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

Diff of /code/trunk/ChangeLog

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

revision 193 by ph10, Mon Jul 30 11:34:18 2007 UTC revision 1359 by ph10, Tue Sep 3 10:10:59 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-07  Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    3.  Make a small performance improvement in strlen16() and strlen32() in
19        pcretest.
20    
21    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23        these cases). There was also one in pcretest.c.
24    
25    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26    
27    6.  In UTF mode, the code for checking whether a group could match an empty
28        string (which is used for indefinitely repeated groups to allow for
29        breaking an infinite loop) was broken when the group contained a repeated
30        negated single-character class with a character that occupied more than one
31        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32        UTF-8 mode). The effect was undefined: the group might or might not be
33        deemed as matching an empty string, or the program might have crashed.
34    
35    7.  The code for checking whether a group could match an empty string was not
36        recognizing that \h, \H, \v, \V, and \R must match a character.
37    
38    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39        an empty string. If it can, pcretest shows this in its information output.
40    
41    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
42        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43        matched by pcre_exec() without using JIT:
44    
45        (a) If the rest of the pattern did not match after a maximal run of
46            grapheme clusters, the code for backing up to try with fewer of them
47            did not always back up over a full grapheme when characters that do not
48            have the modifier quality were involved, e.g. Hangul syllables.
49    
50        (b) If the match point in a subject started with modifier character, and
51            there was no match, the code could incorrectly back up beyond the match
52            point, and potentially beyond the first character in the subject,
53            leading to a segfault or an incorrect match result.
54    
55    10. A conditional group with an assertion condition could lead to PCRE
56        recording an incorrect first data item for a match if no other first data
57        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58        first data item, and therefore matched "ca" after "c" instead of at the
59        start.
60    
61    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63    
64    12. The source of pcregrep now includes z/OS-specific code so that it can be
65        compiled for z/OS as part of the special z/OS distribution.
66    
67    13. Added the -T and -TM options to pcretest.
68    
69    14. The code in pcre_compile.c for creating the table of named capturing groups
70        has been refactored. Instead of creating the table dynamically during the
71        actual compiling pass, the information is remembered during the pre-compile
72        pass (on the stack unless there are more than 20 named groups, in which
73        case malloc() is used) and the whole table is created before the actual
74        compile happens. This has simplified the code (it is now nearly 150 lines
75        shorter) and prepared the way for better handling of references to groups
76        with duplicate names.
77    
78    
79    Version 8.33 28-May-2013
80    ------------------------
81    
82    1.  Added 'U' to some constants that are compared to unsigned integers, to
83        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
84        variables that are added to signed variables, to ensure the result is
85        signed and can be negated.
86    
87    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
88        CMake config files.
89    
90    3.  Revise the creation of config.h.generic so that all boolean macros are
91        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
92        overriding via -D on the command line possible.
93    
94    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
95        to unsigned int is reported to make a quite noticeable speed difference in
96        a specific Windows environment. Testing on Linux did also appear to show
97        some benefit (and it is clearly not harmful). Also fixed the definition of
98        Xop which should be unsigned.
99    
100    5.  Related to (4), changing the definition of the intermediate variable cc
101        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
102        improvements.
103    
104    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
105        unnecessary spaces.
106    
107    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
108        and later.
109    
110    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
111    
112    9.  Optimizing fast_forward_start_bits in JIT.
113    
114    10. Adding support for callouts in JIT, and fixing some issues revealed
115        during this work. Namely:
116    
117        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
118    
119        (b) Minimum length was not checked before the matching is started.
120    
121    11. The value of capture_last that is passed to callouts was incorrect in some
122        cases when there was a capture on one path that was subsequently abandoned
123        after a backtrack. Also, the capture_last value is now reset after a
124        recursion, since all captures are also reset in this case.
125    
126    12. The interpreter no longer returns the "too many substrings" error in the
127        case when an overflowing capture is in a branch that is subsequently
128        abandoned after a backtrack.
129    
130    13. In the pathological case when an offset vector of size 2 is used, pcretest
131        now prints out the matched string after a yield of 0 or 1.
132    
133    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
134        Only supported by the JIT compiler at the moment.
135    
136    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
137    
138    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
139        offset of the starting point of the matching process, provided the offsets
140        vector is large enough.
141    
142    17. The \A escape now records a lookbehind value of 1, though its execution
143        does not actually inspect the previous character. This is to ensure that,
144        in partial multi-segment matching, at least one character from the old
145        segment is retained when a new segment is processed. Otherwise, if there
146        are no lookbehinds in the pattern, \A might match incorrectly at the start
147        of a new segment.
148    
149    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
150    
151    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
152        gives some modest performance improvement in 8-bit mode.
153    
154    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
155        be expressed in certain programming languages using Universal Character
156        Names.
157    
158    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
159        which points out that "non characters" are not "characters that may not
160        appear in Unicode strings" but rather "characters that are reserved for
161        internal use and have only local meaning".
162    
163    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
164        there was a conditional group that depended on an assertion, if the
165        assertion was false, the callout that immediately followed the alternation
166        in the condition was skipped when pcre_exec() was used for matching.
167    
168    23. Allow an explicit callout to be inserted before an assertion that is the
169        condition for a conditional group, for compatibility with automatic
170        callouts, which always insert a callout at this point.
171    
172    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
173        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
174    
175    25. (*PRUNE) is now supported by the JIT compiler.
176    
177    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
178    
179    27. Fix the case where there are two or more SKIPs with arguments that may be
180        ignored.
181    
182    28. (*SKIP) is now supported by the JIT compiler.
183    
184    29. (*THEN) is now supported by the JIT compiler.
185    
186    30. Update RunTest with additional test selector options.
187    
188    31. The way PCRE handles backtracking verbs has been changed in two ways.
189    
190        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
191        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
192        backtracking. In some cases this makes it more Perl-compatible, but Perl's
193        rather obscure rules do not always do the same thing.
194    
195        (2) Previously, backtracking verbs were confined within assertions. This is
196        no longer the case for positive assertions, except for (*ACCEPT). Again,
197        this sometimes improves Perl compatibility, and sometimes does not.
198    
199    32. A number of tests that were in test 2 because Perl did things differently
200        have been moved to test 1, because either Perl or PCRE has changed, and
201        these tests are now compatible.
202    
203    32. Backtracking control verbs are now handled in the same way in JIT and
204        interpreter.
205    
206    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
207        contained a forward subroutine reference caused a compile error.
208    
209    34. Auto-detect and optimize limited repetitions in JIT.
210    
211    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
212        blocking (*UTF) etc.
213    
214    36. In the interpreter, maximizing pattern repetitions for characters and
215        character types now use tail recursion, which reduces stack usage.
216    
217    37. The value of the max lookbehind was not correctly preserved if a compiled
218        and saved regex was reloaded on a host of different endianness.
219    
220    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
221        of the compiled pattern block, expand the flags field from 16 to 32 bits
222        because it was almost full.
223    
224    39. Try madvise first before posix_madvise.
225    
226    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
227        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
228        original change locked out all matches of empty strings. This has been
229        changed so that one match of an empty string per line is recognized.
230        Subsequent searches on the same line (for colouring or for --only-matching,
231        for example) do not recognize empty strings.
232    
233    41. Applied a user patch to fix a number of spelling mistakes in comments.
234    
235    42. Data lines longer than 65536 caused pcretest to crash.
236    
237    43. Clarified the data type for length and startoffset arguments for pcre_exec
238        and pcre_dfa_exec in the function-specific man pages, where they were
239        explicitly stated to be in bytes, never having been updated. I also added
240        some clarification to the pcreapi man page.
241    
242    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
243        a segmentation fault.
244    
245    
246    Version 8.32 30-November-2012
247    -----------------------------
248    
249    1.  Improved JIT compiler optimizations for first character search and single
250        character iterators.
251    
252    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
253        Patch by Daniel Richard G.
254    
255    3.  Single character iterator optimizations in the JIT compiler.
256    
257    4.  Improved JIT compiler optimizations for character ranges.
258    
259    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
260        Reported by Giuseppe D'Angelo.
261    
262    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
263        of a line, was being set incorrectly in cases where .* appeared inside
264        atomic brackets at the start of a pattern, or where there was a subsequent
265        *PRUNE or *SKIP.
266    
267    7.  Improved instruction cache flush for POWER/PowerPC.
268        Patch by Daniel Richard G.
269    
270    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
271        grep:
272    
273        (a) There is now no limit to the number of patterns to be matched.
274    
275        (b) An error is given if a pattern is too long.
276    
277        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
278            are now supported.
279    
280        (d) --exclude-from and --include-from (multiple use) have been added.
281    
282        (e) Exclusions and inclusions now apply to all files and directories, not
283            just to those obtained from scanning a directory recursively.
284    
285        (f) Multiple uses of -f and --file-list are now supported.
286    
287        (g) In a Windows environment, the default for -d has been changed from
288            "read" (the GNU grep default) to "skip", because otherwise the presence
289            of a directory in the file list provokes an error.
290    
291        (h) The documentation has been revised and clarified in places.
292    
293    9.  Improve the matching speed of capturing brackets.
294    
295    10. Changed the meaning of \X so that it now matches a Unicode extended
296        grapheme cluster.
297    
298    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
299        out POSIX threads when JIT support is configured.
300    
301    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
302    
303    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
304        structure could go wrong in environments where size_t is not the same size
305        as int.
306    
307    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
308        set.
309    
310    15. The EBCDIC support had decayed; later updates to the code had included
311        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
312        general tidy up of EBCDIC-related issues, and the documentation was also
313        not quite right. There is now a test that can be run on ASCII systems to
314        check some of the EBCDIC-related things (but is it not a full test).
315    
316    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
317        in a small tidy to the code.
318    
319    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
320    
321    18. If the --only-matching (-o) option in pcregrep is specified multiple
322        times, each one causes appropriate output. For example, -o1 -o2 outputs the
323        substrings matched by the 1st and 2nd capturing parentheses. A separating
324        string can be specified by --om-separator (default empty).
325    
326    19. Improving the first n character searches.
327    
328    20. Turn case lists for horizontal and vertical white space into macros so that
329        they are defined only once.
330    
331    21. This set of changes together give more compatible Unicode case-folding
332        behaviour for characters that have more than one other case when UCP
333        support is available.
334    
335        (a) The Unicode property table now has offsets into a new table of sets of
336            three or more characters that are case-equivalent. The MultiStage2.py
337            script that generates these tables (the pcre_ucd.c file) now scans
338            CaseFolding.txt instead of UnicodeData.txt for character case
339            information.
