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

Diff of /code/trunk/ChangeLog

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

revision 127 by ph10, Mon Mar 19 11:44:45 2007 UTC revision 1478 by ph10, Tue May 27 13:18:31 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 12-Mar-07  Version 8.36 xx-xxx-2014
5    ------------------------
6    
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
9    
10    2.  The tests for quantifiers being too big (greater than 65535) were being
11        applied after reading the number, and stupidly assuming that integer
12        overflow would give a negative number. The tests are now applied as the
13        numbers are read.
14    
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
17    
18    4.  The JIT compiler did not generate match limit checks for certain
19        bracketed expressions with quantifiers. This may lead to exponential
20        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21        issue should be resolved now.
22    
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
25    
26    6.  Inserted two casts and changed some ints to size_t in the light of some
27        reported 64-bit compiler warnings (Bugzilla 1477).
28    
29    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
30        an empty string, which sometimes were behaving incorrectly in the
31        interpreter (though correctly in the JIT matcher). This pcretest input is
32        an example:
33    
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36    
37        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39        for an empty string was breaking the inner loop and carrying on at a lower
40        level, when possessive repeated groups should always return to a higher
41        level as they have no backtrack points in them. The empty string test now
42        occurs at the outer level.
43    
44    
45    Version 8.35 04-April-2014
46    --------------------------
47    
48    1.  A new flag is set, when property checks are present in an XCLASS.
49        When this flag is not set, PCRE can perform certain optimizations
50        such as studying these XCLASS-es.
51    
52    2.  The auto-possessification of character sets were improved: a normal
53        and an extended character set can be compared now. Furthermore
54        the JIT compiler optimizes more character set checks.
55    
56    3.  Got rid of some compiler warnings for potentially uninitialized variables
57        that show up only when compiled with -O2.
58    
59    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
60        of a match later then the end of the match. The pcretest program was not
61        handling the case sensibly - it was outputting from the start to the next
62        binary zero. It now reports this situation in a message, and outputs the
63        text from the end to the start.
64    
65    5.  Fast forward search is improved in JIT. Instead of the first three
66        characters, any three characters with fixed position can be searched.
67        Search order: first, last, middle.
68    
69    6.  Improve character range checks in JIT. Characters are read by an inprecise
70        function now, which returns with an unknown value if the character code is
71        above a certain treshold (e.g: 256). The only limitation is that the value
72        must be bigger than the treshold as well. This function is useful, when
73        the characters above the treshold are handled in the same way.
74    
75    7.  The macros whose names start with RAWUCHAR are placeholders for a future
76        mode in which only the bottom 21 bits of 32-bit data items are used. To
77        make this more memorable for those maintaining the code, the names have
78        been changed to start with UCHAR21, and an extensive comment has been added
79        to their definition.
80    
81    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
82        to the export list in Makefile.am (they were accidentally omitted from the
83        8.34 tarball).
84    
85    9.  The informational output from pcretest used the phrase "starting byte set"
86        which is inappropriate for the 16-bit and 32-bit libraries. As the output
87        for "first char" and "need char" really means "non-UTF-char", I've changed
88        "byte" to "char", and slightly reworded the output. The documentation about
89        these values has also been (I hope) clarified.
90    
91    10. Another JIT related optimization: use table jumps for selecting the correct
92        backtracking path, when more than four alternatives are present inside a
93        bracket.
94    
95    11. Empty match is not possible, when the minimum length is greater than zero,
96        and there is no \K in the pattern. JIT should avoid empty match checks in
97        such cases.
98    
99    12. In a caseless character class with UCP support, when a character with more
100        than one alternative case was not the first character of a range, not all
101        the alternative cases were added to the class. For example, s and \x{17f}
102        are both alternative cases for S: the class [RST] was handled correctly,
103        but [R-T] was not.
104    
105    13. The configure.ac file always checked for pthread support when JIT was
106        enabled. This is not used in Windows, so I have put this test inside a
107        check for the presence of windows.h (which was already tested for).
108    
109    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
110        The algorithm provides a way to skip certain starting offsets, and usually
111        faster than linear prefix searches.
112    
113    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
114        as for 'fr_FR' and 'french'. For some reason, however, it then used the
115        Windows-specific input and output files, which have 'french' screwed in.
116        So this could never have worked. One of the problems with locales is that
117        they aren't always the same. I have now updated RunTest so that it checks
118        the output of the locale test (test 3) against three different output
119        files, and it allows the test to pass if any one of them matches. With luck
120        this should make the test pass on some versions of Solaris where it was
121        failing. Because of the uncertainty, the script did not used to stop if
122        test 3 failed; it now does. If further versions of a French locale ever
123        come to light, they can now easily be added.
124    
125    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
126        there was a message during ./configure, but it did not stop. This now
127        provokes an error. The invalid example in README has been corrected.
128        If a value less than the minimum is given, the minimum value has always
129        been used, but now a warning is given.
130    
131    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
132        was a bug in the test system, which is now fixed. Also, the list of various
133        configurations that are tested for each release did not have one with both
134        16/32 bits and --enable-bar-anycrlf. It now does.
135    
136    18. pcretest was missing "-C bsr" for displaying the \R default setting.
137    
138    19. Little endian PowerPC systems are supported now by the JIT compiler.
139    
140    20. The fast forward newline mechanism could enter to an infinite loop on
141        certain invalid UTF-8 input. Although we don't support these cases
142        this issue can be fixed by a performance optimization.
143    
144    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
145        not take account if existing stack usage. There is now a new global
146        variable called pcre_stack_guard that can be set to point to an external
147        function to check stack availability. It is called at the start of
148        processing every parenthesized group.
149    
150    22. A typo in the code meant that in ungreedy mode the max/min qualifier
151        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
152        not match "ab".
153    
154    23. When UTF was disabled, the JIT program reported some incorrect compile
155        errors. These messages are silenced now.
156    
157    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
158        compiler.
159    
160    25. Change all the temporary files used in RunGrepTest to be different to those
161        used by RunTest so that the tests can be run simultaneously, for example by
162        "make -j check".
163    
164    
165    Version 8.34 15-December-2013
166    -----------------------------
167    
168    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
169        executable memory. Patch inspired by Carsten Klein.
170    
171    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
172        this macro is never tested and has no effect, because the work to support
173        coverage involves only compiling and linking options and special targets in
174        the Makefile. The comment in config.h implied that defining the macro would
175        enable coverage support, which is totally false. There was also support for
176        setting this macro in the CMake files (my fault, I just copied it from
177        configure). SUPPORT_GCOV has now been removed.
178    
179    3.  Make a small performance improvement in strlen16() and strlen32() in
180        pcretest.
181    
182    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
183        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
184        these cases). There was also one in pcretest.c.
185    
186    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
187    
188    6.  In UTF mode, the code for checking whether a group could match an empty
189        string (which is used for indefinitely repeated groups to allow for
190        breaking an infinite loop) was broken when the group contained a repeated
191        negated single-character class with a character that occupied more than one
192        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
193        UTF-8 mode). The effect was undefined: the group might or might not be
194        deemed as matching an empty string, or the program might have crashed.
195    
196    7.  The code for checking whether a group could match an empty string was not
197        recognizing that \h, \H, \v, \V, and \R must match a character.
198    
199    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
200        an empty string. If it can, pcretest shows this in its information output.
201    
202    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
203        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
204        matched by pcre_exec() without using JIT:
205    
206        (a) If the rest of the pattern did not match after a maximal run of
207            grapheme clusters, the code for backing up to try with fewer of them
208            did not always back up over a full grapheme when characters that do not
209            have the modifier quality were involved, e.g. Hangul syllables.
210    
211        (b) If the match point in a subject started with modifier character, and
212            there was no match, the code could incorrectly back up beyond the match
213            point, and potentially beyond the first character in the subject,
214            leading to a segfault or an incorrect match result.
215    
216    10. A conditional group with an assertion condition could lead to PCRE
217        recording an incorrect first data item for a match if no other first data
218        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
219        first data item, and therefore matched "ca" after "c" instead of at the
220        start.
221    
222    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
223        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
224    
225    12. The source of pcregrep now includes z/OS-specific code so that it can be
226        compiled for z/OS as part of the special z/OS distribution.
227    
228    13. Added the -T and -TM options to pcretest.
229    
230    14. The code in pcre_compile.c for creating the table of named capturing groups
231        has been refactored. Instead of creating the table dynamically during the
232        actual compiling pass, the information is remembered during the pre-compile
233        pass (on the stack unless there are more than 20 named groups, in which
234        case malloc() is used) and the whole table is created before the actual
235        compile happens. This has simplified the code (it is now nearly 150 lines
236        shorter) and prepared the way for better handling of references to groups
237        with duplicate names.
238    
239    15. A back reference to a named subpattern when there is more than one of the
240        same name now checks them in the order in which they appear in the pattern.
241        The first one that is set is used for the reference. Previously only the
242        first one was inspected. This change makes PCRE more compatible with Perl.
243    
244    16. Unicode character properties were updated from Unicode 6.3.0.
245    
246    17. The compile-time code for auto-possessification has been refactored, based
247        on a patch by Zoltan Herczeg. It now happens after instead of during
248        compilation. The code is cleaner, and more cases are handled. The option
249        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
250        options in pcretest are provided to set it. It can also be set by
251        (*NO_AUTO_POSSESS) at the start of a pattern.
252    
253    18. The character VT has been added to the default ("C" locale) set of
254        characters that match \s and are generally treated as white space,
255        following this same change in Perl 5.18. There is now no difference between
256        "Perl space" and "POSIX space". Whether VT is treated as white space in
257        other locales depends on the locale.
258    
259    19. The code for checking named groups as conditions, either for being set or
260        for being recursed, has been refactored (this is related to 14 and 15
261        above). Processing unduplicated named groups should now be as fast at
262        numerical groups, and processing duplicated groups should be faster than
263        before.
264    
265    20. Two patches to the CMake build system, by Alexander Barkov:
266    
267          (1) Replace the "source" command by "." in CMakeLists.txt because
268              "source" is a bash-ism.
269    
270          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
271              without these the CMake build does not work on Solaris.
272    
273    21. Perl has changed its handling of \8 and \9. If there is no previously
274        encountered capturing group of those numbers, they are treated as the
275        literal characters 8 and 9 instead of a binary zero followed by the
276        literals. PCRE now does the same.
277    
278    22. Following Perl, added \o{} to specify codepoints in octal, making it
279        possible to specify values greater than 0777 and also making them
280        unambiguous.
281    
282    23. Perl now gives an error for missing closing braces after \x{... instead of
283        treating the string as literal. PCRE now does the same.
284    
285    24. RunTest used to grumble if an inappropriate test was selected explicitly,
286        but just skip it when running all tests. This make it awkward to run ranges
287        of tests when one of them was inappropriate. Now it just skips any
288        inappropriate tests, as it always did when running all tests.
289    
290    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
291        character types such as \d or \w, too many callouts were inserted, and the
292        data that they returned was rubbish.
293    
294    26. In UCP mode, \s was not matching two of the characters that Perl matches,
295        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
296        were matched by \h. The code has now been refactored so that the lists of
297        the horizontal and vertical whitespace characters used for \h and \v (which
298        are defined only in one place) are now also used for \s.
299    
300    27. Add JIT support for the 64 bit TileGX architecture.
301        Patch by Jiong Wang (Tilera Corporation).
302    
303    28. Possessive quantifiers for classes (both explicit and automatically
304        generated) now use special opcodes instead of wrapping in ONCE brackets.
305    
306    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
307        (because it's meaningless), this was not happening when PCRE_CASELESS was
308        set. Not wrong, but inefficient.
309    
310    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
311        properties for \w, \d, etc) is present in a test regex. Otherwise if the
312        test contains no characters greater than 255, Perl doesn't realise it
313        should be using Unicode semantics.
314    
315    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
316        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
317        does in Unicode mode.
318    
319    32. Added the "forbid" facility to pcretest so that putting tests into the
320        wrong test files can sometimes be quickly detected.
321    
322    33. There is now a limit (default 250) on the depth of nesting of parentheses.
323        This limit is imposed to control the amount of system stack used at compile
324        time. It can be changed at build time by --with-parens-nest-limit=xxx or
325        the equivalent in CMake.
326    
327    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
328        errors. Perl warns for these when in warning mode, but PCRE has no facility
329        for giving warnings.
