex to be saved and re-used at a later time by a different
3533        program that might have everything at different addresses.
3534    
3535    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3536        -R library as well as a -L library.
3537    
3538    14. The debugging options of pcretest (-d on the command line or D on a
3539        pattern) showed incorrect output for anything following an extended class
3540        that contained multibyte characters and which was followed by a quantifier.
3541    
3542    15. Added optional support for general category Unicode character properties
3543        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3544        support. It adds about 90K to the size of the library. The meanings of the
3545        inbuilt class escapes such as \d and \s have NOT been changed.
3546    
3547    16. Updated pcredemo.c to include calls to free() to release the memory for the
3548        compiled pattern.
3549    
3550    17. The generated file chartables.c was being created in the source directory
3551        instead of in the building directory. This caused the build to fail if the
3552        source directory was different from the building directory, and was
3553        read-only.
3554    
3555    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3556        file. No doubt somebody will tell me if they don't make sense... Also added
3557        Dan Mooney's comments about building on OpenVMS.
3558    
3559    19. Added support for partial matching via the PCRE_PARTIAL option for
3560        pcre_exec() and the \P data escape in pcretest.
3561    
3562    20. Extended pcretest with 3 new pattern features:
3563    
3564        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3565              write the compiled pattern to the file whose name is "rest-of-line".
3566              This is a straight binary dump of the data, with the saved pointer to
3567              the character tables forced to be NULL. The study data, if any, is
3568              written too. After writing, pcretest reads a new pattern.
3569    
3570        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3571              compiled pattern from the given file. There must not be any
3572              occurrences of "<" in the file name (pretty unlikely); if there are,
3573              pcretest will instead treat the initial "<" as a pattern delimiter.
3574              After reading in the pattern, pcretest goes on to read data lines as
3575              usual.
3576    
3577        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3578              and 16-bit fields in a compiled pattern, to simulate a pattern that
3579              was compiled on a host of opposite endianness.
3580    
3581    21. The pcre-exec() function can now cope with patterns that were compiled on
3582        hosts of opposite endianness, with this restriction:
3583    
3584          As for any compiled expression that is saved and used later, the tables
3585          pointer field cannot be preserved; the extra_data field in the arguments
3586          to pcre_exec() should be used to pass in a tables address if a value
3587          other than the default internal tables were used at compile time.
3588    
3589    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3590        now diagnosed as an error. Previously, most of the time, a negative number
3591        would have been treated as zero, but if in addition "ovector" was passed as
3592        NULL, a crash could occur.
3593    
3594    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3595        new versions from the libtool 1.5 distribution (the last one is a copy of
3596        a file called libtool.m4). This seems to have fixed the need to patch
3597        "configure" to support Darwin 1.3 (which I used to do). However, I still
3598        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3599        workstation).
3600    
3601    24. Changed the PCRE licence to be the more standard "BSD" licence.
3602    
3603    
3604    Version 4.5 01-Dec-03
3605    ---------------------
3606    
3607     1. There has been some re-arrangement of the code for the match() function so
3608        that it can be compiled in a version that does not call itself recursively.
3609        Instead, it keeps those local variables that need separate instances for
3610        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3611        needs to "recurse". Keeping track of where control must go is done by means
3612        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3613        hide most of the details from the main code, and operates only if
3614        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3615        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3616        operating.
3617    
3618        To make it easier for callers to provide specially tailored get/free
3619        functions for this usage, two new functions, pcre_stack_malloc, and
3620        pcre_stack_free, are used. They are always called in strict stacking order,
3621        and the size of block requested is always the same.
3622    
3623        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3624        PCRE has been compiled to use the stack or the heap for recursion. The
3625        -C option of pcretest uses this to show which version is compiled.
3626    
3627        A new data escape \S, is added to pcretest; it causes the amounts of store
3628        obtained and freed by both kinds of malloc/free at match time to be added
3629        to the output.
3630    
3631     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3632        what's available on my current Linux desktop machine.
3633    
3634     3. When matching a UTF-8 string, the test for a valid string at the start has
3635        been extended. If start_offset is not zero, PCRE now checks that it points
3636        to a byte that is the start of a UTF-8 character. If not, it returns
3637        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3638        this is necessary because there may be backward assertions in the pattern.
3639        When matching the same subject several times, it may save resources to use
3640        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3641    
3642     4. The code for checking the validity of UTF-8 strings has been tightened so
3643        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3644        containing "overlong sequences".
3645    
3646     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3647        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3648        should have been "&", but it just so happened that all the cases this let
3649        through by mistake were picked up later in the function.
3650    
3651     6. I had used a variable called "isblank" - this is a C99 function, causing
3652        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3653    
3654     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3655        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3656        so the version is shown; (c) stop "make test" echoing "./RunTest".