340    
341        (b) The code for adding characters or ranges of characters to a character
342            class has been abstracted into a generalized function that also handles
343            case-independence. In UTF-mode with UCP support, this uses the new data
344            to handle characters with more than one other case.
345    
346        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
347            whose other case is greater than 256 are now correctly matched
348            caselessly. Previously, the high codepoint matched the low one, but not
349            vice versa.
350    
351        (d) The processing of \h, \H, \v, and \ in character classes now makes use
352            of the new class addition function, using character lists defined as
353            macros alongside the case definitions of 20 above.
354    
355        (e) Caseless back references now work with characters that have more than
356            one other case.
357    
358        (f) General caseless matching of characters with more than one other case
359            is supported.
360    
361    22. Unicode character properties were updated from Unicode 6.2.0
362    
363    23. Improved CMake support under Windows. Patch by Daniel Richard G.
364    
365    24. Add support for 32-bit character strings, and UTF-32
366    
367    25. Major JIT compiler update (code refactoring and bugfixing).
368        Experimental Sparc 32 support is added.
369    
370    26. Applied a modified version of Daniel Richard G's patch to create
371        pcre.h.generic and config.h.generic by "make" instead of in the
372        PrepareRelease script.
373    
374    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
375        pcre_compile.c when checking for a zero character.
376    
377    28. Introducing a native interface for JIT. Through this interface, the compiled
378        machine code can be directly executed. The purpose of this interface is to
379        provide fast pattern matching, so several sanity checks are not performed.
380        However, feature tests are still performed. The new interface provides
381        1.4x speedup compared to the old one.
382    
383    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
384        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
385        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
386    
387    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
388        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
389        the "old" RFC 2279). Instead, it ended up passing a negative length to
390        pcre_exec().
391    
392    31. Add support for GCC's visibility feature to hide internal functions.
393    
394    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
395        "unknown -C option" after outputting 0 or 1.
396    
397    33. There is now support for generating a code coverage report for the test
398        suite in environments where gcc is the compiler and lcov is installed. This
399        is mainly for the benefit of the developers.
400    
401    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
402        unaddressable using valgrind annotations, allowing valgrind to detect
403        invalid memory accesses. This is mainly for the benefit of the developers.
404    
405    25. (*UTF) can now be used to start a pattern in any of the three libraries.
406    
407    26. Give configure error if --enable-cpp but no C++ compiler found.
408    
409    
410    Version 8.31 06-July-2012
411    -------------------------
412    
413    1.  Fixing a wrong JIT test case and some compiler warnings.
414    
415    2.  Removed a bashism from the RunTest script.
416    
417    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
418        to unsigned type, result still unsigned" that was given by an MS compiler
419        on encountering the code "-sizeof(xxx)".
420    
421    4.  Partial matching support is added to the JIT compiler.
422    
423    5.  Fixed several bugs concerned with partial matching of items that consist
424        of more than one character:
425    
426        (a) /^(..)\1/ did not partially match "aba" because checking references was
427            done on an "all or nothing" basis. This also applied to repeated
428            references.
429    
430        (b) \R did not give a hard partial match if \r was found at the end of the
431            subject.
432    
433        (c) \X did not give a hard partial match after matching one or more
434            characters at the end of the subject.
435    
436        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
437            a partial match for the string "\r".
438    
439        (e) When newline was set to CRLF, the metacharacter "." did not recognize
440            a partial match for a CR character at the end of the subject string.
441    
442    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
443        running pcretest, the text "(JIT)" added to the output whenever JIT is
444        actually used to run the match.
445    
446    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
447        or /S+[+] with a digit between 1 and 7.
448    
449    8.  OP_NOT now supports any UTF character not just single-byte ones.
450    
451    9.  (*MARK) control verb is now supported by the JIT compiler.
452    
453    10. The command "./RunTest list" lists the available tests without actually
454        running any of them. (Because I keep forgetting what they all are.)
455    
456    11. Add PCRE_INFO_MAXLOOKBEHIND.
457    
458    12. Applied a (slightly modified) user-supplied patch that improves performance
459        when the heap is used for recursion (compiled with --disable-stack-for-
460        recursion). Instead of malloc and free for each heap frame each time a
461        logical recursion happens, frames are retained on a chain and re-used where
462        possible. This sometimes gives as much as 30% improvement.
463    
464    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
465        call.
466    
467    14. As documented, (*COMMIT) is now confined to within a positive assertion.
468    
469    15. It is now possible to link pcretest with libedit as an alternative to
470        libreadline.
471    
472    16. (*COMMIT) control verb is now supported by the JIT compiler.
473    
474    17. The Unicode data tables have been updated to Unicode 6.1.0.
475    
476    18. Added --file-list option to pcregrep.
477    
478    19. Added binary file support to pcregrep, including the -a, --binary-files,
479        -I, and --text options.
480    
481    20. The madvise function is renamed for posix_madvise for QNX compatibility
482        reasons. Fixed by Giuseppe D'Angelo.
483    
484    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
485    
486    22. Changed the option for creating symbolic links for 16-bit man pages from
487        -s to -sf so that re-installing does not cause issues.
488    
489    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
490    
491    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
492        matches in certain environments (the workspace was not being correctly
493        retained). Also added to pcre_dfa_exec() a simple plausibility check on
494        some of the workspace data at the beginning of a restart.
495    
496    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
497        was not doing so when it should - probably a typo introduced by SVN 528
498        (change 8.10/14).
499    
500    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
501        \w+ when the character tables indicated that \x{c4} was a word character.
502        There were several related cases, all because the tests for doing a table
503        lookup were testing for characters less than 127 instead of 255.
504    
505    27. If a pattern contains capturing parentheses that are not used in a match,
506        their slots in the ovector are set to -1. For those that are higher than
507        any matched groups, this happens at the end of processing. In the case when
508        there were back references that the ovector was too small to contain
509        (causing temporary malloc'd memory to be used during matching), and the
510        highest capturing number was not used, memory off the end of the ovector
511        was incorrectly being set to -1. (It was using the size of the temporary
512        memory instead of the true size.)
513    
514    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
515        ovector size, it uses memory at the end of the block that it has got.
516    
517    29. Check for an overlong MARK name and give an error at compile time. The
518        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
519    
520    30. JIT compiler update.
521    
522    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
523        Rill for the patch.
524    
525    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
526    
527    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
528    
529    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
530        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
531        was enabled.
532    
533    35. Improve JIT code generation for greedy plus quantifier.
534    
535    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
536        1 to "aa" instead of to an empty string. The bug affected repeated groups
537        that could potentially match an empty string.
538    
539    37. Optimizing single character iterators in JIT.
540    
541    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
542        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
543        codepoints that are too big for the mode are faulted, and in a UTF mode,
544        disallowed codepoints are also faulted.
545    
546    39. If PCRE was compiled with UTF support, in three places in the DFA
547        matcher there was code that should only have been obeyed in UTF mode, but
548        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
549        processing when bytes with values greater than 127 were present. In 16-bit
550        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
551        both cases the values are those that cannot be the first data item in a UTF
552        character. The three items that might have provoked this were recursions,
553        possessively repeated groups, and atomic groups.
554    
555    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
556        and pcregrep, because some OS require shared objects to be explicitly
557        passed to ld, causing the link step to fail if they are not.
558    
559    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
560        mode, patterns that started with \h* or \R* might be incorrectly matched.
561    
562    
563    Version 8.30 04-February-2012
564    -----------------------------
565    
566    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
567        name is defined in ctype.h.
568    
569    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
570        only in quite long subpatterns.
571    
572    3.  Removed the function pcre_info(), which has been obsolete and deprecated
573        since it was replaced by pcre_fullinfo() in February 2000.
574    
575    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
576        match a (*MARK), and the match failed at the start of the subject, a
577        reference to memory before the start of the subject could occur. This bug
578        was introduced by fix 17 of release 8.21.
579    
580    5.  A reference to an unset group with zero minimum repetition was giving
581        totally wrong answers (in non-JavaScript-compatibility mode). For example,
582        /(another)?(\1?)test/ matched against "hello world test". This bug was
583        introduced in release 8.13.
584    
585    6.  Add support for 16-bit character strings (a large amount of work involving
586        many changes and refactorings).
587    
588    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
589        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
590        from a file.
591    
592    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
593        rounding is not applied in this particular case).
594    
595    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
596        if they appear, or are escaped, in patterns.
597    
598    10. Get rid of a number of -Wunused-but-set-variable warnings.
599    
600    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
601        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
602        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
603        also returns the mark "x". This bug applied to capturing parentheses,
604        non-capturing parentheses, and atomic parentheses. It also applied to some
605        assertions.
606    
607    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
608        information out of configure.ac instead of relying on pcre.h.generic, which
609        is not stored in the repository.
610    
611    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
612        -lreadline.
613    
614    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
615    
616    15. Applied Graycode's patch to put the top-level frame on the stack rather
617        than the heap when not using the stack for recursion. This gives a
618        performance improvement in many cases when recursion is not deep.
619    
620    16. Experimental code added to "pcretest -C" to output the stack frame size.
621    
622    
623    Version 8.21 12-Dec-2011
624    ------------------------
625    
626    1.  Updating the JIT compiler.
627    
628    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
629        are added as well.
630    
631    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
632        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
633        calling _pcre_jit_exec. Some extra comments are added.
634    
635    4.  (*MARK) settings inside atomic groups that do not contain any capturing
636        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
637        was introduced by change 18 for 8.20.
638    
639    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
640        ECMA-262 standard.
641    
642    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
643        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
644        This bug was probably introduced by change 9 of 8.13.
645    
646    7.  While fixing 6 above, I noticed that a number of other items were being
647        incorrectly rejected as "not fixed length". This arose partly because newer
648        opcodes had not been added to the fixed-length checking code. I have (a)
649        corrected the bug and added tests for these items, and (b) arranged for an
650        error to occur if an unknown opcode is encountered while checking for fixed
651        length instead of just assuming "not fixed length". The items that were
652        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
653        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
654        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
655    
656    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
657        being incorrectly compiled and would have given unpredicatble results.
658    
659    9.  A possessively repeated subpattern with minimum repeat count greater than
660        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
661        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
662        the first (A) could occur when it should not.
663    
664    10. Add a cast and remove a redundant test from the code.
665    
666    11. JIT should use pcre_malloc/pcre_free for allocation.
667    
668    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
669        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
670        is anything other than /usr, -L is still shown).