330    
331    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
332        However, this was not working for (?!) because it is optimized to (*FAIL),
333        for which PCRE does not allow quantifiers. The optimization is now disabled
334        when a quantifier follows (?!). I can't see any use for this, but it makes
335        things uniform.
336    
337    36. Perl no longer allows group names to start with digits, so I have made this
338        change also in PCRE. It simplifies the code a bit.
339    
340    37. In extended mode, Perl ignores spaces before a + that indicates a
341        possessive quantifier. PCRE allowed a space before the quantifier, but not
342        before the possessive +. It now does.
343    
344    38. The use of \K (reset reported match start) within a repeated possessive
345        group such as (a\Kb)*+ was not working.
346    
347    40. Document that the same character tables must be used at compile time and
348        run time, and that the facility to pass tables to pcre_exec() and
349        pcre_dfa_exec() is for use only with saved/restored patterns.
350    
351    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
352        features for Builds with MSVC:
353    
354        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
355           builds). The .rc files can be used to set FileDescription and many other
356           attributes.
357    
358        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
359           This allows higher-level build scripts which want .pdb files to avoid
360           hard-coding the exact files needed."
361    
362    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
363        mean "start of word" and "end of word", respectively, as a transition aid.
364    
365    43. A minimizing repeat of a class containing codepoints greater than 255 in
366        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
367        compiled to use the heap for recursion.
368    
369    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
370        is configured.
371    
372    
373    Version 8.33 28-May-2013
374    ------------------------
375    
376    1.  Added 'U' to some constants that are compared to unsigned integers, to
377        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
378        variables that are added to signed variables, to ensure the result is
379        signed and can be negated.
380    
381    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
382        CMake config files.
383    
384    3.  Revise the creation of config.h.generic so that all boolean macros are
385        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
386        overriding via -D on the command line possible.
387    
388    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
389        to unsigned int is reported to make a quite noticeable speed difference in
390        a specific Windows environment. Testing on Linux did also appear to show
391        some benefit (and it is clearly not harmful). Also fixed the definition of
392        Xop which should be unsigned.
393    
394    5.  Related to (4), changing the definition of the intermediate variable cc
395        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
396        improvements.
397    
398    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
399        unnecessary spaces.
400    
401    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
402        and later.
403    
404    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
405    
406    9.  Optimizing fast_forward_start_bits in JIT.
407    
408    10. Adding support for callouts in JIT, and fixing some issues revealed
409        during this work. Namely:
410    
411        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
412    
413        (b) Minimum length was not checked before the matching is started.
414    
415    11. The value of capture_last that is passed to callouts was incorrect in some
416        cases when there was a capture on one path that was subsequently abandoned
417        after a backtrack. Also, the capture_last value is now reset after a
418        recursion, since all captures are also reset in this case.
419    
420    12. The interpreter no longer returns the "too many substrings" error in the
421        case when an overflowing capture is in a branch that is subsequently
422        abandoned after a backtrack.
423    
424    13. In the pathological case when an offset vector of size 2 is used, pcretest
425        now prints out the matched string after a yield of 0 or 1.
426    
427    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
428        Only supported by the JIT compiler at the moment.
429    
430    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
431    
432    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
433        offset of the starting point of the matching process, provided the offsets
434        vector is large enough.
435    
436    17. The \A escape now records a lookbehind value of 1, though its execution
437        does not actually inspect the previous character. This is to ensure that,
438        in partial multi-segment matching, at least one character from the old
439        segment is retained when a new segment is processed. Otherwise, if there
440        are no lookbehinds in the pattern, \A might match incorrectly at the start
441        of a new segment.
442    
443    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
444    
445    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
446        gives some modest performance improvement in 8-bit mode.
447    
448    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
449        be expressed in certain programming languages using Universal Character
450        Names.
451    
452    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
453        which points out that "non characters" are not "characters that may not
454        appear in Unicode strings" but rather "characters that are reserved for
455        internal use and have only local meaning".
456    
457    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
458        there was a conditional group that depended on an assertion, if the
459        assertion was false, the callout that immediately followed the alternation
460        in the condition was skipped when pcre_exec() was used for matching.
461    
462    23. Allow an explicit callout to be inserted before an assertion that is the
463        condition for a conditional group, for compatibility with automatic
464        callouts, which always insert a callout at this point.
465    
466    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
467        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
468    
469    25. (*PRUNE) is now supported by the JIT compiler.
470    
471    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
472    
473    27. Fix the case where there are two or more SKIPs with arguments that may be
474        ignored.
475    
476    28. (*SKIP) is now supported by the JIT compiler.
477    
478    29. (*THEN) is now supported by the JIT compiler.
479    
480    30. Update RunTest with additional test selector options.
481    
482    31. The way PCRE handles backtracking verbs has been changed in two ways.
483    
484        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
485        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
486        backtracking. In some cases this makes it more Perl-compatible, but Perl's
487        rather obscure rules do not always do the same thing.
488    
489        (2) Previously, backtracking verbs were confined within assertions. This is
490        no longer the case for positive assertions, except for (*ACCEPT). Again,
491        this sometimes improves Perl compatibility, and sometimes does not.
492    
493    32. A number of tests that were in test 2 because Perl did things differently
494        have been moved to test 1, because either Perl or PCRE has changed, and
495        these tests are now compatible.
496    
497    32. Backtracking control verbs are now handled in the same way in JIT and
498        interpreter.
499    
500    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
501        contained a forward subroutine reference caused a compile error.
502    
503    34. Auto-detect and optimize limited repetitions in JIT.
504    
505    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
506        blocking (*UTF) etc.
507    
508    36. In the interpreter, maximizing pattern repetitions for characters and
509        character types now use tail recursion, which reduces stack usage.
510    
511    37. The value of the max lookbehind was not correctly preserved if a compiled
512        and saved regex was reloaded on a host of different endianness.
513    
514    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
515        of the compiled pattern block, expand the flags field from 16 to 32 bits
516        because it was almost full.
517    
518    39. Try madvise first before posix_madvise.
519    
520    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
521        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
522        original change locked out all matches of empty strings. This has been
523        changed so that one match of an empty string per line is recognized.
524        Subsequent searches on the same line (for colouring or for --only-matching,
525        for example) do not recognize empty strings.
526    
527    41. Applied a user patch to fix a number of spelling mistakes in comments.
528    
529    42. Data lines longer than 65536 caused pcretest to crash.
530    
531    43. Clarified the data type for length and startoffset arguments for pcre_exec
532        and pcre_dfa_exec in the function-specific man pages, where they were
533        explicitly stated to be in bytes, never having been updated. I also added
534        some clarification to the pcreapi man page.
535    
536    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
537        a segmentation fault.
538    
539    
540    Version 8.32 30-November-2012
541    -----------------------------
542    
543    1.  Improved JIT compiler optimizations for first character search and single
544        character iterators.
545    
546    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
547        Patch by Daniel Richard G.
548    
549    3.  Single character iterator optimizations in the JIT compiler.
550    
551    4.  Improved JIT compiler optimizations for character ranges.
552    
553    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
554        Reported by Giuseppe D'Angelo.
555    
556    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
557        of a line, was being set incorrectly in cases where .* appeared inside
558        atomic brackets at the start of a pattern, or where there was a subsequent
559        *PRUNE or *SKIP.
560    
561    7.  Improved instruction cache flush for POWER/PowerPC.
562        Patch by Daniel Richard G.
563    
564    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
565        grep:
566    
567        (a) There is now no limit to the number of patterns to be matched.
568    
569        (b) An error is given if a pattern is too long.
570    
571        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
572            are now supported.
573    
574        (d) --exclude-from and --include-from (multiple use) have been added.
575    
576        (e) Exclusions and inclusions now apply to all files and directories, not
577            just to those obtained from scanning a directory recursively.
578    
579        (f) Multiple uses of -f and --file-list are now supported.
580    
581        (g) In a Windows environment, the default for -d has been changed from
582            "read" (the GNU grep default) to "skip", because otherwise the presence
583            of a directory in the file list provokes an error.
584    
585        (h) The documentation has been revised and clarified in places.
586    
587    9.  Improve the matching speed of capturing brackets.
588    
589    10. Changed the meaning of \X so that it now matches a Unicode extended
590        grapheme cluster.
591    
592    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
593        out POSIX threads when JIT support is configured.
594    
595    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
596    
597    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
598        structure could go wrong in environments where size_t is not the same size
599        as int.
600    
601    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
602        set.
603    
604    15. The EBCDIC support had decayed; later updates to the code had included
605        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
606        general tidy up of EBCDIC-related issues, and the documentation was also
607        not quite right. There is now a test that can be run on ASCII systems to
608        check some of the EBCDIC-related things (but is it not a full test).
609    
610    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
611        in a small tidy to the code.
612    
613    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
614    
615    18. If the --only-matching (-o) option in pcregrep is specified multiple
616        times, each one causes appropriate output. For example, -o1 -o2 outputs the
617        substrings matched by the 1st and 2nd capturing parentheses. A separating
618        string can be specified by --om-separator (default empty).
619    
620    19. Improving the first n character searches.
621    
622    20. Turn case lists for horizontal and vertical white space into macros so that
623        they are defined only once.
624    
625    21. This set of changes together give more compatible Unicode case-folding
626        behaviour for characters that have more than one other case when UCP
627        support is available.
628    
629        (a) The Unicode property table now has offsets into a new table of sets of
630            three or more characters that are case-equivalent. The MultiStage2.py
631            script that generates these tables (the pcre_ucd.c file) now scans
632            CaseFolding.txt instead of UnicodeData.txt for character case
633            information.
634    
635        (b) The code for adding characters or ranges of characters to a character
636            class has been abstracted into a generalized function that also handles
637            case-independence. In UTF-mode with UCP support, this uses the new data
638            to handle characters with more than one other case.
639    
640        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
641            whose other case is greater than 256 are now correctly matched
642            caselessly. Previously, the high codepoint matched the low one, but not
643            vice versa.
644    
645        (d) The processing of \h, \H, \v, and \ in character classes now makes use
646            of the new class addition function, using character lists defined as
647            macros alongside the case definitions of 20 above.
648    
649        (e) Caseless back references now work with characters that have more than
650            one other case.
651    
652        (f) General caseless matching of characters with more than one other case
653            is supported.
654    
655    22. Unicode character properties were updated from Unicode 6.2.0
656    
657    23. Improved CMake support under Windows. Patch by Daniel Richard G.
658    
659    24. Add support for 32-bit character strings, and UTF-32
660    
661    25. Major JIT compiler update (code refactoring and bugfixing).
662        Experimental Sparc 32 support is added.
663    
664    26. Applied a modified version of Daniel Richard G's patch to create
665        pcre.h.generic and config.h.generic by "make" instead of in the
666        PrepareRelease script.
667    
668    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
669        pcre_compile.c when checking for a zero character.
670    
671    28. Introducing a native interface for JIT. Through this interface, the compiled
672        machine code can be directly executed. The purpose of this interface is to
673        provide fast pattern matching, so several sanity checks are not performed.
674        However, feature tests are still performed. The new interface provides
675        1.4x speedup compared to the old one.
676    
677    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
678        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
679        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
680    
681    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
682        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
683        the "old" RFC 2279). Instead, it ended up passing a negative length to
684        pcre_exec().
685    
686    31. Add support for GCC's visibility feature to hide internal functions.
687    
688    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
689        "unknown -C option" after outputting 0 or 1.
690    
691    33. There is now support for generating a code coverage report for the test
692        suite in environments where gcc is the compiler and lcov is installed. This
693        is mainly for the benefit of the developers.
694    
695    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
696        unaddressable using valgrind annotations, allowing valgrind to detect
697        invalid memory accesses. This is mainly for the benefit of the developers.
698    
699    25. (*UTF) can now be used to start a pattern in any of the three libraries.
700    
701    26. Give configure error if --enable-cpp but no C++ compiler found.
702    
703    
704    Version 8.31 06-July-2012
705    -------------------------
706    
707    1.  Fixing a wrong JIT test case and some compiler warnings.
708    
709    2.  Removed a bashism from the RunTest script.
710    
711    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
712        to unsigned type, result still unsigned" that was given by an MS compiler
713        on encountering the code "-sizeof(xxx)".
714    
715    4.  Partial matching support is added to the JIT compiler.
