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

Diff of /code/trunk/ChangeLog

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

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