671    
672    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
673    
674    14. Perl does not support \N without a following name in a [] class; PCRE now
675        also gives an error.
676    
677    15. If a forward reference was repeated with an upper limit of around 2000,
678        it caused the error "internal error: overran compiling workspace". The
679        maximum number of forward references (including repeats) was limited by the
680        internal workspace, and dependent on the LINK_SIZE. The code has been
681        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
682        the default depends on LINK_SIZE. There is a new upper limit (for safety)
683        of around 200,000 forward references. While doing this, I also speeded up
684        the filling in of repeated forward references.
685    
686    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
687        incorrectly expecting the subject to contain another "a" after the start.
688    
689    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
690        in the match, the SKIP should be ignored. This was not happening; instead
691        the SKIP was being treated as NOMATCH. For patterns such as
692        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
693        tested.
694    
695    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
696        now much more compatible with Perl, in particular in cases where the result
697        is a non-match for a non-anchored pattern. For example, if
698        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
699        "m", where previously it did not return a name. A side effect of this
700        change is that for partial matches, the last encountered mark name is
701        returned, as for non matches. A number of tests that were previously not
702        Perl-compatible have been moved into the Perl-compatible test files. The
703        refactoring has had the pleasing side effect of removing one argument from
704        the match() function, thus reducing its stack requirements.
705    
706    19. If the /S+ option was used in pcretest to study a pattern using JIT,
707        subsequent uses of /S (without +) incorrectly behaved like /S+.
708    
709    21. Retrieve executable code size support for the JIT compiler and fixing
710        some warnings.
711    
712    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
713        not work when the shorter character appeared right at the end of the
714        subject string.
715    
716    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
717        systems.
718    
719    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
720        output it when the /M option is used in pcretest.
721    
722    25. The CheckMan script was not being included in the distribution. Also, added
723        an explicit "perl" to run Perl scripts from the PrepareRelease script
724        because this is reportedly needed in Windows.
725    
726    26. If study data was being save in a file and studying had not found a set of
727        "starts with" bytes for the pattern, the data written to the file (though
728        never used) was taken from uninitialized memory and so caused valgrind to
729        complain.
730    
731    27. Updated RunTest.bat as provided by Sheri Pierce.
732    
733    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
734    
735    29. Computation of memory usage for the table of capturing group names was
736        giving an unnecessarily large value.
737    
738    
739    Version 8.20 21-Oct-2011
740    ------------------------
741    
742    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
743        a POSIX class. After further experiments with Perl, which convinced me that
744        Perl has bugs and confusions, a closing square bracket is no longer allowed
745        in a POSIX name. This bug also affected patterns with classes that started
746        with full stops.
747    
748    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
749        captured substring, but while checking the failing first alternative,
750        substring 1 is temporarily captured. If the output vector supplied to
751        pcre_exec() was not big enough for this capture, the yield of the function
752        was still zero ("insufficient space for captured substrings"). This cannot
753        be totally fixed without adding another stack variable, which seems a lot
754        of expense for a edge case. However, I have improved the situation in cases
755        such as /(a)(b)x|abc/ matched against "abc", where the return code
756        indicates that fewer than the maximum number of slots in the ovector have
757        been set.
758    
759    3.  Related to (2) above: when there are more back references in a pattern than
760        slots in the output vector, pcre_exec() uses temporary memory during
761        matching, and copies in the captures as far as possible afterwards. It was
762        using the entire output vector, but this conflicts with the specification
763        that only 2/3 is used for passing back captured substrings. Now it uses
764        only the first 2/3, for compatibility. This is, of course, another edge
765        case.
766    
767    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
768        main code base, and can be used by building with --enable-jit. When this is
769        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
770        runtime --no-jit option is given.
771    
772    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
773        ovector, the return from the function was zero, implying that there were
774        other matches that did not fit. The correct "exactly full" value is now
775        returned.
776    
777    6.  If a subpattern that was called recursively or as a subroutine contained
778        (*PRUNE) or any other control that caused it to give a non-standard return,
779        invalid errors such as "Error -26 (nested recursion at the same subject
780        position)" or even infinite loops could occur.
781    
782    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
783        computing the minimum length on reaching *ACCEPT, and so ended up with the
784        wrong value of 1 rather than 0. Further investigation indicates that
785        computing a minimum subject length in the presence of *ACCEPT is difficult
786        (think back references, subroutine calls), and so I have changed the code
787        so that no minimum is registered for a pattern that contains *ACCEPT.
788    
789    8.  If (*THEN) was present in the first (true) branch of a conditional group,
790        it was not handled as intended. [But see 16 below.]
791    
792    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
793        Sheri Pierce.
794    
795    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
796        the first byte in a match must be "a".
797    
798    11. Change 17 for 8.13 increased the recursion depth for patterns like
799        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
800        pattern contains any instances of (*THEN). If it does not, the old
801        optimizations are restored. It would be nice to do this on a per-group
802        basis, but at the moment that is not feasible.
803    
804    12. In some environments, the output of pcretest -C is CRLF terminated. This
805        broke RunTest's code that checks for the link size. A single white space
806        character after the value is now allowed for.
807    
808    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
809        For "fr", it uses the Windows-specific input and output files.
810    
811    14. If (*THEN) appeared in a group that was called recursively or as a
812        subroutine, it did not work as intended. [But see next item.]
813    
814    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
815        pattern fragments (but not containing any | characters). If A and B are
816        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
817        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
818        D. In other words, Perl considers parentheses that do not contain any |
819        characters to be part of a surrounding alternative, whereas PCRE was
820        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
821        differently. PCRE now behaves in the same way as Perl, except in the case
822        of subroutine/recursion calls such as (?1) which have in any case always
823        been different (but PCRE had them first :-).
824    
825    16. Related to 15 above: Perl does not treat the | in a conditional group as
826        creating alternatives. Such a group is treated in the same way as an
827        ordinary group without any | characters when processing (*THEN). PCRE has
828        been changed to match Perl's behaviour.
829    
830    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
831        RunGrepTest script failed.
832    
833    18. Change 22 for version 13 caused atomic groups to use more stack. This is
834        inevitable for groups that contain captures, but it can lead to a lot of
835        stack use in large patterns. The old behaviour has been restored for atomic
836        groups that do not contain any capturing parentheses.
837    
838    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
839        suppress the check for a minimum subject length at run time. (If it was
840        given to pcre_exec() or pcre_dfa_exec() it did work.)
841    
842    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
843        fail to work when decoding hex characters in data strings in EBCDIC
844        environments.
845    
846    21. It appears that in at least one Mac OS environment, the isxdigit() function
847        is implemented as a macro that evaluates to its argument more than once,
848        contravening the C 90 Standard (I haven't checked a later standard). There
849        was an instance in pcretest which caused it to go wrong when processing
850        \x{...} escapes in subject strings. The has been rewritten to avoid using
851        things like p++ in the argument of isxdigit().
852    
853    
854    Version 8.13 16-Aug-2011
855    ------------------------
856    
857    1.  The Unicode data tables have been updated to Unicode 6.0.0.
858    
859    2.  Two minor typos in pcre_internal.h have been fixed.
860    
861    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
862        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
863        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
864    
865    4.  There were a number of related bugs in the code for matching backrefences
866        caselessly in UTF-8 mode when codes for the characters concerned were
867        different numbers of bytes. For example, U+023A and U+2C65 are an upper
868        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
869        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
870        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
871        2-byte code at the end of the subject (it thought there wasn't enough data
872        left).
873    
874    5.  Comprehensive information about what went wrong is now returned by
875        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
876        as the output vector has at least 2 elements. The offset of the start of
877        the failing character and a reason code are placed in the vector.
878    
879    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
880        now returned is for the first byte of the failing character, instead of the
881        last byte inspected. This is an incompatible change, but I hope it is small
882        enough not to be a problem. It makes the returned offset consistent with
883        pcre_exec() and pcre_dfa_exec().
884    
885    7.  pcretest now gives a text phrase as well as the error number when
886        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
887        failure, the offset and reason code are output.
888    
889    8.  When \R was used with a maximizing quantifier it failed to skip backwards
890        over a \r\n pair if the subsequent match failed. Instead, it just skipped
891        back over a single character (\n). This seems wrong (because it treated the
892        two characters as a single entity when going forwards), conflicts with the
893        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
894        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
895        has been changed.
896    
897    9.  Some internal refactoring has changed the processing so that the handling
898        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
899        time (the PCRE_DOTALL option was changed this way some time ago: version
900        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
901        which was always a bit of a fudge. It also means that there is one less
902        argument for the match() function, which reduces its stack requirements
903        slightly. This change also fixes an incompatibility with Perl: the pattern
904        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
905    
906    10. More internal refactoring has drastically reduced the number of recursive
907        calls to match() for possessively repeated groups such as (abc)++ when
908        using pcre_exec().
909    
910    11. While implementing 10, a number of bugs in the handling of groups were
911        discovered and fixed:
912    
913        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
914        (a|)*(?1) gave a compile-time internal error.
915        ((a|)+)+  did not notice that the outer group could match an empty string.
916        (^a|^)+   was not marked as anchored.
917        (.*a|.*)+ was not marked as matching at start or after a newline.
918    
919    12. Yet more internal refactoring has removed another argument from the match()
920        function. Special calls to this function are now indicated by setting a
921        value in a variable in the "match data" data block.
922    
923    13. Be more explicit in pcre_study() instead of relying on "default" for
924        opcodes that mean there is no starting character; this means that when new
925        ones are added and accidentally left out of pcre_study(), testing should
926        pick them up.
927    
928    14. The -s option of pcretest has been documented for ages as being an old
929        synonym of -m (show memory usage). I have changed it to mean "force study
930        for every regex", that is, assume /S for every regex. This is similar to -i
931        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
932        using it. It makes it easier to run collections of tests with and without
933        study enabled, and thereby test pcre_study() more easily. All the standard
934        tests are now run with and without -s (but some patterns can be marked as
935        "never study" - see 20 below).
936    
937    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
938        restoration of the capturing data to the outer values was not happening
939        correctly.
940    
941    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
942        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
943        pattern had matched an empty string, and so incorrectly returned a no
944        match.
945    
946    17. There was optimizing code for the last branch of non-capturing parentheses,
947        and also for the obeyed branch of a conditional subexpression, which used
948        tail recursion to cut down on stack usage. Unfortunately, now that there is
949        the possibility of (*THEN) occurring in these branches, tail recursion is
950        no longer possible because the return has to be checked for (*THEN). These
951        two optimizations have therefore been removed. [But see 8.20/11 above.]