716    
717    5.  Fixed several bugs concerned with partial matching of items that consist
718        of more than one character:
719    
720        (a) /^(..)\1/ did not partially match "aba" because checking references was
721            done on an "all or nothing" basis. This also applied to repeated
722            references.
723    
724        (b) \R did not give a hard partial match if \r was found at the end of the
725            subject.
726    
727        (c) \X did not give a hard partial match after matching one or more
728            characters at the end of the subject.
729    
730        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
731            a partial match for the string "\r".
732    
733        (e) When newline was set to CRLF, the metacharacter "." did not recognize
734            a partial match for a CR character at the end of the subject string.
735    
736    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
737        running pcretest, the text "(JIT)" added to the output whenever JIT is
738        actually used to run the match.
739    
740    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
741        or /S+[+] with a digit between 1 and 7.
742    
743    8.  OP_NOT now supports any UTF character not just single-byte ones.
744    
745    9.  (*MARK) control verb is now supported by the JIT compiler.
746    
747    10. The command "./RunTest list" lists the available tests without actually
748        running any of them. (Because I keep forgetting what they all are.)
749    
750    11. Add PCRE_INFO_MAXLOOKBEHIND.
751    
752    12. Applied a (slightly modified) user-supplied patch that improves performance
753        when the heap is used for recursion (compiled with --disable-stack-for-
754        recursion). Instead of malloc and free for each heap frame each time a
755        logical recursion happens, frames are retained on a chain and re-used where
756        possible. This sometimes gives as much as 30% improvement.
757    
758    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
759        call.
760    
761    14. As documented, (*COMMIT) is now confined to within a positive assertion.
762    
763    15. It is now possible to link pcretest with libedit as an alternative to
764        libreadline.
765    
766    16. (*COMMIT) control verb is now supported by the JIT compiler.
767    
768    17. The Unicode data tables have been updated to Unicode 6.1.0.
769    
770    18. Added --file-list option to pcregrep.
771    
772    19. Added binary file support to pcregrep, including the -a, --binary-files,
773        -I, and --text options.
774    
775    20. The madvise function is renamed for posix_madvise for QNX compatibility
776        reasons. Fixed by Giuseppe D'Angelo.
777    
778    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
779    
780    22. Changed the option for creating symbolic links for 16-bit man pages from
781        -s to -sf so that re-installing does not cause issues.
782    
783    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
784    
785    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
786        matches in certain environments (the workspace was not being correctly
787        retained). Also added to pcre_dfa_exec() a simple plausibility check on
788        some of the workspace data at the beginning of a restart.
789    
790    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
791        was not doing so when it should - probably a typo introduced by SVN 528
792        (change 8.10/14).
793    
794    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
795        \w+ when the character tables indicated that \x{c4} was a word character.
796        There were several related cases, all because the tests for doing a table
797        lookup were testing for characters less than 127 instead of 255.
798    
799    27. If a pattern contains capturing parentheses that are not used in a match,
800        their slots in the ovector are set to -1. For those that are higher than
801        any matched groups, this happens at the end of processing. In the case when
802        there were back references that the ovector was too small to contain
803        (causing temporary malloc'd memory to be used during matching), and the
804        highest capturing number was not used, memory off the end of the ovector
805        was incorrectly being set to -1. (It was using the size of the temporary
806        memory instead of the true size.)
807    
808    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
809        ovector size, it uses memory at the end of the block that it has got.
810    
811    29. Check for an overlong MARK name and give an error at compile time. The
812        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
813    
814    30. JIT compiler update.
815    
816    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
817        Rill for the patch.
818    
819    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
820    
821    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
822    
823    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
824        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
825        was enabled.
826    
827    35. Improve JIT code generation for greedy plus quantifier.
828    
829    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
830        1 to "aa" instead of to an empty string. The bug affected repeated groups
831        that could potentially match an empty string.
832    
833    37. Optimizing single character iterators in JIT.
834    
835    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
836        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
837        codepoints that are too big for the mode are faulted, and in a UTF mode,
838        disallowed codepoints are also faulted.
839    
840    39. If PCRE was compiled with UTF support, in three places in the DFA
841        matcher there was code that should only have been obeyed in UTF mode, but
842        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
843        processing when bytes with values greater than 127 were present. In 16-bit
844        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
845        both cases the values are those that cannot be the first data item in a UTF
846        character. The three items that might have provoked this were recursions,
847        possessively repeated groups, and atomic groups.
848    
849    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
850        and pcregrep, because some OS require shared objects to be explicitly
851        passed to ld, causing the link step to fail if they are not.
852    
853    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
854        mode, patterns that started with \h* or \R* might be incorrectly matched.
855    
856    
857    Version 8.30 04-February-2012
858    -----------------------------
859    
860    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
861        name is defined in ctype.h.
862    
863    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
864        only in quite long subpatterns.
865    
866    3.  Removed the function pcre_info(), which has been obsolete and deprecated
867        since it was replaced by pcre_fullinfo() in February 2000.
868    
869    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
870        match a (*MARK), and the match failed at the start of the subject, a
871        reference to memory before the start of the subject could occur. This bug
872        was introduced by fix 17 of release 8.21.
873    
874    5.  A reference to an unset group with zero minimum repetition was giving
875        totally wrong answers (in non-JavaScript-compatibility mode). For example,
876        /(another)?(\1?)test/ matched against "hello world test". This bug was
877        introduced in release 8.13.
878    
879    6.  Add support for 16-bit character strings (a large amount of work involving
880        many changes and refactorings).
881    
882    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
883        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
884        from a file.
885    
886    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
887        rounding is not applied in this particular case).
888    
889    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
890        if they appear, or are escaped, in patterns.
891    
892    10. Get rid of a number of -Wunused-but-set-variable warnings.
893    
894    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
895        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
896        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
897        also returns the mark "x". This bug applied to capturing parentheses,
898        non-capturing parentheses, and atomic parentheses. It also applied to some
899        assertions.
900    
901    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
902        information out of configure.ac instead of relying on pcre.h.generic, which
903        is not stored in the repository.
904    
905    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
906        -lreadline.
907    
908    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
909    
910    15. Applied Graycode's patch to put the top-level frame on the stack rather
911        than the heap when not using the stack for recursion. This gives a
912        performance improvement in many cases when recursion is not deep.
913    
914    16. Experimental code added to "pcretest -C" to output the stack frame size.
915    
916    
917    Version 8.21 12-Dec-2011
918    ------------------------
919    
920    1.  Updating the JIT compiler.
921    
922    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
923        are added as well.
924    
925    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
926        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
927        calling _pcre_jit_exec. Some extra comments are added.
928    
929    4.  (*MARK) settings inside atomic groups that do not contain any capturing
930        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
931        was introduced by change 18 for 8.20.
932    
933    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
934        ECMA-262 standard.
935    
936    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
937        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
938        This bug was probably introduced by change 9 of 8.13.
939    
940    7.  While fixing 6 above, I noticed that a number of other items were being
941        incorrectly rejected as "not fixed length". This arose partly because newer
942        opcodes had not been added to the fixed-length checking code. I have (a)
943        corrected the bug and added tests for these items, and (b) arranged for an
944        error to occur if an unknown opcode is encountered while checking for fixed
945        length instead of just assuming "not fixed length". The items that were
946        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
947        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
948        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
949    
950    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
951        being incorrectly compiled and would have given unpredicatble results.
952    
953    9.  A possessively repeated subpattern with minimum repeat count greater than
954        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
955        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
956        the first (A) could occur when it should not.
957    
958    10. Add a cast and remove a redundant test from the code.
959    
960    11. JIT should use pcre_malloc/pcre_free for allocation.
961    
962    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
963        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
964        is anything other than /usr, -L is still shown).
965    
966    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
967    
968    14. Perl does not support \N without a following name in a [] class; PCRE now
969        also gives an error.
970    
971    15. If a forward reference was repeated with an upper limit of around 2000,
972        it caused the error "internal error: overran compiling workspace". The
973        maximum number of forward references (including repeats) was limited by the
974        internal workspace, and dependent on the LINK_SIZE. The code has been
975        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
976        the default depends on LINK_SIZE. There is a new upper limit (for safety)
977        of around 200,000 forward references. While doing this, I also speeded up
978        the filling in of repeated forward references.
979    
980    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
981        incorrectly expecting the subject to contain another "a" after the start.
982    
983    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
984        in the match, the SKIP should be ignored. This was not happening; instead
985        the SKIP was being treated as NOMATCH. For patterns such as
986        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
987        tested.
988    
989    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
990        now much more compatible with Perl, in particular in cases where the result
991        is a non-match for a non-anchored pattern. For example, if
992        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
993        "m", where previously it did not return a name. A side effect of this
994        change is that for partial matches, the last encountered mark name is
995        returned, as for non matches. A number of tests that were previously not
996        Perl-compatible have been moved into the Perl-compatible test files. The
997        refactoring has had the pleasing side effect of removing one argument from
998        the match() function, thus reducing its stack requirements.
999    
1000    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1001        subsequent uses of /S (without +) incorrectly behaved like /S+.
1002    
1003    21. Retrieve executable code size support for the JIT compiler and fixing
1004        some warnings.
1005    
1006    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1007        not work when the shorter character appeared right at the end of the
1008        subject string.
1009    
1010    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1011        systems.
1012    
1013    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1014        output it when the /M option is used in pcretest.
1015    
1016    25. The CheckMan script was not being included in the distribution. Also, added
1017        an explicit "perl" to run Perl scripts from the PrepareRelease script
1018        because this is reportedly needed in Windows.
1019    
1020    26. If study data was being save in a file and studying had not found a set of
1021        "starts with" bytes for the pattern, the data written to the file (though
1022        never used) was taken from uninitialized memory and so caused valgrind to
1023        complain.
1024    
1025    27. Updated RunTest.bat as provided by Sheri Pierce.
1026    
1027    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1028    
1029    29. Computation of memory usage for the table of capturing group names was
1030        giving an unnecessarily large value.
1031    
1032    
1033    Version 8.20 21-Oct-2011
1034    ------------------------
1035    
1036    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1037        a POSIX class. After further experiments with Perl, which convinced me that
1038        Perl has bugs and confusions, a closing square bracket is no longer allowed
1039        in a POSIX name. This bug also affected patterns with classes that started
1040        with full stops.
1041    
1042    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1043        captured substring, but while checking the failing first alternative,
1044        substring 1 is temporarily captured. If the output vector supplied to
1045        pcre_exec() was not big enough for this capture, the yield of the function
1046        was still zero ("insufficient space for captured substrings"). This cannot
1047        be totally fixed without adding another stack variable, which seems a lot
1048        of expense for a edge case. However, I have improved the situation in cases
1049        such as /(a)(b)x|abc/ matched against "abc", where the return code
1050        indicates that fewer than the maximum number of slots in the ovector have
1051        been set.
1052    
1053    3.  Related to (2) above: when there are more back references in a pattern than
1054        slots in the output vector, pcre_exec() uses temporary memory during
1055        matching, and copies in the captures as far as possible afterwards. It was
1056        using the entire output vector, but this conflicts with the specification
1057        that only 2/3 is used for passing back captured substrings. Now it uses
1058        only the first 2/3, for compatibility. This is, of course, another edge
1059        case.
1060    
1061    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1062        main code base, and can be used by building with --enable-jit. When this is
1063        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1064        runtime --no-jit option is given.
1065    
1066    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1067        ovector, the return from the function was zero, implying that there were
1068        other matches that did not fit. The correct "exactly full" value is now
1069        returned.
1070    
1071    6.  If a subpattern that was called recursively or as a subroutine contained
1072        (*PRUNE) or any other control that caused it to give a non-standard return,
1073        invalid errors such as "Error -26 (nested recursion at the same subject
1074        position)" or even infinite loops could occur.
1075    
1076    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1077        computing the minimum length on reaching *ACCEPT, and so ended up with the
1078        wrong value of 1 rather than 0. Further investigation indicates that
1079        computing a minimum subject length in the presence of *ACCEPT is difficult
1080        (think back references, subroutine calls), and so I have changed the code
1081        so that no minimum is registered for a pattern that contains *ACCEPT.
1082    
1083    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1084        it was not handled as intended. [But see 16 below.]
1085    
1086    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1087        Sheri Pierce.
1088    
1089    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1090        the first byte in a match must be "a".
