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

Diff of /code/trunk/ChangeLog

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

revision 182 by ph10, Wed Jun 13 15:09:54 2007 UTC revision 1528 by ph10, Sat Feb 28 11:48:09 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 13-June-07  Version 8.37 xx-xxx-2015
5    ------------------------
6    
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13    
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16    
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
19    
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
22    
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
26    
27    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
28        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30    
31    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
32        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33    
34    8. When a pattern is compiled, it remembers the highest back reference so that
35       when matching, if the ovector is too small, extra memory can be obtained to
36       use instead. A conditional subpattern whose condition is a check on a
37       capture having happened, such as, for example in the pattern
38       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39       setting the highest backreference number. This mattered only if pcre_exec()
40       was called with an ovector that was too small to hold the capture, and there
41       was no other kind of back reference (a situation which is probably quite
42       rare). The effect of the bug was that the condition was always treated as
43       FALSE when the capture could not be consulted, leading to a incorrect
44       behaviour by pcre2_match(). This bug has been fixed.
45    
46    9. A reference to a duplicated named group (either a back reference or a test
47       for being set in a conditional) that occurred in a part of the pattern where
48       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49       to be incorrectly calculated, leading to overwriting.
50    
51    10. A mutually recursive set of back references such as (\2)(\1) caused a
52        segfault at study time (while trying to find the minimum matching length).
53        The infinite loop is now broken (with the minimum length unset, that is,
54        zero).
55    
56    11. If an assertion that was used as a condition was quantified with a minimum
57        of zero, matching went wrong. In particular, if the whole group had
58        unlimited repetition and could match an empty string, a segfault was
59        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60        assertions to be quantified, but not if they are being used as conditions,
61        so the above pattern is faulted by Perl. PCRE has now been changed so that
62        it also rejects such patterns.
63    
64    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65        failed to allow the zero-repeat case if pcre2_exec() was called with an
66        ovector too small to capture the group.
67    
68    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69        Red Hat Product Security:
70    
71        (a) A crash if /K and /F were both set with the option to save the compiled
72        pattern.
73    
74        (b) Another crash if the option to print captured substrings in a callout
75        was combined with setting a null ovector, for example \O\C+ as a subject
76        string.
77    
78    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79        forward reference repeated a large (but limited) number of times within a
80        repeated outer group that has a zero minimum quantifier, caused incorrect
81        code to be compiled, leading to the error "internal error:
82        previously-checked referenced subpattern not found" when an incorrect
83        memory address was read. This bug was reported as "heap overflow",
84        discovered by Kai Lu of Fortinet's FortiGuard Labs.
85    
86    
87    Version 8.36 26-September-2014
88    ------------------------------
89    
90    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
91        -Wmissing-field-initializers and -Wunused-parameter.
92    
93    2.  The tests for quantifiers being too big (greater than 65535) were being
94        applied after reading the number, and stupidly assuming that integer
95        overflow would give a negative number. The tests are now applied as the
96        numbers are read.
97    
98    3.  Tidy code in pcre_exec.c where two branches that used to be different are
99        now the same.
100    
101    4.  The JIT compiler did not generate match limit checks for certain
102        bracketed expressions with quantifiers. This may lead to exponential
103        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
104        issue should be resolved now.
105    
106    5.  Fixed an issue, which occures when nested alternatives are optimized
107        with table jumps.
108    
109    6.  Inserted two casts and changed some ints to size_t in the light of some
110        reported 64-bit compiler warnings (Bugzilla 1477).
111    
112    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
113        an empty string, which sometimes were behaving incorrectly in the
114        interpreter (though correctly in the JIT matcher). This pcretest input is
115        an example:
116    
117          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
118          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
119    
120        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
121        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
122        for an empty string was breaking the inner loop and carrying on at a lower
123        level, when possessive repeated groups should always return to a higher
124        level as they have no backtrack points in them. The empty string test now
125        occurs at the outer level.
126    
127    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
128        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
129    
130    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
131        doesn't).
132    
133    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
134        to hold a pattern to be incorrectly computed (too small) when there were
135        named back references to duplicated names. This could cause "internal
136        error: code overflow" or "double free or corruption" or other memory
137        handling errors.
138    
139    11. When named subpatterns had the same prefixes, back references could be
140        confused. For example, in this pattern:
141    
142          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
143    
144        the reference to 'Name' was incorrectly treated as a reference to a
145        duplicate name.
146    
147    12. A pattern such as /^s?c/mi8 where the optional character has more than
148        one "other case" was incorrectly compiled such that it would only try to
149        match starting at "c".