952    
953    18. If a pattern containing \R was studied, it was assumed that \R always
954        matched two bytes, thus causing the minimum subject length to be
955        incorrectly computed because \R can also match just one byte.
956    
957    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
958        was incorrectly computed.
959    
960    20. If /S is present twice on a test pattern in pcretest input, it now
961        *disables* studying, thereby overriding the use of -s on the command line
962        (see 14 above). This is necessary for one or two tests to keep the output
963        identical in both cases.
964    
965    21. When (*ACCEPT) was used in an assertion that matched an empty string and
966        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
967    
968    22. When an atomic group that contained a capturing parenthesis was
969        successfully matched, but the branch in which it appeared failed, the
970        capturing was not being forgotten if a higher numbered group was later
971        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
972        group 1 to "a", when in fact it should be unset. This applied to multi-
973        branched capturing and non-capturing groups, repeated or not, and also to
974        positive assertions (capturing in negative assertions does not happen
975        in PCRE) and also to nested atomic groups.
976    
977    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
978        subject after a captured substring, to make it easier to tell which of a
979        number of identical substrings has been captured.
980    
981    24. The way atomic groups are processed by pcre_exec() has been changed so that
982        if they are repeated, backtracking one repetition now resets captured
983        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
984        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
985        "aaa". Previously, it would have been "a". As part of this code
986        refactoring, the way recursive calls are handled has also been changed.
987    
988    25. If an assertion condition captured any substrings, they were not passed
989        back unless some other capturing happened later. For example, if
990        (?(?=(a))a) was matched against "a", no capturing was returned.
991    
992    26. When studying a pattern that contained subroutine calls or assertions,
993        the code for finding the minimum length of a possible match was handling
994        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
995        group 1 called group 2 while simultaneously a separate group 2 called group
996        1). A stack overflow occurred in this case. I have fixed this by limiting
997        the recursion depth to 10.
998    
999    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1000        Fortmann. This supports explicit test numbers on the command line, and has
1001        argument validation and error reporting.
1002    
1003    28. An instance of \X with an unlimited repeat could fail if at any point the
1004        first character it looked at was a mark character.
1005    
1006    29. Some minor code refactoring concerning Unicode properties and scripts
1007        should reduce the stack requirement of match() slightly.
1008    
1009    30. Added the '=' option to pcretest to check the setting of unused capturing
1010        slots at the end of the pattern, which are documented as being -1, but are
1011        not included in the return count.
1012    
1013    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1014        compiled something random. Now it gives a compile-time error (as does
1015        Perl).
1016    
1017    32. A *MARK encountered during the processing of a positive assertion is now
1018        recorded and passed back (compatible with Perl).
1019    
1020    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1021        had alternative anchored branches, the search for a second match in a line
1022        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1023        matched the line "0102" twice. The same bug affected patterns that started
1024        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1025        twice.
1026    
1027    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1028        does, and because of capturing effects, quantifying parenthesized
1029        assertions may at times be useful. Quantifiers are now allowed for
1030        parenthesized assertions.
1031    
1032    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1033    
1034    36. \g was being checked for fancy things in a character class, when it should
1035        just be a literal "g".
1036    
1037    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1038        appearance of a nested POSIX class supersedes an apparent external class.
1039        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1040        unescaped square brackets may also appear as part of class names. For
1041        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1042        more like Perl. (But see 8.20/1 above.)
1043    
1044    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1045        was because it thought it was \N{name}, which is not supported).
1046    
1047    39. Add minix to OS list not supporting the -S option in pcretest.
1048    
1049    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1050        cannot analyze patterns in sufficient detail to catch mutual recursions
1051        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1052        subgroup is called recursively as a subpattern for a second time at the
1053        same position in the subject string. In previous releases this might have
1054        been caught by the recursion limit, or it might have run out of stack.
1055    
1056    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1057        happen only once. PCRE was, however incorrectly giving a compile time error
1058        "recursive call could loop indefinitely" because it cannot analyze the
1059        pattern in sufficient detail. The compile time test no longer happens when
1060        PCRE is compiling a conditional subpattern, but actual runaway loops are
1061        now caught at runtime (see 40 above).
1062    
1063    42. It seems that Perl allows any characters other than a closing parenthesis
1064        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1065        has been changed to be the same.
1066    
1067    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1068        as not to get warnings when autogen.sh is called. Also changed
1069        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1070    
1071    44. To help people who use pcregrep to scan files containing exceedingly long
1072        lines, the following changes have been made:
1073    
1074        (a) The default value of the buffer size parameter has been increased from
1075            8K to 20K. (The actual buffer used is three times this size.)
1076    
1077        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1078            PCRE is built.
1079    
1080        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1081            to be set at run time.
1082    
1083        (d) Numerical values in pcregrep options can be followed by K or M, for
1084            example --buffer-size=50K.
1085    
1086        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1087            given and the return code is set to 2.
1088    
1089    45. Add a pointer to the latest mark to the callout data block.
1090    
1091    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1092        partial match of an empty string instead of no match. This was specific to
1093        the use of ".".
1094    
1095    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1096        complete match instead of a partial match. This bug was dependent on both
1097        the PCRE_UTF8 and PCRE_DOTALL options being set.
1098    
1099    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1100        starting byte set, because \b was not being ignored.
1101    
1102    
1103    Version 8.12 15-Jan-2011
1104    ------------------------
1105    
1106    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1107        checks for such things as part of the documentation building process.
1108    
1109    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1110        --match-limit and --recursion-limit options (added for 8.11). In
1111        particular, this made one of the standard tests fail. (The integer value
1112        went into the wrong half of a long int.)
1113    
1114    3.  If the --colour option was given to pcregrep with -v (invert match), it
1115        did strange things, either producing crazy output, or crashing. It should,
1116        of course, ignore a request for colour when reporting lines that do not
1117        match.
1118    
1119    4.  Another pcregrep bug caused similar problems if --colour was specified with
1120        -M (multiline) and the pattern match finished with a line ending.
1121    
1122    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1123        matched in multiline mode, the following line was shown as part of the
1124        match. This seems wrong, so I have changed it.
1125    
1126    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1127        the check for further matches in the same line (so they could be coloured)
1128        to overrun the end of the current line. If another match was found, it was
1129        incorrectly shown (and then shown again when found in the next line).
1130    
1131    7.  If pcregrep was compiled under Windows, there was a reference to the
1132        function pcregrep_exit() before it was defined. I am assuming this was
1133        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1134        reported by a user. I've moved the definition above the reference.
1135    
1136    
1137    Version 8.11 10-Dec-2010
1138    ------------------------
1139    
1140    1.  (*THEN) was not working properly if there were untried alternatives prior
1141        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1142        backtracked to try for "b" instead of moving to the next alternative branch
1143        at the same level (in this case, to look for "c"). The Perl documentation
1144        is clear that when (*THEN) is backtracked onto, it goes to the "next
1145        alternative in the innermost enclosing group".
1146    
1147    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1148        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1149        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1150        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1151        (*THEN).
1152    
1153    3.  If \s appeared in a character class, it removed the VT character from
1154        the class, even if it had been included by some previous item, for example
1155        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1156        of \s, but is part of the POSIX "space" class.)
1157    
1158    4.  A partial match never returns an empty string (because you can always
1159        match an empty string at the end of the subject); however the checking for
1160        an empty string was starting at the "start of match" point. This has been
1161        changed to the "earliest inspected character" point, because the returned
1162        data for a partial match starts at this character. This means that, for
1163        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1164        (previously it gave "no match").
1165    
1166    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1167        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1168        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1169        has an implication that the given string is incomplete (because a partial
1170        match is preferred over a full match). For this reason, these items now
1171        give a partial match in this situation. [Aside: previously, the one case
1172        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1173        match rather than a full match, which was wrong by the old rules, but is
1174        now correct.]
1175    
1176    6.  There was a bug in the handling of #-introduced comments, recognized when
1177        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1178        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1179        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1180        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1181        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1182        places in pcre_compile().
1183    
1184    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1185        comments when looking ahead for named forward references to subpatterns,
1186        the only newline sequence it recognized was NL. It now handles newlines
1187        according to the set newline convention.
1188    
1189    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1190        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1191        cater for a lack of strerror(). These oversights have been fixed.
1192    
1193    9.  Added --match-limit and --recursion-limit to pcregrep.
1194    
1195    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1196    
1197    11. When the -o option was used, pcregrep was setting a return code of 1, even
1198        when matches were found, and --line-buffered was not being honoured.
1199    
1200    12. Added an optional parentheses number to the -o and --only-matching options
1201        of pcregrep.
1202    
1203    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1204        can match an empty string. The code to do it in pcretest and pcredemo
1205        needed fixing:
1206    
1207        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1208            only one byte after an empty string match just before CRLF (this case
1209            just got forgotten; "any" and "anycrlf" were OK).
1210    
1211        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1212            mode when an empty string match preceded an ASCII character followed by
1213            a non-ASCII character. (The code for advancing by one character rather
1214            than one byte was nonsense.)
1215    
1216        (c) The pcredemo.c sample program did not have any code at all to handle
1217            the cases when CRLF is a valid newline sequence.
1218    
1219    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1220        as a starting offset was within the subject string. There is now a new
1221        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1222        negative or greater than the length of the string. In order to test this,
1223        pcretest is extended to allow the setting of negative starting offsets.
1224    
1225    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1226        starting offset points to the beginning of a UTF-8 character was
1227        unnecessarily clumsy. I tidied it up.
1228    
1229    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1230        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1231    
1232    17. Nobody had reported that the --include_dir option, which was added in
1233        release 7.7 should have been called --include-dir (hyphen, not underscore)
1234        for compatibility with GNU grep. I have changed it to --include-dir, but
1235        left --include_dir as an undocumented synonym, and the same for
1236        --exclude-dir, though that is not available in GNU grep, at least as of
1237        release 2.5.4.
1238    
1239    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1240        characters from a string of bytes) have been redefined so as not to use
1241        loops, in order to improve performance in some environments. At the same
1242        time, I abstracted some of the common code into auxiliary macros to save
1243        repetition (this should not affect the compiled code).
1244    
1245    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1246        compile-time error is now given if \c is not followed by an ASCII
1247        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1248        different, and any byte value is allowed.)
1249    
1250    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1251        START_OPTIMIZE option, which is now allowed at compile time - but just
1252        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1253        to pcregrep and other applications that have no direct access to PCRE
1254        options. The new /Y option in pcretest sets this option when calling
1255        pcre_compile().