1091    
1092    11. Change 17 for 8.13 increased the recursion depth for patterns like
1093        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1094        pattern contains any instances of (*THEN). If it does not, the old
1095        optimizations are restored. It would be nice to do this on a per-group
1096        basis, but at the moment that is not feasible.
1097    
1098    12. In some environments, the output of pcretest -C is CRLF terminated. This
1099        broke RunTest's code that checks for the link size. A single white space
1100        character after the value is now allowed for.
1101    
1102    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1103        For "fr", it uses the Windows-specific input and output files.
1104    
1105    14. If (*THEN) appeared in a group that was called recursively or as a
1106        subroutine, it did not work as intended. [But see next item.]
1107    
1108    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1109        pattern fragments (but not containing any | characters). If A and B are
1110        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1111        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1112        D. In other words, Perl considers parentheses that do not contain any |
1113        characters to be part of a surrounding alternative, whereas PCRE was
1114        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1115        differently. PCRE now behaves in the same way as Perl, except in the case
1116        of subroutine/recursion calls such as (?1) which have in any case always
1117        been different (but PCRE had them first :-).
1118    
1119    16. Related to 15 above: Perl does not treat the | in a conditional group as
1120        creating alternatives. Such a group is treated in the same way as an
1121        ordinary group without any | characters when processing (*THEN). PCRE has
1122        been changed to match Perl's behaviour.
1123    
1124    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1125        RunGrepTest script failed.
1126    
1127    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1128        inevitable for groups that contain captures, but it can lead to a lot of
1129        stack use in large patterns. The old behaviour has been restored for atomic
1130        groups that do not contain any capturing parentheses.
1131    
1132    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1133        suppress the check for a minimum subject length at run time. (If it was
1134        given to pcre_exec() or pcre_dfa_exec() it did work.)
1135    
1136    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1137        fail to work when decoding hex characters in data strings in EBCDIC
1138        environments.
1139    
1140    21. It appears that in at least one Mac OS environment, the isxdigit() function
1141        is implemented as a macro that evaluates to its argument more than once,
1142        contravening the C 90 Standard (I haven't checked a later standard). There
1143        was an instance in pcretest which caused it to go wrong when processing
1144        \x{...} escapes in subject strings. The has been rewritten to avoid using
1145        things like p++ in the argument of isxdigit().
1146    
1147    
1148    Version 8.13 16-Aug-2011
1149    ------------------------
1150    
1151    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1152    
1153    2.  Two minor typos in pcre_internal.h have been fixed.
1154    
1155    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1156        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1157        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1158    
1159    4.  There were a number of related bugs in the code for matching backrefences
1160        caselessly in UTF-8 mode when codes for the characters concerned were
1161        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1162        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1163        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1164        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1165        2-byte code at the end of the subject (it thought there wasn't enough data
1166        left).
1167    
1168    5.  Comprehensive information about what went wrong is now returned by
1169        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1170        as the output vector has at least 2 elements. The offset of the start of
1171        the failing character and a reason code are placed in the vector.
1172    
1173    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1174        now returned is for the first byte of the failing character, instead of the
1175        last byte inspected. This is an incompatible change, but I hope it is small
1176        enough not to be a problem. It makes the returned offset consistent with
1177        pcre_exec() and pcre_dfa_exec().
1178    
1179    7.  pcretest now gives a text phrase as well as the error number when
1180        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1181        failure, the offset and reason code are output.
1182    
1183    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1184        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1185        back over a single character (\n). This seems wrong (because it treated the
1186        two characters as a single entity when going forwards), conflicts with the
1187        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1188        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1189        has been changed.
1190    
1191    9.  Some internal refactoring has changed the processing so that the handling
1192        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1193        time (the PCRE_DOTALL option was changed this way some time ago: version
1194        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1195        which was always a bit of a fudge. It also means that there is one less
1196        argument for the match() function, which reduces its stack requirements
1197        slightly. This change also fixes an incompatibility with Perl: the pattern
1198        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1199    
1200    10. More internal refactoring has drastically reduced the number of recursive
1201        calls to match() for possessively repeated groups such as (abc)++ when
1202        using pcre_exec().
1203    
1204    11. While implementing 10, a number of bugs in the handling of groups were
1205        discovered and fixed:
1206    
1207        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1208        (a|)*(?1) gave a compile-time internal error.
1209        ((a|)+)+  did not notice that the outer group could match an empty string.
1210        (^a|^)+   was not marked as anchored.
1211        (.*a|.*)+ was not marked as matching at start or after a newline.
1212    
1213    12. Yet more internal refactoring has removed another argument from the match()
1214        function. Special calls to this function are now indicated by setting a
1215        value in a variable in the "match data" data block.
1216    
1217    13. Be more explicit in pcre_study() instead of relying on "default" for
1218        opcodes that mean there is no starting character; this means that when new
1219        ones are added and accidentally left out of pcre_study(), testing should
1220        pick them up.
1221    
1222    14. The -s option of pcretest has been documented for ages as being an old
1223        synonym of -m (show memory usage). I have changed it to mean "force study
1224        for every regex", that is, assume /S for every regex. This is similar to -i
1225        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1226        using it. It makes it easier to run collections of tests with and without
1227        study enabled, and thereby test pcre_study() more easily. All the standard
1228        tests are now run with and without -s (but some patterns can be marked as
1229        "never study" - see 20 below).
1230    
1231    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1232        restoration of the capturing data to the outer values was not happening
1233        correctly.
1234    
1235    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1236        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1237        pattern had matched an empty string, and so incorrectly returned a no
1238        match.
1239    
1240    17. There was optimizing code for the last branch of non-capturing parentheses,
1241        and also for the obeyed branch of a conditional subexpression, which used
1242        tail recursion to cut down on stack usage. Unfortunately, now that there is
1243        the possibility of (*THEN) occurring in these branches, tail recursion is
1244        no longer possible because the return has to be checked for (*THEN). These
1245        two optimizations have therefore been removed. [But see 8.20/11 above.]
1246    
1247    18. If a pattern containing \R was studied, it was assumed that \R always
1248        matched two bytes, thus causing the minimum subject length to be
1249        incorrectly computed because \R can also match just one byte.
1250    
1251    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1252        was incorrectly computed.
1253    
1254    20. If /S is present twice on a test pattern in pcretest input, it now
1255        *disables* studying, thereby overriding the use of -s on the command line
1256        (see 14 above). This is necessary for one or two tests to keep the output
1257        identical in both cases.
1258    
1259    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1260        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1261    
1262    22. When an atomic group that contained a capturing parenthesis was
1263        successfully matched, but the branch in which it appeared failed, the
1264        capturing was not being forgotten if a higher numbered group was later
1265        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1266        group 1 to "a", when in fact it should be unset. This applied to multi-
1267        branched capturing and non-capturing groups, repeated or not, and also to
1268        positive assertions (capturing in negative assertions does not happen
1269        in PCRE) and also to nested atomic groups.
1270    
1271    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1272        subject after a captured substring, to make it easier to tell which of a
1273        number of identical substrings has been captured.
1274    
1275    24. The way atomic groups are processed by pcre_exec() has been changed so that
1276        if they are repeated, backtracking one repetition now resets captured
1277        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1278        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1279        "aaa". Previously, it would have been "a". As part of this code
1280        refactoring, the way recursive calls are handled has also been changed.
1281    
1282    25. If an assertion condition captured any substrings, they were not passed
1283        back unless some other capturing happened later. For example, if
1284        (?(?=(a))a) was matched against "a", no capturing was returned.
1285    
1286    26. When studying a pattern that contained subroutine calls or assertions,
1287        the code for finding the minimum length of a possible match was handling
1288        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1289        group 1 called group 2 while simultaneously a separate group 2 called group
1290        1). A stack overflow occurred in this case. I have fixed this by limiting
1291        the recursion depth to 10.
1292    
1293    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1294        Fortmann. This supports explicit test numbers on the command line, and has
1295        argument validation and error reporting.
1296    
1297    28. An instance of \X with an unlimited repeat could fail if at any point the
1298        first character it looked at was a mark character.
1299    
1300    29. Some minor code refactoring concerning Unicode properties and scripts
1301        should reduce the stack requirement of match() slightly.
1302    
1303    30. Added the '=' option to pcretest to check the setting of unused capturing
1304        slots at the end of the pattern, which are documented as being -1, but are
1305        not included in the return count.
1306    
1307    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1308        compiled something random. Now it gives a compile-time error (as does
1309        Perl).
1310    
1311    32. A *MARK encountered during the processing of a positive assertion is now
1312        recorded and passed back (compatible with Perl).
1313    
1314    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1315        had alternative anchored branches, the search for a second match in a line
1316        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1317        matched the line "0102" twice. The same bug affected patterns that started
1318        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1319        twice.
1320    
1321    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1322        does, and because of capturing effects, quantifying parenthesized
1323        assertions may at times be useful. Quantifiers are now allowed for
1324        parenthesized assertions.
1325    
1326    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1327    
1328    36. \g was being checked for fancy things in a character class, when it should
1329        just be a literal "g".
1330    
1331    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1332        appearance of a nested POSIX class supersedes an apparent external class.
1333        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1334        unescaped square brackets may also appear as part of class names. For
1335        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1336        more like Perl. (But see 8.20/1 above.)
1337    
1338    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1339        was because it thought it was \N{name}, which is not supported).
1340    
1341    39. Add minix to OS list not supporting the -S option in pcretest.
1342    
1343    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1344        cannot analyze patterns in sufficient detail to catch mutual recursions
1345        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1346        subgroup is called recursively as a subpattern for a second time at the
1347        same position in the subject string. In previous releases this might have
1348        been caught by the recursion limit, or it might have run out of stack.
1349    
1350    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1351        happen only once. PCRE was, however incorrectly giving a compile time error
1352        "recursive call could loop indefinitely" because it cannot analyze the
1353        pattern in sufficient detail. The compile time test no longer happens when
1354        PCRE is compiling a conditional subpattern, but actual runaway loops are
1355        now caught at runtime (see 40 above).
1356    
1357    42. It seems that Perl allows any characters other than a closing parenthesis
1358        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1359        has been changed to be the same.
1360    
1361    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1362        as not to get warnings when autogen.sh is called. Also changed
1363        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1364    
1365    44. To help people who use pcregrep to scan files containing exceedingly long
1366        lines, the following changes have been made:
1367    
1368        (a) The default value of the buffer size parameter has been increased from
1369            8K to 20K. (The actual buffer used is three times this size.)
1370    
1371        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1372            PCRE is built.
1373    
1374        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1375            to be set at run time.
1376    
1377        (d) Numerical values in pcregrep options can be followed by K or M, for
1378            example --buffer-size=50K.
1379    
1380        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1381            given and the return code is set to 2.
1382    
1383    45. Add a pointer to the latest mark to the callout data block.
1384    
1385    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1386        partial match of an empty string instead of no match. This was specific to
1387        the use of ".".
1388    
1389    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1390        complete match instead of a partial match. This bug was dependent on both
1391        the PCRE_UTF8 and PCRE_DOTALL options being set.
1392    
1393    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1394        starting byte set, because \b was not being ignored.
1395    
1396    
1397    Version 8.12 15-Jan-2011
1398    ------------------------
1399    
1400    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1401        checks for such things as part of the documentation building process.
1402    
1403    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1404        --match-limit and --recursion-limit options (added for 8.11). In
1405        particular, this made one of the standard tests fail. (The integer value
1406        went into the wrong half of a long int.)
1407    
1408    3.  If the --colour option was given to pcregrep with -v (invert match), it
1409        did strange things, either producing crazy output, or crashing. It should,
1410        of course, ignore a request for colour when reporting lines that do not
1411        match.
1412    
1413    4.  Another pcregrep bug caused similar problems if --colour was specified with
1414        -M (multiline) and the pattern match finished with a line ending.
1415    
1416    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1417        matched in multiline mode, the following line was shown as part of the
1418        match. This seems wrong, so I have changed it.
1419    
1420    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1421        the check for further matches in the same line (so they could be coloured)
1422        to overrun the end of the current line. If another match was found, it was
1423        incorrectly shown (and then shown again when found in the next line).
1424    
1425    7.  If pcregrep was compiled under Windows, there was a reference to the
1426        function pcregrep_exit() before it was defined. I am assuming this was
1427        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1428        reported by a user. I've moved the definition above the reference.