3657    
3658     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3659    
3660     9. The prototype for memmove() for systems that don't have it was using
3661        size_t, but the inclusion of the header that defines size_t was later. I've
3662        moved the #includes for the C headers earlier to avoid this.
3663    
3664    10. Added some adjustments to the code to make it easier to compiler on certain
3665        special systems:
3666    
3667          (a) Some "const" qualifiers were missing.
3668          (b) Added the macro EXPORT before all exported functions; by default this
3669              is defined to be empty.
3670          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3671              that it reads its output file name as an argument instead of writing
3672              to the standard output and assuming this can be redirected.
3673    
3674    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3675        class containing characters with values greater than 255, PCRE compilation
3676        went into a loop.
3677    
3678    12. A recursive reference to a subpattern that was within another subpattern
3679        that had a minimum quantifier of zero caused PCRE to crash. For example,
3680        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3681        recursion. If the recursively-called subpattern itself had a zero repeat,
3682        that was OK.
3683    
3684    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3685        buffer into which it was copied (for escape processing) was still set at
3686        1024, so long lines caused crashes.
3687    
3688    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3689        "internal error: code overflow...". This applied to any character class
3690        that was followed by a possessive quantifier.
3691    
3692    15. Modified the Makefile to add libpcre.la as a prerequisite for
3693        libpcreposix.la because I was told this is needed for a parallel build to
3694        work.
3695    
3696    16. If a pattern that contained .* following optional items at the start was
3697        studied, the wrong optimizing data was generated, leading to matching
3698        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3699        matching string must start with a or b or c. The correct conclusion for
3700        this pattern is that a match can start with any character.
3701    
3702    
3703    Version 4.4 13-Aug-03
3704    ---------------------
3705    
3706     1. In UTF-8 mode, a character class containing characters with values between
3707        127 and 255 was not handled correctly if the compiled pattern was studied.
3708        In fixing this, I have also improved the studying algorithm for such
3709        classes (slightly).
3710    
3711     2. Three internal functions had redundant arguments passed to them. Removal
3712        might give a very teeny performance improvement.
3713    
3714     3. Documentation bug: the value of the capture_top field in a callout is *one
3715        more than* the number of the hightest numbered captured substring.
3716    
3717     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3718        in incorrectly linking with a previously installed version. They now link
3719        explicitly with libpcre.la.
3720    
3721     5. configure.in no longer needs to recognize Cygwin specially.
3722    
3723     6. A problem in pcre.in for Windows platforms is fixed.
3724    
3725     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3726        pcretest, it used to include the size of the study block as part of its
3727        output. Unfortunately, the structure contains a field that has a different
3728        size on different hardware architectures. This meant that the tests that
3729        showed this size failed. As the block is currently always of a fixed size,
3730        this information isn't actually particularly useful in pcretest output, so
3731        I have just removed it.
3732    
3733     8. Three pre-processor statements accidentally did not start in column 1.
3734        Sadly, there are *still* compilers around that complain, even though
3735        standard C has not required this for well over a decade. Sigh.
3736    
3737     9. In pcretest, the code for checking callouts passed small integers in the
3738        callout_data field, which is a void * field. However, some picky compilers
3739        complained about the casts involved for this on 64-bit systems. Now
3740        pcretest passes the address of the small integer instead, which should get
3741        rid of the warnings.
3742    
3743    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3744        both compile and run time, and gives an error if an invalid UTF-8 sequence
3745        is found. There is a option for disabling this check in cases where the
3746        string is known to be correct and/or the maximum performance is wanted.
3747    
3748    11. In response to a bug report, I changed one line in Makefile.in from
3749    
3750            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3751        to
3752            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3753    
3754        to look similar to other lines, but I have no way of telling whether this
3755        is the right thing to do, as I do not use Windows. No doubt I'll get told
3756        if it's wrong...
3757    
3758    
3759    Version 4.3 21-May-03
3760    ---------------------
3761    
3762    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3763       Makefile.
3764    
3765    2. Some refactoring to improve the quality of the code:
3766    
3767       (i)   The utf8_table... variables are now declared "const".
3768    
3769       (ii)  The code for \cx, which used the "case flipping" table to upper case
3770             lower case letters, now just substracts 32. This is ASCII-specific,
3771             but the whole concept of \cx is ASCII-specific, so it seems
3772             reasonable.
3773    
3774       (iii) PCRE was using its character types table to recognize decimal and
3775             hexadecimal digits in the pattern. This is silly, because it handles
3776             only 0-9, a-f, and A-F, but the character types table is locale-
3777             specific, which means strange things might happen. A private
3778             table is now used for this - though it costs 256 bytes, a table is
3779             much faster than multiple explicit tests. Of course, the standard
3780             character types table is still used for matching digits in subject
3781             strings against \d.
3782    
3783       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3784             ending in _t are). So I've renamed it as ESC_tee.
3785    
3786    3. The first argument for regexec() in the POSIX wrapper should have been
3787       defined as "const".