1256    
1257    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1258        back references. Groups containing recursive back references were forced to
1259        be atomic by that change, but in the case of named groups, the amount of
1260        memory required was incorrectly computed, leading to "Failed: internal
1261        error: code overflow". This has been fixed.
1262    
1263    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1264        pcretest.c, to avoid build problems in some Borland environments.
1265    
1266    
1267    Version 8.10 25-Jun-2010
1268    ------------------------
1269    
1270    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1271        THEN.
1272    
1273    2.  (*ACCEPT) was not working when inside an atomic group.
1274    
1275    3.  Inside a character class, \B is treated as a literal by default, but
1276        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1277        causes the error). The code is unchanged, but I tidied the documentation.
1278    
1279    4.  Inside a character class, PCRE always treated \R and \X as literals,
1280        whereas Perl faults them if its -w option is set. I have changed PCRE so
1281        that it faults them when PCRE_EXTRA is set.
1282    
1283    5.  Added support for \N, which always matches any character other than
1284        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1285    
1286    6.  When compiling pcregrep with newer versions of gcc which may have
1287        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1288        declared with attribute warn_unused_result" were given. Just casting the
1289        result to (void) does not stop the warnings; a more elaborate fudge is
1290        needed. I've used a macro to implement this.
1291    
1292    7.  Minor change to pcretest.c to avoid a compiler warning.
1293    
1294    8.  Added four artifical Unicode properties to help with an option to make
1295        \s etc use properties (see next item). The new properties are: Xan
1296        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1297    
1298    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1299        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1300        this option. Modified pcretest to add /W to test this facility. Added
1301        REG_UCP to make it available via the POSIX interface.
1302    
1303    10. Added --line-buffered to pcregrep.
1304    
1305    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1306        studied, and the match started with a letter with a code point greater than
1307        127 whose first byte was different to the first byte of the other case of
1308        the letter, the other case of this starting letter was not recognized
1309        (#976).
1310    
1311    12. If a pattern that was studied started with a repeated Unicode property
1312        test, for example, \p{Nd}+, there was the theoretical possibility of
1313        setting up an incorrect bitmap of starting bytes, but fortunately it could
1314        not have actually happened in practice until change 8 above was made (it
1315        added property types that matched character-matching opcodes).
1316    
1317    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1318        possible starting bytes for non-anchored patterns.
1319    
1320    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1321        \R, and also a number of cases that involve Unicode properties, both
1322        explicit and implicit when PCRE_UCP is set.
1323    
1324    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1325        input, it could crash or give wrong results if characters with values
1326        greater than 0xc0 were present in the subject string. (Detail: it assumed
1327        UTF-8 input when processing these items.)
1328    
1329    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1330        size_t is 64-bit (#991).
1331    
1332    17. Added a check for running out of memory when PCRE is compiled with
1333        --disable-stack-for-recursion (#990).
1334    
1335    18. If the last data line in a file for pcretest does not have a newline on
1336        the end, a newline was missing in the output.
1337    
1338    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1339        less than 128) in its various bitmaps. However, there is a facility for
1340        generating tables according to the current locale when PCRE is compiled. It
1341        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1342        characters, are recognized by isspace() and therefore were getting set in
1343        these tables, and indeed these tables seem to approximate to ISO 8859. This
1344        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1345        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1346        which of course cannot start UTF-8 characters. I have changed the code so
1347        that only real ASCII characters (less than 128) and the correct starting
1348        bytes for UTF-8 encodings are set for characters greater than 127 when in
1349        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1350        altogether.)
1351    
1352    20. Added the /T option to pcretest so as to be able to run tests with non-
1353        standard character tables, thus making it possible to include the tests
1354        used for 19 above in the standard set of tests.
1355    
1356    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1357        reference to a subpattern the other side of a comment that contains an
1358        opening parenthesis caused either an internal compiling error, or a
1359        reference to the wrong subpattern.
1360    
1361    
1362    Version 8.02 19-Mar-2010
1363    ------------------------
1364    
1365    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1366    
1367    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1368        configured.
1369    
1370    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1371        original author of that file, following a query about its status.
1372    
1373    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1374        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1375    
1376    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1377        quantifier applied to a forward-referencing subroutine call, could compile
1378        incorrect code or give the error "internal error: previously-checked
1379        referenced subpattern not found".
1380    
1381    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1382        variables to point to external functions. For these systems, therefore,
1383        pcre_malloc etc. are now initialized to local functions that call the
1384        relevant global functions.
1385    
1386    7.  There were two entries missing in the vectors called coptable and poptable
1387        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1388        I've fixed the data, and added a kludgy way of testing at compile time that
1389        the lengths are correct (equal to the number of opcodes).
1390    
1391    8.  Following on from 7, I added a similar kludge to check the length of the
1392        eint vector in pcreposix.c.
1393    
1394    9.  Error texts for pcre_compile() are held as one long string to avoid too
1395        much relocation at load time. To find a text, the string is searched,
1396        counting zeros. There was no check for running off the end of the string,
1397        which could happen if a new error number was added without updating the
1398        string.
1399    
1400    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1401    
1402    11. \K was not working if it appeared in an atomic group or in a group that
1403        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1404        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1405        the assertion is positive, but not if it is negative.
1406    
1407    12. Change 11 fortuitously reduced the size of the stack frame used in the
1408        "match()" function of pcre_exec.c by one pointer. Forthcoming
1409        implementation of support for (*MARK) will need an extra pointer on the
1410        stack; I have reserved it now, so that the stack frame size does not
1411        decrease.
1412    
1413    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1414        item in branch that calls a recursion is a subroutine call - as in the
1415        second branch in the above example - was incorrectly given the compile-
1416        time error "recursive call could loop indefinitely" because pcre_compile()
1417        was not correctly checking the subroutine for matching a non-empty string.
1418    
1419    14. The checks for overrunning compiling workspace could trigger after an
1420        overrun had occurred. This is a "should never occur" error, but it can be
1421        triggered by pathological patterns such as hundreds of nested parentheses.
1422        The checks now trigger 100 bytes before the end of the workspace.
1423    
1424    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1425    
1426    
1427    Version 8.01 19-Jan-2010
1428    ------------------------
1429    
1430    1.  If a pattern contained a conditional subpattern with only one branch (in
1431        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1432        computed the wrong minimum data length (which is of course zero for such
1433        subpatterns). This could cause incorrect "no match" results.
1434    
1435    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1436        the pattern is reset in the first branch, pcre_compile() failed with
1437        "internal error: code overflow at offset...". This happened only when
1438        the reset was to the original external option setting. (An optimization
1439        abstracts leading options settings into an external setting, which was the
1440        cause of this.)
1441    
1442    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1443        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1444        assertion pattern did not match (meaning that the assertion was true), it
1445        was incorrectly treated as false if the SKIP had been reached during the
1446        matching. This also applied to assertions used as conditions.
1447    
1448    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1449        assertion subpattern, including such a pattern used as a condition,
1450        unpredictable results occurred, instead of the error return
1451        PCRE_ERROR_DFA_UITEM.
1452    
1453    5.  The C++ GlobalReplace function was not working like Perl for the special
1454        situation when an empty string is matched. It now does the fancy magic
1455        stuff that is necessary.
1456    
1457    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1458        removed. (These were left over from very, very early versions of PCRE.)
1459    
1460    7.  Some cosmetic changes to the code to make life easier when compiling it
1461        as part of something else:
1462    
1463        (a) Change DEBUG to PCRE_DEBUG.
1464    
1465        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1466            called "current" as "current_branch", to prevent a collision with the
1467            Linux macro when compiled as a kernel module.
1468    
1469        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1470            prevent a collision with the Linux macro when compiled as a kernel
1471            module.
1472    
1473    8.  In pcre_compile() there are some checks for integer overflows that used to
1474        cast potentially large values to (double). This has been changed to that
1475        when building, a check for int64_t is made, and if it is found, it is used
1476        instead, thus avoiding the use of floating point arithmetic. (There is no
1477        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1478        double.
1479    
1480    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1481        2005 (difference between two addresses compared to an unsigned value).
1482    
1483    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1484        custom one, because of the following reported problem in Windows:
1485    
1486          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1487              under Win32.
1488          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1489              therefore missing the function definition.
1490          - The compiler thus generates a "C" signature for the test function.
1491          - The linker fails to find the "C" function.
1492          - PCRE fails to configure if asked to do so against libbz2.
1493    
1494    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1495        messages were output:
1496    
1497          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1498          rerunning libtoolize, to keep the correct libtool macros in-tree.
1499          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1500    
1501        I have done both of these things.
1502    
1503    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1504        most of the time, it *can* run out if it is given a pattern that contains a
1505        runaway infinite recursion. I updated the discussion in the pcrestack man
1506        page.
1507    
1508    13. Now that we have gone to the x.xx style of version numbers, the minor
1509        version may start with zero. Using 08 or 09 is a bad idea because users
1510        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1511        interpreted as invalid octal numbers. I've updated the previous comment in
1512        configure.ac, and also added a check that gives an error if 08 or 09 are
1513        used.
1514    
1515    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1516        causing partial matching to fail when the end of the subject matched \W
1517        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1518    
1519    15. There were some discrepancies between the declarations in pcre_internal.h
1520        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1521        their definitions. The declarations used "const uschar *" and the
1522        definitions used USPTR. Even though USPTR is normally defined as "const
1523        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1524        reported that: "This difference in casting confuses some C++ compilers, for
1525        example, SunCC recognizes above declarations as different functions and
1526        generates broken code for hbpcre." I have changed the declarations to use
1527        USPTR.
1528    
1529    16. GNU libtool is named differently on some systems. The autogen.sh script now
1530        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1531        (FreeBSD).
1532    
1533    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1534        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1535        comment: "Figure out how to create a longlong from a string: strtoll and
1536        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1537        instance, but it only takes 2 args instead of 3!"
1538    
1539    18. A subtle bug concerned with back references has been fixed by a change of
1540        specification, with a corresponding code fix. A pattern such as
1541        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1542        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1543        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1544        same bug. Such groups have to be quantified to be useful, or contained
1545        inside another quantified group. (If there's no repetition, the reference
1546        can never match.) The problem arises because, having left the group and
1547        moved on to the rest of the pattern, a later failure that backtracks into
1548        the group uses the captured value from the final iteration of the group
1549        rather than the correct earlier one. I have fixed this in PCRE by forcing
1550        any group that contains a reference to itself to be an atomic group; that
1551        is, there cannot be any backtracking into it once it has completed. This is
1552        similar to recursive and subroutine calls.