1429    
1430    
1431    Version 8.11 10-Dec-2010
1432    ------------------------
1433    
1434    1.  (*THEN) was not working properly if there were untried alternatives prior
1435        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1436        backtracked to try for "b" instead of moving to the next alternative branch
1437        at the same level (in this case, to look for "c"). The Perl documentation
1438        is clear that when (*THEN) is backtracked onto, it goes to the "next
1439        alternative in the innermost enclosing group".
1440    
1441    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1442        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1443        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1444        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1445        (*THEN).
1446    
1447    3.  If \s appeared in a character class, it removed the VT character from
1448        the class, even if it had been included by some previous item, for example
1449        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1450        of \s, but is part of the POSIX "space" class.)
1451    
1452    4.  A partial match never returns an empty string (because you can always
1453        match an empty string at the end of the subject); however the checking for
1454        an empty string was starting at the "start of match" point. This has been
1455        changed to the "earliest inspected character" point, because the returned
1456        data for a partial match starts at this character. This means that, for
1457        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1458        (previously it gave "no match").
1459    
1460    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1461        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1462        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1463        has an implication that the given string is incomplete (because a partial
1464        match is preferred over a full match). For this reason, these items now
1465        give a partial match in this situation. [Aside: previously, the one case
1466        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1467        match rather than a full match, which was wrong by the old rules, but is
1468        now correct.]
1469    
1470    6.  There was a bug in the handling of #-introduced comments, recognized when
1471        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1472        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1473        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1474        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1475        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1476        places in pcre_compile().
1477    
1478    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1479        comments when looking ahead for named forward references to subpatterns,
1480        the only newline sequence it recognized was NL. It now handles newlines
1481        according to the set newline convention.
1482    
1483    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1484        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1485        cater for a lack of strerror(). These oversights have been fixed.
1486    
1487    9.  Added --match-limit and --recursion-limit to pcregrep.
1488    
1489    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1490    
1491    11. When the -o option was used, pcregrep was setting a return code of 1, even
1492        when matches were found, and --line-buffered was not being honoured.
1493    
1494    12. Added an optional parentheses number to the -o and --only-matching options
1495        of pcregrep.
1496    
1497    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1498        can match an empty string. The code to do it in pcretest and pcredemo
1499        needed fixing:
1500    
1501        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1502            only one byte after an empty string match just before CRLF (this case
1503            just got forgotten; "any" and "anycrlf" were OK).
1504    
1505        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1506            mode when an empty string match preceded an ASCII character followed by
1507            a non-ASCII character. (The code for advancing by one character rather
1508            than one byte was nonsense.)
1509    
1510        (c) The pcredemo.c sample program did not have any code at all to handle
1511            the cases when CRLF is a valid newline sequence.
1512    
1513    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1514        as a starting offset was within the subject string. There is now a new
1515        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1516        negative or greater than the length of the string. In order to test this,
1517        pcretest is extended to allow the setting of negative starting offsets.
1518    
1519    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1520        starting offset points to the beginning of a UTF-8 character was
1521        unnecessarily clumsy. I tidied it up.
1522    
1523    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1524        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1525    
1526    17. Nobody had reported that the --include_dir option, which was added in
1527        release 7.7 should have been called --include-dir (hyphen, not underscore)
1528        for compatibility with GNU grep. I have changed it to --include-dir, but
1529        left --include_dir as an undocumented synonym, and the same for
1530        --exclude-dir, though that is not available in GNU grep, at least as of
1531        release 2.5.4.
1532    
1533    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1534        characters from a string of bytes) have been redefined so as not to use
1535        loops, in order to improve performance in some environments. At the same
1536        time, I abstracted some of the common code into auxiliary macros to save
1537        repetition (this should not affect the compiled code).
1538    
1539    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1540        compile-time error is now given if \c is not followed by an ASCII
1541        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1542        different, and any byte value is allowed.)
1543    
1544    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1545        START_OPTIMIZE option, which is now allowed at compile time - but just
1546        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1547        to pcregrep and other applications that have no direct access to PCRE
1548        options. The new /Y option in pcretest sets this option when calling
1549        pcre_compile().
1550    
1551    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1552        back references. Groups containing recursive back references were forced to
1553        be atomic by that change, but in the case of named groups, the amount of
1554        memory required was incorrectly computed, leading to "Failed: internal
1555        error: code overflow". This has been fixed.
1556    
1557    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1558        pcretest.c, to avoid build problems in some Borland environments.
1559    
1560    
1561    Version 8.10 25-Jun-2010
1562    ------------------------
1563    
1564    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1565        THEN.
1566    
1567    2.  (*ACCEPT) was not working when inside an atomic group.
1568    
1569    3.  Inside a character class, \B is treated as a literal by default, but
1570        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1571        causes the error). The code is unchanged, but I tidied the documentation.
1572    
1573    4.  Inside a character class, PCRE always treated \R and \X as literals,
1574        whereas Perl faults them if its -w option is set. I have changed PCRE so
1575        that it faults them when PCRE_EXTRA is set.
1576    
1577    5.  Added support for \N, which always matches any character other than
1578        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1579    
1580    6.  When compiling pcregrep with newer versions of gcc which may have
1581        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1582        declared with attribute warn_unused_result" were given. Just casting the
1583        result to (void) does not stop the warnings; a more elaborate fudge is
1584        needed. I've used a macro to implement this.
1585    
1586    7.  Minor change to pcretest.c to avoid a compiler warning.
1587    
1588    8.  Added four artifical Unicode properties to help with an option to make
1589        \s etc use properties (see next item). The new properties are: Xan
1590        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1591    
1592    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1593        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1594        this option. Modified pcretest to add /W to test this facility. Added
1595        REG_UCP to make it available via the POSIX interface.
1596    
1597    10. Added --line-buffered to pcregrep.
1598    
1599    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1600        studied, and the match started with a letter with a code point greater than
1601        127 whose first byte was different to the first byte of the other case of
1602        the letter, the other case of this starting letter was not recognized
1603        (#976).
1604    
1605    12. If a pattern that was studied started with a repeated Unicode property
1606        test, for example, \p{Nd}+, there was the theoretical possibility of
1607        setting up an incorrect bitmap of starting bytes, but fortunately it could
1608        not have actually happened in practice until change 8 above was made (it
1609        added property types that matched character-matching opcodes).
1610    
1611    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1612        possible starting bytes for non-anchored patterns.
1613    
1614    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1615        \R, and also a number of cases that involve Unicode properties, both
1616        explicit and implicit when PCRE_UCP is set.
1617    
1618    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1619        input, it could crash or give wrong results if characters with values
1620        greater than 0xc0 were present in the subject string. (Detail: it assumed
1621        UTF-8 input when processing these items.)
1622    
1623    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1624        size_t is 64-bit (#991).
1625    
1626    17. Added a check for running out of memory when PCRE is compiled with
1627        --disable-stack-for-recursion (#990).
1628    
1629    18. If the last data line in a file for pcretest does not have a newline on
1630        the end, a newline was missing in the output.
1631    
1632    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1633        less than 128) in its various bitmaps. However, there is a facility for
1634        generating tables according to the current locale when PCRE is compiled. It
1635        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1636        characters, are recognized by isspace() and therefore were getting set in
1637        these tables, and indeed these tables seem to approximate to ISO 8859. This
1638        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1639        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1640        which of course cannot start UTF-8 characters. I have changed the code so
1641        that only real ASCII characters (less than 128) and the correct starting
1642        bytes for UTF-8 encodings are set for characters greater than 127 when in
1643        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1644        altogether.)
1645    
1646    20. Added the /T option to pcretest so as to be able to run tests with non-
1647        standard character tables, thus making it possible to include the tests
1648        used for 19 above in the standard set of tests.
1649    
1650    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1651        reference to a subpattern the other side of a comment that contains an
1652        opening parenthesis caused either an internal compiling error, or a
1653        reference to the wrong subpattern.
1654    
1655    
1656    Version 8.02 19-Mar-2010
1657    ------------------------
1658    
1659    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1660    
1661    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1662        configured.
1663    
1664    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1665        original author of that file, following a query about its status.
1666    
1667    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1668        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1669    
1670    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1671        quantifier applied to a forward-referencing subroutine call, could compile
1672        incorrect code or give the error "internal error: previously-checked
1673        referenced subpattern not found".
1674    
1675    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1676        variables to point to external functions. For these systems, therefore,
1677        pcre_malloc etc. are now initialized to local functions that call the
1678        relevant global functions.
1679    
1680    7.  There were two entries missing in the vectors called coptable and poptable
1681        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1682        I've fixed the data, and added a kludgy way of testing at compile time that
1683        the lengths are correct (equal to the number of opcodes).
1684    
1685    8.  Following on from 7, I added a similar kludge to check the length of the
1686        eint vector in pcreposix.c.
1687    
1688    9.  Error texts for pcre_compile() are held as one long string to avoid too
1689        much relocation at load time. To find a text, the string is searched,
1690        counting zeros. There was no check for running off the end of the string,
1691        which could happen if a new error number was added without updating the
1692        string.
1693    
1694    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1695    
1696    11. \K was not working if it appeared in an atomic group or in a group that
1697        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1698        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1699        the assertion is positive, but not if it is negative.
1700    
1701    12. Change 11 fortuitously reduced the size of the stack frame used in the
1702        "match()" function of pcre_exec.c by one pointer. Forthcoming
1703        implementation of support for (*MARK) will need an extra pointer on the
1704        stack; I have reserved it now, so that the stack frame size does not
1705        decrease.
1706    
1707    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1708        item in branch that calls a recursion is a subroutine call - as in the
1709        second branch in the above example - was incorrectly given the compile-
1710        time error "recursive call could loop indefinitely" because pcre_compile()
1711        was not correctly checking the subroutine for matching a non-empty string.
1712    
1713    14. The checks for overrunning compiling workspace could trigger after an
1714        overrun had occurred. This is a "should never occur" error, but it can be
1715        triggered by pathological patterns such as hundreds of nested parentheses.
1716        The checks now trigger 100 bytes before the end of the workspace.
1717    
1718    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1719    
1720    
1721    Version 8.01 19-Jan-2010
1722    ------------------------
1723    
1724    1.  If a pattern contained a conditional subpattern with only one branch (in
1725        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1726        computed the wrong minimum data length (which is of course zero for such
1727        subpatterns). This could cause incorrect "no match" results.
1728    
1729    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1730        the pattern is reset in the first branch, pcre_compile() failed with
1731        "internal error: code overflow at offset...". This happened only when
1732        the reset was to the original external option setting. (An optimization
1733        abstracts leading options settings into an external setting, which was the
1734        cause of this.)
1735    
1736    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1737        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1738        assertion pattern did not match (meaning that the assertion was true), it
1739        was incorrectly treated as false if the SKIP had been reached during the
1740        matching. This also applied to assertions used as conditions.
1741    
1742    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1743        assertion subpattern, including such a pattern used as a condition,
1744        unpredictable results occurred, instead of the error return
1745        PCRE_ERROR_DFA_UITEM.
1746    
1747    5.  The C++ GlobalReplace function was not working like Perl for the special
1748        situation when an empty string is matched. It now does the fancy magic
1749        stuff that is necessary.
1750    
1751    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1752        removed. (These were left over from very, very early versions of PCRE.)
1753    
1754    7.  Some cosmetic changes to the code to make life easier when compiling it
1755        as part of something else:
1756    
1757        (a) Change DEBUG to PCRE_DEBUG.
1758    
1759        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1760            called "current" as "current_branch", to prevent a collision with the
1761            Linux macro when compiled as a kernel module.
1762    
1763        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1764            prevent a collision with the Linux macro when compiled as a kernel
1765            module.
1766    
1767    8.  In pcre_compile() there are some checks for integer overflows that used to
1768        cast potentially large values to (double). This has been changed to that
1769        when building, a check for int64_t is made, and if it is found, it is used
1770        instead, thus avoiding the use of floating point arithmetic. (There is no
1771        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1772        double.
1773    
1774    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1775        2005 (difference between two addresses compared to an unsigned value).
1776    
1777    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1778        custom one, because of the following reported problem in Windows:
1779    
1780          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1781              under Win32.
1782          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1783              therefore missing the function definition.
1784          - The compiler thus generates a "C" signature for the test function.
1785          - The linker fails to find the "C" function.
1786          - PCRE fails to configure if asked to do so against libbz2.
1787    
1788    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1789        messages were output:
1790    
1791          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1792          rerunning libtoolize, to keep the correct libtool macros in-tree.
1793          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1794    
1795        I have done both of these things.
1796    
1797    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1798        most of the time, it *can* run out if it is given a pattern that contains a
1799        runaway infinite recursion. I updated the discussion in the pcrestack man
1800        page.
1801    
1802    13. Now that we have gone to the x.xx style of version numbers, the minor
1803        version may start with zero. Using 08 or 09 is a bad idea because users
1804        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1805        interpreted as invalid octal numbers. I've updated the previous comment in
1806        configure.ac, and also added a check that gives an error if 08 or 09 are
1807        used.
1808    
1809    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1810        causing partial matching to fail when the end of the subject matched \W
1811        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1812    
1813    15. There were some discrepancies between the declarations in pcre_internal.h
1814        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1815        their definitions. The declarations used "const uschar *" and the
1816        definitions used USPTR. Even though USPTR is normally defined as "const
1817        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1818        reported that: "This difference in casting confuses some C++ compilers, for
1819        example, SunCC recognizes above declarations as different functions and
1820        generates broken code for hbpcre." I have changed the declarations to use
1821        USPTR.
1822    
1823    16. GNU libtool is named differently on some systems. The autogen.sh script now
1824        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1825        (FreeBSD).
1826    
1827    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1828        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1829        comment: "Figure out how to create a longlong from a string: strtoll and
1830        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1831        instance, but it only takes 2 args instead of 3!"
1832    
1833    18. A subtle bug concerned with back references has been fixed by a change of
1834        specification, with a corresponding code fix. A pattern such as
1835        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1836        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1837        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1838        same bug. Such groups have to be quantified to be useful, or contained
1839        inside another quantified group. (If there's no repetition, the reference
1840        can never match.) The problem arises because, having left the group and
1841        moved on to the rest of the pattern, a later failure that backtracks into
1842        the group uses the captured value from the final iteration of the group
1843        rather than the correct earlier one. I have fixed this in PCRE by forcing
1844        any group that contains a reference to itself to be an atomic group; that
1845        is, there cannot be any backtracking into it once it has completed. This is
1846        similar to recursive and subroutine calls.
1847    
1848    
1849    Version 8.00 19-Oct-09
1850    ----------------------
1851    
1852    1.  The table for translating pcre_compile() error codes into POSIX error codes
1853        was out-of-date, and there was no check on the pcre_compile() error code
1854        being within the table. This could lead to an OK return being given in
1855        error.
1856    
1857    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1858        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1859        in a Windows environment.
1860    
1861    3.  The pcregrep --count option prints the count for each file even when it is
1862        zero, as does GNU grep. However, pcregrep was also printing all files when
1863        --files-with-matches was added. Now, when both options are given, it prints
1864        counts only for those files that have at least one match. (GNU grep just
1865        prints the file name in this circumstance, but including the count seems
1866        more useful - otherwise, why use --count?) Also ensured that the
1867        combination -clh just lists non-zero counts, with no names.
1868    
1869    4.  The long form of the pcregrep -F option was incorrectly implemented as
1870        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1871        but it seems right to fix it, and I didn't think it was worth preserving
1872        the old behaviour.
1873    
1874    5.  The command line items --regex=pattern and --regexp=pattern were not
1875        recognized by pcregrep, which required --regex pattern or --regexp pattern
1876        (with a space rather than an '='). The man page documented the '=' forms,
1877        which are compatible with GNU grep; these now work.
1878    
1879    6.  No libpcreposix.pc file was created for pkg-config; there was just
1880        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1881    
1882    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1883        when UCP support is not needed, by modifying the Python script that
1884        generates it from Unicode data files. This should not matter if the module
1885        is correctly used as a library, but I received one complaint about 50K of
1886        unwanted data. My guess is that the person linked everything into his
1887        program rather than using a library. Anyway, it does no harm.
1888    
1889    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1890        was a minimum greater than 1 for a wide character in a possessive
1891        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1892        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1893        character. Chaos in the form of incorrect output or a compiling loop could
1894        result.
1895    
1896    9.  The restrictions on what a pattern can contain when partial matching is
1897        requested for pcre_exec() have been removed. All patterns can now be
1898        partially matched by this function. In addition, if there are at least two
1899        slots in the offset vector, the offset of the earliest inspected character
1900        for the match and the offset of the end of the subject are set in them when
1901        PCRE_ERROR_PARTIAL is returned.
1902    
1903    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1904        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1905        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1906        and may be more useful for multi-segment matching.
1907    
1908    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1909        used to be given if ever the end of the subject was reached; now it is
1910        given only if matching could not proceed because another character was
1911        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1912        string "Z", which now yields "no match" instead of "partial match". In the
1913        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1914        final character ended with (*FAIL).
1915    
1916    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1917        if the pattern had a "must contain" character that was already found in the
1918        earlier partial match, unless partial matching was again requested. For
1919        example, with the pattern /dog.(body)?/, the "must contain" character is
1920        "g". If the first part-match was for the string "dog", restarting with
1921        "sbody" failed. This bug has been fixed.
1922    
1923    13. The string returned by pcre_dfa_exec() after a partial match has been
1924        changed so that it starts at the first inspected character rather than the
1925        first character of the match. This makes a difference only if the pattern
1926        starts with a lookbehind assertion or \b or \B (\K is not supported by
1927        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1928        matching functions compatible, and I think it's the right thing to do.
1929    
1930    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1931        so that the demonstration program is easily available in environments where
1932        PCRE has not been installed from source.
1933    
1934    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1935        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1936        library.
1937    
1938    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1939        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1940        is not the first non-POSIX option to be added. Clearly some people find
1941        these options useful.
1942    
1943    17. If a caller to the POSIX matching function regexec() passes a non-zero
1944        value for nmatch with a NULL value for pmatch, the value of
1945        nmatch is forced to zero.
1946    
1947    18. RunGrepTest did not have a test for the availability of the -u option of
1948        the diff command, as RunTest does. It now checks in the same way as
1949        RunTest, and also checks for the -b option.
1950    
1951    19. If an odd number of negated classes containing just a single character
1952        interposed, within parentheses, between a forward reference to a named
1953        subpattern and the definition of the subpattern, compilation crashed with
1954        an internal error, complaining that it could not find the referenced
1955        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1956        [The bug was that it was starting one character too far in when skipping
1957        over the character class, thus treating the ] as data rather than
1958        terminating the class. This meant it could skip too much.]
1959    
1960    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1961        /g option in pcretest when the pattern contains \K, which makes it possible
1962        to have an empty string match not at the start, even when the pattern is
1963        anchored. Updated pcretest and pcredemo to use this option.
1964    
1965    21. If the maximum number of capturing subpatterns in a recursion was greater
1966        than the maximum at the outer level, the higher number was returned, but
1967        with unset values at the outer level. The correct (outer level) value is
1968        now given.
1969    
1970    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1971        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1972        make it do so. The string so far is captured, making this feature
1973        compatible with Perl.
1974    
1975    23. The tests have been re-organized, adding tests 11 and 12, to make it
1976        possible to check the Perl 5.10 features against Perl 5.10.
1977    
1978    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1979        pattern matches a fixed length string. PCRE did not allow this; now it
1980        does. Neither allows recursion.
1981    
1982    25. I finally figured out how to implement a request to provide the minimum
1983        length of subject string that was needed in order to match a given pattern.
1984        (It was back references and recursion that I had previously got hung up
1985        on.) This code has now been added to pcre_study(); it finds a lower bound
1986        to the length of subject needed. It is not necessarily the greatest lower
1987        bound, but using it to avoid searching strings that are too short does give
1988        some useful speed-ups. The value is available to calling programs via
1989        pcre_fullinfo().
1990    
1991    26. While implementing 25, I discovered to my embarrassment that pcretest had
1992        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1993        study optimizations had never been tested with that matching function.
1994        Oops. What is worse, even when it was passed study data, there was a bug in
1995        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1996        were also very few tests of studied patterns with pcre_dfa_exec().
1997    
1998    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1999        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2000        on the other side of the coin, they are no longer allowed to have different
2001        names, because these cannot be distinguished in PCRE, and this has caused
2002        confusion. (This is a difference from Perl.)
2003    
2004    28. When duplicate subpattern names are present (necessarily with different
2005        numbers, as required by 27 above), and a test is made by name in a
2006        conditional pattern, either for a subpattern having been matched, or for
2007        recursion in such a pattern, all the associated numbered subpatterns are
2008        tested, and the overall condition is true if the condition is true for any
2009        one of them. This is the way Perl works, and is also more like the way
2010        testing by number works.
2011    
2012    
2013    Version 7.9 11-Apr-09
2014    ---------------------
2015    
2016    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2017        (pcretest), all targets were linked against these libraries. This included
2018        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2019        libraries. This caused unwanted dependencies to be created. This problem
2020        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2021        pcretest is linked with readline.
2022    
2023    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2024        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2025        moved outside it again, because FALSE and TRUE are already defined in AIX,
2026        but BOOL is not.
2027    
2028    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2029        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2030    
2031    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2032        hyphens) following file names and line numbers when outputting matching
2033        lines. This is not true; no spaces are inserted. I have also clarified the
2034        wording for the --colour (or --color) option.
2035    
2036    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2037        was not coloured; this is different to GNU grep, so I have changed it to be
2038        the same.
2039    
2040    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2041        each matching line was coloured. Now it goes on to look for further matches
2042        of any of the test patterns, which is the same behaviour as GNU grep.
2043    
2044    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2045        doesn't make sense to accept an empty string match in pcregrep, so I have
2046        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2047        seems to be how GNU grep behaves. [But see later change 40 for release
2048        8.33.]
2049    
2050    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2051        start or after a newline", because the conditional assertion was not being
2052        correctly handled. The rule now is that both the assertion and what follows
2053        in the first alternative must satisfy the test.
2054    
2055    9.  If auto-callout was enabled in a pattern with a conditional group whose
2056        condition was an assertion, PCRE could crash during matching, both with
2057        pcre_exec() and pcre_dfa_exec().
2058    
2059    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2060        used for matching.
2061    
2062    11. Unicode property support in character classes was not working for
2063        characters (bytes) greater than 127 when not in UTF-8 mode.
2064    
2065    12. Added the -M command line option to pcretest.
2066    
2067    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2068    
2069    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2070    
2071    16. Added comments and documentation about mis-use of no_arg in the C++
2072        wrapper.
2073    
2074    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2075        from Martin Jerabek that uses macro names for all relevant character and
2076        string constants.
2077    
2078    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2079        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2080        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2081        these, but not everybody uses configure.
2082    
2083    19. A conditional group that had only one branch was not being correctly
2084        recognized as an item that could match an empty string. This meant that an
2085        enclosing group might also not be so recognized, causing infinite looping
2086        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2087        with the subject "ab", where knowledge that the repeated group can match
2088        nothing is needed in order to break the loop.
2089    
2090    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2091        exec(), but without supplying a callout function, matching went wrong.
2092    
2093    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2094        leak if the size of the offset vector was greater than 30. When the vector
2095        is smaller, the saved offsets during recursion go onto a local stack
2096        vector, but for larger vectors malloc() is used. It was failing to free
2097        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2098        error, in fact).
2099    
2100    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2101        heapframe that is used only when UTF-8 support is enabled. This caused no
2102        problem, but was untidy.
2103    
2104    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2105        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2106        included within another project.
2107    
2108    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2109        slightly modified by me:
2110    
2111          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2112              not building pcregrep.
2113    
2114          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2115              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2116    
2117    25. Forward references, both numeric and by name, in patterns that made use of
2118        duplicate group numbers, could behave incorrectly or give incorrect errors,
2119        because when scanning forward to find the reference group, PCRE was not
2120        taking into account the duplicate group numbers. A pattern such as
2121        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2122    
2123    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2124        the feature of a custom pointer more persuasive (as requested by a user).
2125    
2126    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2127        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2128        pre-defined.
2129    
2130    28. Added support for (*UTF8) at the start of a pattern.
2131    
2132    29. Arrange for flags added by the "release type" setting in CMake to be shown
2133        in the configuration summary.