1553    
1554    
1555    Version 8.00 19-Oct-09
1556    ----------------------
1557    
1558    1.  The table for translating pcre_compile() error codes into POSIX error codes
1559        was out-of-date, and there was no check on the pcre_compile() error code
1560        being within the table. This could lead to an OK return being given in
1561        error.
1562    
1563    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1564        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1565        in a Windows environment.
1566    
1567    3.  The pcregrep --count option prints the count for each file even when it is
1568        zero, as does GNU grep. However, pcregrep was also printing all files when
1569        --files-with-matches was added. Now, when both options are given, it prints
1570        counts only for those files that have at least one match. (GNU grep just
1571        prints the file name in this circumstance, but including the count seems
1572        more useful - otherwise, why use --count?) Also ensured that the
1573        combination -clh just lists non-zero counts, with no names.
1574    
1575    4.  The long form of the pcregrep -F option was incorrectly implemented as
1576        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1577        but it seems right to fix it, and I didn't think it was worth preserving
1578        the old behaviour.
1579    
1580    5.  The command line items --regex=pattern and --regexp=pattern were not
1581        recognized by pcregrep, which required --regex pattern or --regexp pattern
1582        (with a space rather than an '='). The man page documented the '=' forms,
1583        which are compatible with GNU grep; these now work.
1584    
1585    6.  No libpcreposix.pc file was created for pkg-config; there was just
1586        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1587    
1588    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1589        when UCP support is not needed, by modifying the Python script that
1590        generates it from Unicode data files. This should not matter if the module
1591        is correctly used as a library, but I received one complaint about 50K of
1592        unwanted data. My guess is that the person linked everything into his
1593        program rather than using a library. Anyway, it does no harm.
1594    
1595    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1596        was a minimum greater than 1 for a wide character in a possessive
1597        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1598        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1599        character. Chaos in the form of incorrect output or a compiling loop could
1600        result.
1601    
1602    9.  The restrictions on what a pattern can contain when partial matching is
1603        requested for pcre_exec() have been removed. All patterns can now be
1604        partially matched by this function. In addition, if there are at least two
1605        slots in the offset vector, the offset of the earliest inspected character
1606        for the match and the offset of the end of the subject are set in them when
1607        PCRE_ERROR_PARTIAL is returned.
1608    
1609    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1610        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1611        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1612        and may be more useful for multi-segment matching.
1613    
1614    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1615        used to be given if ever the end of the subject was reached; now it is
1616        given only if matching could not proceed because another character was
1617        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1618        string "Z", which now yields "no match" instead of "partial match". In the
1619        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1620        final character ended with (*FAIL).
1621    
1622    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1623        if the pattern had a "must contain" character that was already found in the
1624        earlier partial match, unless partial matching was again requested. For
1625        example, with the pattern /dog.(body)?/, the "must contain" character is
1626        "g". If the first part-match was for the string "dog", restarting with
1627        "sbody" failed. This bug has been fixed.
1628    
1629    13. The string returned by pcre_dfa_exec() after a partial match has been
1630        changed so that it starts at the first inspected character rather than the
1631        first character of the match. This makes a difference only if the pattern
1632        starts with a lookbehind assertion or \b or \B (\K is not supported by
1633        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1634        matching functions compatible, and I think it's the right thing to do.
1635    
1636    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1637        so that the demonstration program is easily available in environments where
1638        PCRE has not been installed from source.
1639    
1640    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1641        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1642        library.
1643    
1644    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1645        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1646        is not the first non-POSIX option to be added. Clearly some people find
1647        these options useful.
1648    
1649    17. If a caller to the POSIX matching function regexec() passes a non-zero
1650        value for nmatch with a NULL value for pmatch, the value of
1651        nmatch is forced to zero.
1652    
1653    18. RunGrepTest did not have a test for the availability of the -u option of
1654        the diff command, as RunTest does. It now checks in the same way as
1655        RunTest, and also checks for the -b option.
1656    
1657    19. If an odd number of negated classes containing just a single character
1658        interposed, within parentheses, between a forward reference to a named
1659        subpattern and the definition of the subpattern, compilation crashed with
1660        an internal error, complaining that it could not find the referenced
1661        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1662        [The bug was that it was starting one character too far in when skipping
1663        over the character class, thus treating the ] as data rather than
1664        terminating the class. This meant it could skip too much.]
1665    
1666    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1667        /g option in pcretest when the pattern contains \K, which makes it possible
1668        to have an empty string match not at the start, even when the pattern is
1669        anchored. Updated pcretest and pcredemo to use this option.
1670    
1671    21. If the maximum number of capturing subpatterns in a recursion was greater
1672        than the maximum at the outer level, the higher number was returned, but
1673        with unset values at the outer level. The correct (outer level) value is
1674        now given.
1675    
1676    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1677        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1678        make it do so. The string so far is captured, making this feature
1679        compatible with Perl.
1680    
1681    23. The tests have been re-organized, adding tests 11 and 12, to make it
1682        possible to check the Perl 5.10 features against Perl 5.10.
1683    
1684    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1685        pattern matches a fixed length string. PCRE did not allow this; now it
1686        does. Neither allows recursion.
1687    
1688    25. I finally figured out how to implement a request to provide the minimum
1689        length of subject string that was needed in order to match a given pattern.
1690        (It was back references and recursion that I had previously got hung up
1691        on.) This code has now been added to pcre_study(); it finds a lower bound
1692        to the length of subject needed. It is not necessarily the greatest lower
1693        bound, but using it to avoid searching strings that are too short does give
1694        some useful speed-ups. The value is available to calling programs via
1695        pcre_fullinfo().
1696    
1697    26. While implementing 25, I discovered to my embarrassment that pcretest had
1698        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1699        study optimizations had never been tested with that matching function.
1700        Oops. What is worse, even when it was passed study data, there was a bug in
1701        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1702        were also very few tests of studied patterns with pcre_dfa_exec().
1703    
1704    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1705        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1706        on the other side of the coin, they are no longer allowed to have different
1707        names, because these cannot be distinguished in PCRE, and this has caused
1708        confusion. (This is a difference from Perl.)
1709    
1710    28. When duplicate subpattern names are present (necessarily with different
1711        numbers, as required by 27 above), and a test is made by name in a
1712        conditional pattern, either for a subpattern having been matched, or for
1713        recursion in such a pattern, all the associated numbered subpatterns are
1714        tested, and the overall condition is true if the condition is true for any
1715        one of them. This is the way Perl works, and is also more like the way
1716        testing by number works.
1717    
1718    
1719    Version 7.9 11-Apr-09
1720    ---------------------
1721    
1722    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1723        (pcretest), all targets were linked against these libraries. This included
1724        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1725        libraries. This caused unwanted dependencies to be created. This problem
1726        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1727        pcretest is linked with readline.
1728    
1729    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1730        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1731        moved outside it again, because FALSE and TRUE are already defined in AIX,
1732        but BOOL is not.
1733    
1734    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1735        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1736    
1737    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1738        hyphens) following file names and line numbers when outputting matching
1739        lines. This is not true; no spaces are inserted. I have also clarified the
1740        wording for the --colour (or --color) option.
1741    
1742    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1743        was not coloured; this is different to GNU grep, so I have changed it to be
1744        the same.
1745    
1746    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1747        each matching line was coloured. Now it goes on to look for further matches
1748        of any of the test patterns, which is the same behaviour as GNU grep.
1749    
1750    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1751        doesn't make sense to accept an empty string match in pcregrep, so I have
1752        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1753        seems to be how GNU grep behaves. [But see later change 40 for release
1754        8.33.]
1755    
1756    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1757        start or after a newline", because the conditional assertion was not being
1758        correctly handled. The rule now is that both the assertion and what follows
1759        in the first alternative must satisfy the test.
1760    
1761    9.  If auto-callout was enabled in a pattern with a conditional group whose
1762        condition was an assertion, PCRE could crash during matching, both with
1763        pcre_exec() and pcre_dfa_exec().
1764    
1765    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1766        used for matching.
1767    
1768    11. Unicode property support in character classes was not working for
1769        characters (bytes) greater than 127 when not in UTF-8 mode.
1770    
1771    12. Added the -M command line option to pcretest.
1772    
1773    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1774    
1775    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1776    
1777    16. Added comments and documentation about mis-use of no_arg in the C++
1778        wrapper.
1779    
1780    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1781        from Martin Jerabek that uses macro names for all relevant character and
1782        string constants.
1783    
1784    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1785        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1786        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1787        these, but not everybody uses configure.
1788    
1789    19. A conditional group that had only one branch was not being correctly
1790        recognized as an item that could match an empty string. This meant that an
1791        enclosing group might also not be so recognized, causing infinite looping
1792        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1793        with the subject "ab", where knowledge that the repeated group can match
1794        nothing is needed in order to break the loop.
1795    
1796    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1797        exec(), but without supplying a callout function, matching went wrong.
1798    
1799    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1800        leak if the size of the offset vector was greater than 30. When the vector
1801        is smaller, the saved offsets during recursion go onto a local stack
1802        vector, but for larger vectors malloc() is used. It was failing to free
1803        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1804        error, in fact).
1805    
1806    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1807        heapframe that is used only when UTF-8 support is enabled. This caused no
1808        problem, but was untidy.
1809    
1810    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1811        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1812        included within another project.
1813    
1814    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1815        slightly modified by me:
1816    
1817          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1818              not building pcregrep.
1819    
1820          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1821              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1822    
1823    25. Forward references, both numeric and by name, in patterns that made use of
1824        duplicate group numbers, could behave incorrectly or give incorrect errors,
1825        because when scanning forward to find the reference group, PCRE was not
1826        taking into account the duplicate group numbers. A pattern such as
1827        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1828    
1829    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1830        the feature of a custom pointer more persuasive (as requested by a user).
1831    
1832    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1833        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1834        pre-defined.
1835    
1836    28. Added support for (*UTF8) at the start of a pattern.
1837    
1838    29. Arrange for flags added by the "release type" setting in CMake to be shown
1839        in the configuration summary.
1840    
1841    
1842    Version 7.8 05-Sep-08
1843    ---------------------
1844    
1845    1.  Replaced UCP searching code with optimized version as implemented for Ad
1846        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1847        stage table and inline lookup instead of a function, giving speed ups of 2
1848        to 5 times on some simple patterns that I tested. Permission was given to
1849        distribute the MultiStage2.py script that generates the tables (it's not in
1850        the tarball, but is in the Subversion repository).