2134    
2135    
2136    Version 7.8 05-Sep-08
2137    ---------------------
2138    
2139    1.  Replaced UCP searching code with optimized version as implemented for Ad
2140        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2141        stage table and inline lookup instead of a function, giving speed ups of 2
2142        to 5 times on some simple patterns that I tested. Permission was given to
2143        distribute the MultiStage2.py script that generates the tables (it's not in
2144        the tarball, but is in the Subversion repository).
2145    
2146    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2147        scripts.
2148    
2149    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2150        a group with a zero qualifier. The result of the study could be incorrect,
2151        or the function might crash, depending on the pattern.
2152    
2153    4.  Caseless matching was not working for non-ASCII characters in back
2154        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2155        It now works when Unicode Property Support is available.
2156    
2157    5.  In pcretest, an escape such as \x{de} in the data was always generating
2158        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2159        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2160        truncation.
2161    
2162    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2163    
2164    7.  Added two (int) casts to pcregrep when printing the difference of two
2165        pointers, in case they are 64-bit values.
2166    
2167    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2168        test 2 if it fails.
2169    
2170    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2171        and a #define of that name to empty if it is not externally set. This is to
2172        allow users of MSVC to set it if necessary.
2173    
2174    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2175        the convenience functions in the pcre_get.c source file.
2176    
2177    11. An option change at the start of a pattern that had top-level alternatives
2178        could cause overwriting and/or a crash. This command provoked a crash in
2179        some environments:
2180    
2181          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2182    
2183        This potential security problem was recorded as CVE-2008-2371.
2184    
2185    12. For a pattern where the match had to start at the beginning or immediately
2186        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2187        pcre_dfa_exec() could read past the end of the passed subject if there was
2188        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2189        pcretest so that it places the subject at the end of its malloc-ed buffer.
2190    
2191    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2192        exec() might read past the end of the data buffer in UTF-8 mode.
2193    
2194    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2195        the data contained the byte 0x85 as part of a UTF-8 character within its
2196        first line. This applied both to normal and DFA matching.
2197    
2198    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2199        /^[^d]*?$/8 failed to match "abc".
2200    
2201    16. Added a missing copyright notice to pcrecpp_internal.h.
2202    
2203    17. Make it more clear in the documentation that values returned from
2204        pcre_exec() in ovector are byte offsets, not character counts.
2205    
2206    18. Tidied a few places to stop certain compilers from issuing warnings.
2207    
2208    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2209        supplied by Stefan Weber. I made a further small update for 7.8 because
2210        there is a change of source arrangements: the pcre_searchfuncs.c module is
2211        replaced by pcre_ucd.c.
2212    
2213    
2214    Version 7.7 07-May-08
2215    ---------------------
2216    
2217    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2218        a string to a long long, pretend we don't even have a long long." This is
2219        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2220    
2221    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2222        pre-7.6 versions, which defined a global no_arg variable instead of putting
2223        it in the RE class. (See also #8 below.)
2224    
2225    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2226        Lopes.
2227    
2228    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2229    
2230        (1) The include/exclude patterns were being applied to the whole pathnames
2231            of files, instead of just to the final components.
2232    
2233        (2) If there was more than one level of directory, the subdirectories were
2234            skipped unless they satisfied the include/exclude conditions. This is
2235            inconsistent with GNU grep (and could even be seen as contrary to the
2236            pcregrep specification - which I improved to make it absolutely clear).
2237            The action now is always to scan all levels of directory, and just
2238            apply the include/exclude patterns to regular files.
2239    
2240    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2241        --exclude_dir in the tests to avoid scanning .svn directories.
2242    
2243    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2244        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2245        doesn't support NULs in patterns.
2246    
2247    7.  Added some missing "const"s to declarations of static tables in
2248        pcre_compile.c and pcre_dfa_exec.c.
2249    
2250    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2251        caused by fix #2  above. (Subsequently also a second patch to fix the
2252        first patch. And a third patch - this was a messy problem.)
2253    
2254    9.  Applied Craig's patch to remove the use of push_back().
2255    
2256    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2257        matching function regexec().
2258    
2259    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2260        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2261        references. PCRE supports relative numbers with this syntax (I don't think
2262        Oniguruma does).
2263    
2264    12. Previously, a group with a zero repeat such as (...){0} was completely
2265        omitted from the compiled regex. However, this means that if the group
2266        was called as a subroutine from elsewhere in the pattern, things went wrong
2267        (an internal error was given). Such groups are now left in the compiled
2268        pattern, with a new opcode that causes them to be skipped at execution
2269        time.
2270    
2271    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2272        to the way PCRE behaves:
2273    
2274        (a) A lone ] character is dis-allowed (Perl treats it as data).
2275    
2276        (b) A back reference to an unmatched subpattern matches an empty string
2277            (Perl fails the current match path).
2278    
2279        (c) A data ] in a character class must be notated as \] because if the
2280            first data character in a class is ], it defines an empty class. (In
2281            Perl it is not possible to have an empty class.) The empty class []
2282            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2283            The negative empty class [^] matches any one character, independently
2284            of the DOTALL setting.
2285    
2286    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2287        non-existent subpattern following a character class starting with ']' and
2288        containing () gave an internal compiling error instead of "reference to
2289        non-existent subpattern". Fortunately, when the pattern did exist, the
2290        compiled code was correct. (When scanning forwards to check for the
2291        existence of the subpattern, it was treating the data ']' as terminating
2292        the class, so got the count wrong. When actually compiling, the reference
2293        was subsequently set up correctly.)
2294    
2295    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2296        it was being rejected as not supported by pcre_dfa_exec(), even though
2297        other assertions are supported. I have made pcre_dfa_exec() support
2298        (*FAIL).
2299    
2300    16. The implementation of 13c above involved the invention of a new opcode,
2301        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2302        cannot be changed at match time, I realized I could make a small
2303        improvement to matching performance by compiling OP_ALLANY instead of
2304        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2305        on the OP_ANY path.
2306    
2307    17. Compiling pcretest on Windows with readline support failed without the
2308        following two fixes: (1) Make the unistd.h include conditional on
2309        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2310    
2311    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2312        ncurses library to be included for pcretest when ReadLine support is
2313        requested, but also to allow for it to be overridden. This patch came from
2314        Daniel Bergström.
2315    
2316    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2317        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2318        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2319        spotting this.
2320    
2321    
2322    Version 7.6 28-Jan-08
2323    ---------------------
2324    
2325    1.  A character class containing a very large number of characters with
2326        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2327        overflow.
2328    
2329    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2330        HAVE_LONG_LONG is not defined.
2331    
2332    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2333        bring them up to date and include new features. This patch includes:
2334    
2335        - Fixed PH's badly added libz and libbz2 support.
2336        - Fixed a problem with static linking.
2337        - Added pcredemo. [But later removed - see 7 below.]
2338        - Fixed dftables problem and added an option.
2339        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2340            HAVE_LONG_LONG.
2341        - Added readline support for pcretest.
2342        - Added an listing of the option settings after cmake has run.
2343    
2344    4.  A user submitted a patch to Makefile that makes it easy to create
2345        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2346        Makefile.am that cause it to include this special target, without
2347        affecting anything else. Note that the same mingw target plus all
2348        the other distribution libraries and programs are now supported
2349        when configuring with CMake (see 6 below) instead of with
2350        Configure/Make.
2351    
2352    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2353        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2354        exported in the Windows port". It has not yet been confirmed that the patch
2355        solves the problem, but it does no harm.
2356    
2357    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2358        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2359        with CMake, and also correct the comment about stack recursion.
2360    
2361    7.  Remove the automatic building of pcredemo from the ./configure system and
2362        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2363        of a program that users should build themselves after PCRE is installed, so
2364        building it automatically is not really right. What is more, it gave
2365        trouble in some build environments.
2366    
2367    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2368    
2369    
2370    Version 7.5 10-Jan-08
2371    ---------------------
2372    
2373    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2374        values in parens when parsing an RE using the C++ wrapper."
2375    
2376    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2377        Characters greater than 255 were excluded from the class instead of being
2378        included.
2379    
2380    3.  The same bug as (2) above applied to negated POSIX classes such as
2381        [:^space:].
2382    
2383    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2384        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2385        I have changed it.
2386    
2387    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2388        first named subpattern) and a construct such as (?&a) would reference the
2389        first named subpattern whose name started with "a" (in other words, the
2390        length check was missing). Both these problems are fixed. "Subpattern name
2391        expected" is now given for (?&) (a zero-length name), and this patch also
2392        makes it give the same error for \k'' (previously it complained that that
2393        was a reference to a non-existent subpattern).
2394    
2395    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2396        this is right because (?- can be followed by option settings as well as by
2397        digits. I have, however, made the messages clearer.
2398    
2399    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2400        than the number used in the conditional) now cause a compile-time error.
2401        This is actually not compatible with Perl, which accepts such patterns, but
2402        treats the conditional as always being FALSE (as PCRE used to), but it
2403        seems to me that giving a diagnostic is better.
2404    
2405    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2406        and messages.
2407    
2408    9.  Fix two occurrences of "backslash" in comments that should have been
2409        "backspace".
2410    
2411    10. Remove two redundant lines of code that can never be obeyed (their function
2412        was moved elsewhere).
2413    
2414    11. The program that makes PCRE's Unicode character property table had a bug
2415        which caused it to generate incorrect table entries for sequences of
2416        characters that have the same character type, but are in different scripts.
2417        It amalgamated them into a single range, with the script of the first of
2418        them. In other words, some characters were in the wrong script. There were
2419        thirteen such cases, affecting characters in the following ranges:
2420    
2421          U+002b0 - U+002c1
2422          U+0060c - U+0060d
2423          U+0061e - U+00612
2424          U+0064b - U+0065e
2425          U+0074d - U+0076d
2426          U+01800 - U+01805
2427          U+01d00 - U+01d77
2428          U+01d9b - U+01dbf
2429          U+0200b - U+0200f
2430          U+030fc - U+030fe
2431          U+03260 - U+0327f
2432          U+0fb46 - U+0fbb1
2433          U+10450 - U+1049d
2434    
2435    12. The -o option (show only the matching part of a line) for pcregrep was not
2436        compatible with GNU grep in that, if there was more than one match in a
2437        line, it showed only the first of them. It now behaves in the same way as
2438        GNU grep.
2439    
2440    13. If the -o and -v options were combined for pcregrep, it printed a blank
2441        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2442        does the same. The return code can be used to tell if there were any
2443        non-matching lines.
2444    
2445    14. Added --file-offsets and --line-offsets to pcregrep.
2446    
2447    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2448        infinitely looping recursion. The bug was that positive lookaheads were not
2449        being skipped when checking for a possible empty match (negative lookaheads
2450        and both kinds of lookbehind were skipped).
2451    
2452    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2453        inclusion of <windows.h> to before rather than after the definition of
2454        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2455    
2456    17. Specifying a possessive quantifier with a specific limit for a Unicode
2457        character property caused pcre_compile() to compile bad code, which led at
2458        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2459        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2460        caused the error; without that there was no problem.
2461    
2462    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2463    
2464    19. Added --enable-pcretest-libreadline.
2465    
2466    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2467        RE::GlobalReplace(). As a result, the number of replacements returned was
2468        double what it should be. I removed one of the increments, but Craig sent a
2469        later patch that removed the other one (the right fix) and added unit tests
2470        that check the return values (which was not done before).
2471    
2472    21. Several CMake things:
2473    
2474        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2475            the names libpcre and libpcreposix, not just pcre and pcreposix.
2476    
2477        (2) The above change means that pcretest and pcregrep are now correctly
2478            linked with the newly-built libraries, not previously installed ones.
2479    
2480        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2481    
2482    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2483        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2484        UTF-8 newline character). The key issue is that the pattern starts .*;
2485        this means that the match must be either at the beginning, or after a
2486        newline. The bug was in the code for advancing after a failed match and
2487        checking that the new position followed a newline. It was not taking
2488        account of UTF-8 characters correctly.
2489    
2490    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2491        character classes. PCRE was not treating the sequence [:...:] as a
2492        character class unless the ... were all letters. Perl, however, seems to
2493        allow any characters between [: and :], though of course it rejects as
2494        unknown any "names" that contain non-letters, because all the known class
2495        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2496        for example, whereas PCRE did not - it did not recognize a POSIX character
2497        class. This seemed a bit dangerous, so the code has been changed to be
2498        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2499        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2500        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2501        Perl does, and where it didn't before.
2502    
2503    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2504        Windows environments %n is disabled by default.