1851    
1852    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1853        scripts.
1854    
1855    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1856        a group with a zero qualifier. The result of the study could be incorrect,
1857        or the function might crash, depending on the pattern.
1858    
1859    4.  Caseless matching was not working for non-ASCII characters in back
1860        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1861        It now works when Unicode Property Support is available.
1862    
1863    5.  In pcretest, an escape such as \x{de} in the data was always generating
1864        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1865        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1866        truncation.
1867    
1868    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1869    
1870    7.  Added two (int) casts to pcregrep when printing the difference of two
1871        pointers, in case they are 64-bit values.
1872    
1873    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1874        test 2 if it fails.
1875    
1876    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1877        and a #define of that name to empty if it is not externally set. This is to
1878        allow users of MSVC to set it if necessary.
1879    
1880    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1881        the convenience functions in the pcre_get.c source file.
1882    
1883    11. An option change at the start of a pattern that had top-level alternatives
1884        could cause overwriting and/or a crash. This command provoked a crash in
1885        some environments:
1886    
1887          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1888    
1889        This potential security problem was recorded as CVE-2008-2371.
1890    
1891    12. For a pattern where the match had to start at the beginning or immediately
1892        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1893        pcre_dfa_exec() could read past the end of the passed subject if there was
1894        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1895        pcretest so that it places the subject at the end of its malloc-ed buffer.
1896    
1897    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1898        exec() might read past the end of the data buffer in UTF-8 mode.
1899    
1900    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1901        the data contained the byte 0x85 as part of a UTF-8 character within its
1902        first line. This applied both to normal and DFA matching.
1903    
1904    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1905        /^[^d]*?$/8 failed to match "abc".
1906    
1907    16. Added a missing copyright notice to pcrecpp_internal.h.
1908    
1909    17. Make it more clear in the documentation that values returned from
1910        pcre_exec() in ovector are byte offsets, not character counts.
1911    
1912    18. Tidied a few places to stop certain compilers from issuing warnings.
1913    
1914    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1915        supplied by Stefan Weber. I made a further small update for 7.8 because
1916        there is a change of source arrangements: the pcre_searchfuncs.c module is
1917        replaced by pcre_ucd.c.
1918    
1919    
1920    Version 7.7 07-May-08
1921    ---------------------
1922    
1923    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1924        a string to a long long, pretend we don't even have a long long." This is
1925        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1926    
1927    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1928        pre-7.6 versions, which defined a global no_arg variable instead of putting
1929        it in the RE class. (See also #8 below.)
1930    
1931    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1932        Lopes.
1933    
1934    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1935    
1936        (1) The include/exclude patterns were being applied to the whole pathnames
1937            of files, instead of just to the final components.
1938    
1939        (2) If there was more than one level of directory, the subdirectories were
1940            skipped unless they satisfied the include/exclude conditions. This is
1941            inconsistent with GNU grep (and could even be seen as contrary to the
1942            pcregrep specification - which I improved to make it absolutely clear).
1943            The action now is always to scan all levels of directory, and just
1944            apply the include/exclude patterns to regular files.
1945    
1946    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1947        --exclude_dir in the tests to avoid scanning .svn directories.
1948    
1949    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1950        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1951        doesn't support NULs in patterns.
1952    
1953    7.  Added some missing "const"s to declarations of static tables in
1954        pcre_compile.c and pcre_dfa_exec.c.
1955    
1956    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1957        caused by fix #2  above. (Subsequently also a second patch to fix the
1958        first patch. And a third patch - this was a messy problem.)
1959    
1960    9.  Applied Craig's patch to remove the use of push_back().
1961    
1962    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1963        matching function regexec().
1964    
1965    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1966        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1967        references. PCRE supports relative numbers with this syntax (I don't think
1968        Oniguruma does).
1969    
1970    12. Previously, a group with a zero repeat such as (...){0} was completely
1971        omitted from the compiled regex. However, this means that if the group
1972        was called as a subroutine from elsewhere in the pattern, things went wrong
1973        (an internal error was given). Such groups are now left in the compiled
1974        pattern, with a new opcode that causes them to be skipped at execution
1975        time.
1976    
1977    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1978        to the way PCRE behaves:
1979    
1980        (a) A lone ] character is dis-allowed (Perl treats it as data).
1981    
1982        (b) A back reference to an unmatched subpattern matches an empty string
1983            (Perl fails the current match path).
1984    
1985        (c) A data ] in a character class must be notated as \] because if the
1986            first data character in a class is ], it defines an empty class. (In
1987            Perl it is not possible to have an empty class.) The empty class []
1988            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1989            The negative empty class [^] matches any one character, independently
1990            of the DOTALL setting.
1991    
1992    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1993        non-existent subpattern following a character class starting with ']' and
1994        containing () gave an internal compiling error instead of "reference to
1995        non-existent subpattern". Fortunately, when the pattern did exist, the
1996        compiled code was correct. (When scanning forwards to check for the
1997        existence of the subpattern, it was treating the data ']' as terminating
1998        the class, so got the count wrong. When actually compiling, the reference
1999        was subsequently set up correctly.)
2000    
2001    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2002        it was being rejected as not supported by pcre_dfa_exec(), even though
2003        other assertions are supported. I have made pcre_dfa_exec() support
2004        (*FAIL).
2005    
2006    16. The implementation of 13c above involved the invention of a new opcode,
2007        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2008        cannot be changed at match time, I realized I could make a small
2009        improvement to matching performance by compiling OP_ALLANY instead of
2010        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2011        on the OP_ANY path.
2012    
2013    17. Compiling pcretest on Windows with readline support failed without the
2014        following two fixes: (1) Make the unistd.h include conditional on
2015        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2016    
2017    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2018        ncurses library to be included for pcretest when ReadLine support is
2019        requested, but also to allow for it to be overridden. This patch came from
2020        Daniel Bergström.
2021    
2022    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2023        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2024        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2025        spotting this.
2026    
2027    
2028    Version 7.6 28-Jan-08
2029    ---------------------
2030    
2031    1.  A character class containing a very large number of characters with
2032        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2033        overflow.
2034    
2035    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2036        HAVE_LONG_LONG is not defined.
2037    
2038    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2039        bring them up to date and include new features. This patch includes:
2040    
2041        - Fixed PH's badly added libz and libbz2 support.
2042        - Fixed a problem with static linking.
2043        - Added pcredemo. [But later removed - see 7 below.]
2044        - Fixed dftables problem and added an option.
2045        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2046            HAVE_LONG_LONG.
2047        - Added readline support for pcretest.
2048        - Added an listing of the option settings after cmake has run.
2049    
2050    4.  A user submitted a patch to Makefile that makes it easy to create
2051        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2052        Makefile.am that cause it to include this special target, without
2053        affecting anything else. Note that the same mingw target plus all
2054        the other distribution libraries and programs are now supported
2055        when configuring with CMake (see 6 below) instead of with
2056        Configure/Make.
2057    
2058    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2059        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2060        exported in the Windows port". It has not yet been confirmed that the patch
2061        solves the problem, but it does no harm.
2062    
2063    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2064        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2065        with CMake, and also correct the comment about stack recursion.
2066    
2067    7.  Remove the automatic building of pcredemo from the ./configure system and
2068        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2069        of a program that users should build themselves after PCRE is installed, so
2070        building it automatically is not really right. What is more, it gave
2071        trouble in some build environments.
2072    
2073    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2074    
2075    
2076    Version 7.5 10-Jan-08
2077    ---------------------
2078    
2079    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2080        values in parens when parsing an RE using the C++ wrapper."
2081    
2082    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2083        Characters greater than 255 were excluded from the class instead of being
2084        included.
2085    
2086    3.  The same bug as (2) above applied to negated POSIX classes such as
2087        [:^space:].
2088    
2089    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2090        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2091        I have changed it.
2092    
2093    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2094        first named subpattern) and a construct such as (?&a) would reference the
2095        first named subpattern whose name started with "a" (in other words, the
2096        length check was missing). Both these problems are fixed. "Subpattern name
2097        expected" is now given for (?&) (a zero-length name), and this patch also
2098        makes it give the same error for \k'' (previously it complained that that
2099        was a reference to a non-existent subpattern).
2100    
2101    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2102        this is right because (?- can be followed by option settings as well as by
2103        digits. I have, however, made the messages clearer.
2104    
2105    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2106        than the number used in the conditional) now cause a compile-time error.
2107        This is actually not compatible with Perl, which accepts such patterns, but
2108        treats the conditional as always being FALSE (as PCRE used to), but it
2109        seems to me that giving a diagnostic is better.
2110    
2111    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2112        and messages.
2113    
2114    9.  Fix two occurrences of "backslash" in comments that should have been
2115        "backspace".
2116    
2117    10. Remove two redundant lines of code that can never be obeyed (their function
2118        was moved elsewhere).
2119    
2120    11. The program that makes PCRE's Unicode character property table had a bug
2121        which caused it to generate incorrect table entries for sequences of
2122        characters that have the same character type, but are in different scripts.
2123        It amalgamated them into a single range, with the script of the first of
2124        them. In other words, some characters were in the wrong script. There were
2125        thirteen such cases, affecting characters in the following ranges:
2126    
2127          U+002b0 - U+002c1
2128          U+0060c - U+0060d
2129          U+0061e - U+00612
2130          U+0064b - U+0065e
2131          U+0074d - U+0076d
2132          U+01800 - U+01805
2133          U+01d00 - U+01d77
2134          U+01d9b - U+01dbf
2135          U+0200b - U+0200f
2136          U+030fc - U+030fe
2137          U+03260 - U+0327f
2138          U+0fb46 - U+0fbb1
2139          U+10450 - U+1049d
2140    
2141    12. The -o option (show only the matching part of a line) for pcregrep was not
2142        compatible with GNU grep in that, if there was more than one match in a
2143        line, it showed only the first of them. It now behaves in the same way as
2144        GNU grep.
2145    
2146    13. If the -o and -v options were combined for pcregrep, it printed a blank
2147        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2148        does the same. The return code can be used to tell if there were any
2149        non-matching lines.
2150    
2151    14. Added --file-offsets and --line-offsets to pcregrep.
2152    
2153    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2154        infinitely looping recursion. The bug was that positive lookaheads were not
2155        being skipped when checking for a possible empty match (negative lookaheads
2156        and both kinds of lookbehind were skipped).
2157    
2158    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2159        inclusion of <windows.h> to before rather than after the definition of
2160        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2161    
2162    17. Specifying a possessive quantifier with a specific limit for a Unicode
2163        character property caused pcre_compile() to compile bad code, which led at
2164        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2165        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2166        caused the error; without that there was no problem.
2167    
2168    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2169    
2170    19. Added --enable-pcretest-libreadline.
2171    
2172    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2173        RE::GlobalReplace(). As a result, the number of replacements returned was
2174        double what it should be. I removed one of the increments, but Craig sent a
2175        later patch that removed the other one (the right fix) and added unit tests
2176        that check the return values (which was not done before).
2177    
2178    21. Several CMake things:
2179    
2180        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2181            the names libpcre and libpcreposix, not just pcre and pcreposix.
2182    
2183        (2) The above change means that pcretest and pcregrep are now correctly
2184            linked with the newly-built libraries, not previously installed ones.
2185    
2186        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2187    
2188    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2189        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2190        UTF-8 newline character). The key issue is that the pattern starts .*;
2191        this means that the match must be either at the beginning, or after a
2192        newline. The bug was in the code for advancing after a failed match and
2193        checking that the new position followed a newline. It was not taking
2194        account of UTF-8 characters correctly.
2195    
2196    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2197        character classes. PCRE was not treating the sequence [:...:] as a
2198        character class unless the ... were all letters. Perl, however, seems to
2199        allow any characters between [: and :], though of course it rejects as
2200        unknown any "names" that contain non-letters, because all the known class
2201        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2202        for example, whereas PCRE did not - it did not recognize a POSIX character
2203        class. This seemed a bit dangerous, so the code has been changed to be
2204        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2205        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2206        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2207        Perl does, and where it didn't before.
2208    
2209    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2210        Windows environments %n is disabled by default.
2211    
2212    
2213    Version 7.4 21-Sep-07
2214    ---------------------
2215    
2216    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2217        means that a class such as [\s] counted as "explicit reference to CR or
2218        LF". That isn't really right - the whole point of the change was to try to
2219        help when there was an actual mention of one of the two characters. So now
2220        the change happens only if \r or \n (or a literal CR or LF) character is
2221        encountered.
2222    
2223    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2224        of both had grown to the point where there were only 3 bits left.
2225        Fortunately, there was spare space in the data structure, and so I have
2226        moved the internal flags into a new 16-bit field to free up more option
2227        bits.
2228    
2229    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2230        but did not set the internal JCHANGED flag - either of these is enough to
2231        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2232        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2233        start sets both bits.
2234    
2235    4.  Added options (at build time, compile time, exec time) to change \R from
2236        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2237    
2238    5.  doc/pcresyntax.html was missing from the distribution.
2239    
2240    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2241        compatibility, even though it is no longer used.
2242    
2243    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2244        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2245        windows.h file is present (where different names are used). [This was
2246        reversed later after testing - see 16 below.]
2247    
2248    8.  Changed all #include <config.h> to #include "config.h". There were also
2249        some further <pcre.h> cases that I changed to "pcre.h".
2250    
2251    9.  When pcregrep was used with the --colour option, it missed the line ending
2252        sequence off the lines that it output.
2253    
2254    10. It was pointed out to me that arrays of string pointers cause lots of
2255        relocations when a shared library is dynamically loaded. A technique of
2256        using a single long string with a table of offsets can drastically reduce
2257        these. I have refactored PCRE in four places to do this. The result is
2258        dramatic:
2259    
2260          Originally:                          290
2261          After changing UCP table:            187
2262          After changing error message table:   43
2263          After changing table of "verbs"       36
2264          After changing table of Posix names   22
2265    
2266        Thanks to the folks working on Gregex for glib for this insight.
2267    
2268    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2269        unicode-properties was also set.
2270    
2271    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2272    
2273    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2274        checked only for CRLF.
2275    
2276    14. Added casts to pcretest.c to avoid compiler warnings.
2277    
2278    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2279    
2280    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2281        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2282        entirely. This removes changes made in 7 above.
2283    
2284    17. The CMake files have been updated, and there is now more information about
2285        building with CMake in the NON-UNIX-USE document.
2286    
2287    
2288    Version 7.3 28-Aug-07
2289  ---------------------  ---------------------
2290    
2291   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 9  Version 7.3 05-Jul-07 Line 2293  Version 7.3 05-Jul-07
2293      brackets there is not right, since it causes compilers to look for an      brackets there is not right, since it causes compilers to look for an
2294      installed pcre.h, not the version that is in the source that is being      installed pcre.h, not the version that is in the source that is being
2295      compiled (which of course may be different). I have changed it back to:      compiled (which of course may be different). I have changed it back to:
2296    
2297        #include "pcre.h"        #include "pcre.h"
2298    
2299      I have a vague recollection that the change was concerned with compiling in      I have a vague recollection that the change was concerned with compiling in
2300      different directories, but in the new build system, that is taken care of      different directories, but in the new build system, that is taken care of
2301      by the VPATH setting the Makefile.      by the VPATH setting the Makefile.
2302    
2303   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed   2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2304      when the subject happened to end in the byte 0x85 (e.g. if the last      when the subject happened to end in the byte 0x85 (e.g. if the last
2305      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline      character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
# Line 23  Version 7.3 05-Jul-07 Line 2307  Version 7.3 05-Jul-07
2307      of another character. The bug was that, for an unlimited repeat of . in      of another character. The bug was that, for an unlimited repeat of . in
2308      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by      not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2309      characters when looking for a newline.      characters when looking for a newline.
   
  3. A small performance improvement in the DOTALL UTF-8 mode .* case.  
   
  4. Debugging: adjusted the names of opcodes for different kinds of parentheses  
     in debug output.  
   
  5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp  
     unittest when running under MinGW.  
2310    
2311     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2312    
2313     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2314        in debug output.
2315    
2316     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2317        long printing in the pcrecpp unittest when running under MinGW.
2318    
2319     6. ESC_K was left out of the EBCDIC table.
2320    
2321     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2322        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2323        limit also applies to "virtual nesting" when a pattern is recursive, and in
2324        this case 1000 isn't so big. I have been able to remove this limit at the
2325        expense of backing off one optimization in certain circumstances. Normally,
2326        when pcre_exec() would call its internal match() function recursively and
2327        immediately return the result unconditionally, it uses a "tail recursion"
2328        feature to save stack. However, when a subpattern that can match an empty
2329        string has an unlimited repetition quantifier, it no longer makes this
2330        optimization. That gives it a stack frame in which to save the data for
2331        checking that an empty string has been matched. Previously this was taken
2332        from the 1000-entry workspace that had been reserved. So now there is no
2333        explicit limit, but more stack is used.
2334    
2335     8. Applied Daniel's patches to solve problems with the import/export magic
2336        syntax that is required for Windows, and which was going wrong for the
2337        pcreposix and pcrecpp parts of the library. These were overlooked when this
2338        problem was solved for the main library.
2339    
2340     9. There were some crude static tests to avoid integer overflow when computing
2341        the size of patterns that contain repeated groups with explicit upper
2342        limits. As the maximum quantifier is 65535, the maximum group length was
2343        set at 30,000 so that the product of these two numbers did not overflow a
2344        32-bit integer. However, it turns out that people want to use groups that
2345        are longer than 30,000 bytes (though not repeat them that many times).
2346        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2347        made it possible to implement the integer overflow checks in a much more
2348        dynamic way, which I have now done. The artificial limitation on group
2349        length has been removed - we now have only the limit on the total length of
2350        the compiled pattern, which depends on the LINK_SIZE setting.
2351    
2352    10. Fixed a bug in the documentation for get/copy named substring when
2353        duplicate names are permitted. If none of the named substrings are set, the
2354        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2355        empty string.
2356    
2357    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2358        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2359        because the ] is interpreted as the first data character and the
2360        terminating ] is not found. PCRE has been made compatible with Perl in this
2361        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2362        cause memory overwriting.
2363    
2364    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2365        string has been matched (to stop an infinite loop). It was not recognizing
2366        a conditional subpattern that could match an empty string if that
2367        subpattern was within another subpattern. For example, it looped when
2368        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2369        condition was not nested. This bug has been fixed.
2370    
2371    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2372        past the start of the subject in the presence of bytes with the top bit
2373        set, for example "\x8aBCD".
2374    
2375    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2376        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2377    
2378    14. Optimized (?!) to (*FAIL).
2379    
2380    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2381        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2382        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2383        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2384        does: it's just the validity check that is more restrictive.
2385    
2386    16. Inserted checks for integer overflows during escape sequence (backslash)
2387        processing, and also fixed erroneous offset values for syntax errors during
2388        backslash processing.
2389    
2390    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2391        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2392    
2393    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2394        caused an overrun.
2395    
2396    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2397        something other than just ASCII characters) inside a group that had an
2398        unlimited repeat caused a loop at compile time (while checking to see
2399        whether the group could match an empty string).
2400    
2401    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2402        [\P{Any}] did so. (Error in the code for printing property names.)
2403    
2404    21. An orphan \E inside a character class could cause a crash.
2405    
2406    22. A repeated capturing bracket such as (A)? could cause a wild memory
2407        reference during compilation.
2408    
2409    23. There are several functions in pcre_compile() that scan along a compiled
2410        expression for various reasons (e.g. to see if it's fixed length for look
2411        behind). There were bugs in these functions when a repeated \p or \P was
2412        present in the pattern. These operators have additional parameters compared
2413        with \d, etc, and these were not being taken into account when moving along
2414        the compiled data. Specifically:
2415    
2416        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2417            length.
2418    
2419        (b) An item such as \pL+ within a repeated group could cause crashes or
2420            loops.
2421    
2422        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2423            "reference to non-existent subpattern" error.
2424    
2425        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2426    
2427    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2428        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2429    
2430    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2431    
2432    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2433        character were causing crashes (broken optimization).
2434    
2435    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2436        \p or \P) caused a compile-time loop.
2437    
2438    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2439        break. For example, the unstudied pattern [\r\n]A does not match the string
2440        "\r\nA" because change 7.0/46 below moves the current point on by two
2441        characters after failing to match at the start. However, the pattern \nA
2442        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2443        the same is true. There doesn't seem any very clean way out of this, but
2444        what I have chosen to do makes the common cases work: PCRE now takes note
2445        of whether there can be an explicit match for \r or \n anywhere in the
2446        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2447        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2448        pattern has explicit CR or LF references.
2449    
2450    29. Added (*CR) etc for changing newline setting at start of pattern.
2451    
2452    
2453  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.193  
changed lines
  Added in v.1359

  ViewVC Help
Powered by ViewVC 1.1.5