2505    
2506    
2507    Version 7.4 21-Sep-07
2508    ---------------------
2509    
2510    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2511        means that a class such as [\s] counted as "explicit reference to CR or
2512        LF". That isn't really right - the whole point of the change was to try to
2513        help when there was an actual mention of one of the two characters. So now
2514        the change happens only if \r or \n (or a literal CR or LF) character is
2515        encountered.
2516    
2517    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2518        of both had grown to the point where there were only 3 bits left.
2519        Fortunately, there was spare space in the data structure, and so I have
2520        moved the internal flags into a new 16-bit field to free up more option
2521        bits.
2522    
2523    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2524        but did not set the internal JCHANGED flag - either of these is enough to
2525        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2526        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2527        start sets both bits.
2528    
2529    4.  Added options (at build time, compile time, exec time) to change \R from
2530        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2531    
2532    5.  doc/pcresyntax.html was missing from the distribution.
2533    
2534    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2535        compatibility, even though it is no longer used.
2536    
2537    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2538        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2539        windows.h file is present (where different names are used). [This was
2540        reversed later after testing - see 16 below.]
2541    
2542    8.  Changed all #include <config.h> to #include "config.h". There were also
2543        some further <pcre.h> cases that I changed to "pcre.h".
2544    
2545    9.  When pcregrep was used with the --colour option, it missed the line ending
2546        sequence off the lines that it output.
2547    
2548    10. It was pointed out to me that arrays of string pointers cause lots of
2549        relocations when a shared library is dynamically loaded. A technique of
2550        using a single long string with a table of offsets can drastically reduce
2551        these. I have refactored PCRE in four places to do this. The result is
2552        dramatic:
2553    
2554          Originally:                          290
2555          After changing UCP table:            187
2556          After changing error message table:   43
2557          After changing table of "verbs"       36
2558          After changing table of Posix names   22
2559    
2560        Thanks to the folks working on Gregex for glib for this insight.
2561    
2562    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2563        unicode-properties was also set.
2564    
2565    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2566    
2567    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2568        checked only for CRLF.
2569    
2570    14. Added casts to pcretest.c to avoid compiler warnings.
2571    
2572    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2573    
2574    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2575        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2576        entirely. This removes changes made in 7 above.
2577    
2578    17. The CMake files have been updated, and there is now more information about
2579        building with CMake in the NON-UNIX-USE document.
2580    
2581    
2582    Version 7.3 28-Aug-07
2583    ---------------------
2584    
2585     1. In the rejigging of the build system that eventually resulted in 7.1, the
2586        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2587        brackets there is not right, since it causes compilers to look for an
2588        installed pcre.h, not the version that is in the source that is being
2589        compiled (which of course may be different). I have changed it back to:
2590    
2591          #include "pcre.h"
2592    
2593        I have a vague recollection that the change was concerned with compiling in
2594        different directories, but in the new build system, that is taken care of
2595        by the VPATH setting the Makefile.
2596    
2597     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2598        when the subject happened to end in the byte 0x85 (e.g. if the last
2599        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2600        characters but of course it shouldn't be taken as a newline when it is part
2601        of another character. The bug was that, for an unlimited repeat of . in
2602        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2603        characters when looking for a newline.
2604    
2605     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2606    
2607     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2608        in debug output.
2609    
2610     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2611        long printing in the pcrecpp unittest when running under MinGW.
2612    
2613     6. ESC_K was left out of the EBCDIC table.
2614    
2615     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2616        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2617        limit also applies to "virtual nesting" when a pattern is recursive, and in
2618        this case 1000 isn't so big. I have been able to remove this limit at the
2619        expense of backing off one optimization in certain circumstances. Normally,
2620        when pcre_exec() would call its internal match() function recursively and
2621        immediately return the result unconditionally, it uses a "tail recursion"
2622        feature to save stack. However, when a subpattern that can match an empty
2623        string has an unlimited repetition quantifier, it no longer makes this
2624        optimization. That gives it a stack frame in which to save the data for
2625        checking that an empty string has been matched. Previously this was taken
2626        from the 1000-entry workspace that had been reserved. So now there is no
2627        explicit limit, but more stack is used.
2628    
2629     8. Applied Daniel's patches to solve problems with the import/export magic
2630        syntax that is required for Windows, and which was going wrong for the
2631        pcreposix and pcrecpp parts of the library. These were overlooked when this
2632        problem was solved for the main library.
2633    
2634     9. There were some crude static tests to avoid integer overflow when computing
2635        the size of patterns that contain repeated groups with explicit upper
2636        limits. As the maximum quantifier is 65535, the maximum group length was
2637        set at 30,000 so that the product of these two numbers did not overflow a
2638        32-bit integer. However, it turns out that people want to use groups that
2639        are longer than 30,000 bytes (though not repeat them that many times).
2640        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2641        made it possible to implement the integer overflow checks in a much more
2642        dynamic way, which I have now done. The artificial limitation on group
2643        length has been removed - we now have only the limit on the total length of
2644        the compiled pattern, which depends on the LINK_SIZE setting.
2645    
2646    10. Fixed a bug in the documentation for get/copy named substring when
2647        duplicate names are permitted. If none of the named substrings are set, the
2648        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2649        empty string.
2650    
2651    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2652        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2653        because the ] is interpreted as the first data character and the
2654        terminating ] is not found. PCRE has been made compatible with Perl in this
2655        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2656        cause memory overwriting.
2657    
2658    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2659        string has been matched (to stop an infinite loop). It was not recognizing
2660        a conditional subpattern that could match an empty string if that
2661        subpattern was within another subpattern. For example, it looped when
2662        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2663        condition was not nested. This bug has been fixed.
2664    
2665    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2666        past the start of the subject in the presence of bytes with the top bit
2667        set, for example "\x8aBCD".
2668    
2669    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2670        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2671    
2672    14. Optimized (?!) to (*FAIL).
2673    
2674    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2675        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2676        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2677        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2678        does: it's just the validity check that is more restrictive.
2679    
2680    16. Inserted checks for integer overflows during escape sequence (backslash)
2681        processing, and also fixed erroneous offset values for syntax errors during
2682        backslash processing.
2683    
2684    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2685        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2686    
2687    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2688        caused an overrun.
2689    
2690    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2691        something other than just ASCII characters) inside a group that had an
2692        unlimited repeat caused a loop at compile time (while checking to see
2693        whether the group could match an empty string).
2694    
2695    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2696        [\P{Any}] did so. (Error in the code for printing property names.)
2697    
2698    21. An orphan \E inside a character class could cause a crash.
2699    
2700    22. A repeated capturing bracket such as (A)? could cause a wild memory
2701        reference during compilation.
2702    
2703    23. There are several functions in pcre_compile() that scan along a compiled
2704        expression for various reasons (e.g. to see if it's fixed length for look
2705        behind). There were bugs in these functions when a repeated \p or \P was
2706        present in the pattern. These operators have additional parameters compared
2707        with \d, etc, and these were not being taken into account when moving along
2708        the compiled data. Specifically:
2709    
2710        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2711            length.
2712    
2713        (b) An item such as \pL+ within a repeated group could cause crashes or
2714            loops.
2715    
2716        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2717            "reference to non-existent subpattern" error.
2718    
2719        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2720    
2721    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2722        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2723    
2724    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2725    
2726    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2727        character were causing crashes (broken optimization).
2728    
2729    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2730        \p or \P) caused a compile-time loop.
2731    
2732    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2733        break. For example, the unstudied pattern [\r\n]A does not match the string
2734        "\r\nA" because change 7.0/46 below moves the current point on by two
2735        characters after failing to match at the start. However, the pattern \nA
2736        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2737        the same is true. There doesn't seem any very clean way out of this, but
2738        what I have chosen to do makes the common cases work: PCRE now takes note
2739        of whether there can be an explicit match for \r or \n anywhere in the
2740        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2741        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2742        pattern has explicit CR or LF references.
2743    
2744    29. Added (*CR) etc for changing newline setting at start of pattern.
2745    
2746    
2747    Version 7.2 19-Jun-07
2748    ---------------------
2749    
2750     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2751        which is apparently normally available under Windows.
2752    
2753     2. Re-jig the pcregrep tests with different newline settings in an attempt
2754        to make them independent of the local environment's newline setting.
2755    
2756     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2757    
2758     4. Some of the "internals" tests were previously cut out when the link size
2759        was not 2, because the output contained actual offsets. The recent new
2760        "Z" feature of pcretest means that these can be cut out, making the tests
2761        usable with all link sizes.
2762    
2763     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2764        stack recursion. This gives a massive performance boost under BSD, but just
2765        a small improvement under Linux. However, it saves one field in the frame
2766        in all cases.
2767    
2768     6. Added more features from the forthcoming Perl 5.10:
2769    
2770        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2771            recursion call. It refers to the nth most recently opened parentheses.
2772    
2773        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2774            to be opened parentheses.
2775    
2776        (c) Conditions that refer to capturing parentheses can be specified
2777            relatively, for example, (?(-2)... or (?(+3)...
2778    
2779        (d) \K resets the start of the current match so that everything before
2780            is not part of it.
2781    
2782        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2783    
2784        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2785            reference syntax.
2786    
2787        (g) (?| introduces a group in which the numbering of parentheses in each
2788            alternative starts with the same number.
2789    
2790        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2791    
2792     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2793        PCRE_INFO_JCHANGED.
2794    
2795     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2796        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2797        for detecting groups that can match an empty string.
2798    
2799     9. A pattern with a very large number of alternatives (more than several
2800        hundred) was running out of internal workspace during the pre-compile
2801        phase, where pcre_compile() figures out how much memory will be needed. A
2802        bit of new cunning has reduced the workspace needed for groups with
2803        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2804        workspace instead of running out of the 4096 that are available.
2805    
2806    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2807    
2808    11. Applied patch from Google to remove an optimization that didn't quite work.
2809        The report of the bug said:
2810    
2811          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2812          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2813          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2814    
2815    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2816        it matched the wrong number of bytes.
2817    
2818    
2819    Version 7.1 24-Apr-07
2820  ---------------------  ---------------------
2821    
2822   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
# Line 33  Version 7.1 12-Mar-07 Line 2848  Version 7.1 12-Mar-07
2848   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan   5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2849      Weber: (1) pcre_internal.h was missing some function renames; (2) updated      Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2850      makevp.bat for the current PCRE, using the additional files      makevp.bat for the current PCRE, using the additional files
2851      makevp-compile.txt, makevp-linklib.txt, and pcregexp.pas.      makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2852    
2853   6. A Windows user reported a minor discrepancy with test 2, which turned out   6. A Windows user reported a minor discrepancy with test 2, which turned out
2854      to be caused by a trailing space on an input line that had got lost in his      to be caused by a trailing space on an input line that had got lost in his
# Line 116  Version 7.1 12-Mar-07 Line 2931  Version 7.1 12-Mar-07
2931  14. If erroroffset was passed as NULL to pcre_compile, it provoked a  14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2932      segmentation fault instead of returning the appropriate error message.      segmentation fault instead of returning the appropriate error message.
2933    
2934    15. In multiline mode when the newline sequence was set to "any", the pattern
2935        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2936        This doesn't seem right; it now treats the CRLF combination as the line
2937        ending, and so does not match in that case. It's only a pattern such as ^$
2938        that would hit this one: something like ^ABC$ would have failed after \r
2939        and then tried again after \r\n.
2940    
2941    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2942        in an attempt to make files that differ only in their line terminators
2943        compare equal. This works on Linux.
2944    
2945    17. Under certain error circumstances pcregrep might try to free random memory
2946        as it exited. This is now fixed, thanks to valgrind.
2947    
2948    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2949        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2950        was because its rules for how to advance for /g after matching an empty
2951        string at the end of a line did not allow for this case. They now check for
2952        it specially.
2953    
2954    20. pcretest is supposed to handle patterns and data of any length, by
2955        extending its buffers when necessary. It was getting this wrong when the
2956        buffer for a data line had to be extended.
2957    
2958    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2959        CRLF as a newline sequence.
2960    
2961    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2962        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2963        I have nevertheless tidied it up.
2964    
2965    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2966    
2967    24. Added a man page for pcre-config.
2968    
2969    
2970  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
2971  ---------------------  ---------------------

Legend:
Removed from v.127  
changed lines
  Added in v.1478

  ViewVC Help
Powered by ViewVC 1.1.5