150    
151    13. When a pattern starting with \s was studied, VT was not included in the
152        list of possible starting characters; this should have been part of the
153        8.34/18 patch.
154    
155    14. If a character class started [\Qx]... where x is any character, the class
156        was incorrectly terminated at the ].
157    
158    15. If a pattern that started with a caseless match for a character with more
159        than one "other case" was studied, PCRE did not set up the starting code
160        unit bit map for the list of possible characters. Now it does. This is an
161        optimization improvement, not a bug fix.
162    
163    16. The Unicode data tables have been updated to Unicode 7.0.0.
164    
165    17. Fixed a number of memory leaks in pcregrep.
166    
167    18. Avoid a compiler warning (from some compilers) for a function call with
168        a cast that removes "const" from an lvalue by using an intermediate
169        variable (to which the compiler does not object).
170    
171    19. Incorrect code was compiled if a group that contained an internal recursive
172        back reference was optional (had quantifier with a minimum of zero). This
173        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
174        caused segmentation faults because of stack overflows at compile time.
175    
176    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
177        group that is quantified with an indefinite repeat, caused a compile-time
178        loop which used up all the system stack and provoked a segmentation fault.
179        This was not the same bug as 19 above.
180    
181    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
182        Patch by Mike Frysinger.
183    
184    
185    Version 8.35 04-April-2014
186    --------------------------
187    
188    1.  A new flag is set, when property checks are present in an XCLASS.
189        When this flag is not set, PCRE can perform certain optimizations
190        such as studying these XCLASS-es.
191    
192    2.  The auto-possessification of character sets were improved: a normal
193        and an extended character set can be compared now. Furthermore
194        the JIT compiler optimizes more character set checks.
195    
196    3.  Got rid of some compiler warnings for potentially uninitialized variables
197        that show up only when compiled with -O2.
198    
199    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
200        of a match later then the end of the match. The pcretest program was not
201        handling the case sensibly - it was outputting from the start to the next
202        binary zero. It now reports this situation in a message, and outputs the
203        text from the end to the start.
204    
205    5.  Fast forward search is improved in JIT. Instead of the first three
206        characters, any three characters with fixed position can be searched.
207        Search order: first, last, middle.
208    
209    6.  Improve character range checks in JIT. Characters are read by an inprecise
210        function now, which returns with an unknown value if the character code is
211        above a certain threshold (e.g: 256). The only limitation is that the value
212        must be bigger than the threshold as well. This function is useful when
213        the characters above the threshold are handled in the same way.
214    
215    7.  The macros whose names start with RAWUCHAR are placeholders for a future
216        mode in which only the bottom 21 bits of 32-bit data items are used. To
217        make this more memorable for those maintaining the code, the names have
218        been changed to start with UCHAR21, and an extensive comment has been added
219        to their definition.
220    
221    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
222        to the export list in Makefile.am (they were accidentally omitted from the
223        8.34 tarball).
224    
225    9.  The informational output from pcretest used the phrase "starting byte set"
226        which is inappropriate for the 16-bit and 32-bit libraries. As the output
227        for "first char" and "need char" really means "non-UTF-char", I've changed
228        "byte" to "char", and slightly reworded the output. The documentation about
229        these values has also been (I hope) clarified.
230    
231    10. Another JIT related optimization: use table jumps for selecting the correct
232        backtracking path, when more than four alternatives are present inside a
233        bracket.
234    
235    11. Empty match is not possible, when the minimum length is greater than zero,
236        and there is no \K in the pattern. JIT should avoid empty match checks in
237        such cases.
238    
239    12. In a caseless character class with UCP support, when a character with more
240        than one alternative case was not the first character of a range, not all
241        the alternative cases were added to the class. For example, s and \x{17f}
242        are both alternative cases for S: the class [RST] was handled correctly,
243        but [R-T] was not.
244    
245    13. The configure.ac file always checked for pthread support when JIT was
246        enabled. This is not used in Windows, so I have put this test inside a
247        check for the presence of windows.h (which was already tested for).
248    
249    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
250        The algorithm provides a way to skip certain starting offsets, and usually
251        faster than linear prefix searches.
252    
253    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
254        as for 'fr_FR' and 'french'. For some reason, however, it then used the
255        Windows-specific input and output files, which have 'french' screwed in.
256        So this could never have worked. One of the problems with locales is that
257        they aren't always the same. I have now updated RunTest so that it checks
258        the output of the locale test (test 3) against three different output
259        files, and it allows the test to pass if any one of them matches. With luck
260        this should make the test pass on some versions of Solaris where it was
261        failing. Because of the uncertainty, the script did not used to stop if
262        test 3 failed; it now does. If further versions of a French locale ever
263        come to light, they can now easily be added.
264    
265    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
266        there was a message during ./configure, but it did not stop. This now
267        provokes an error. The invalid example in README has been corrected.
268        If a value less than the minimum is given, the minimum value has always
269        been used, but now a warning is given.
270    
271    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
272        was a bug in the test system, which is now fixed. Also, the list of various
273        configurations that are tested for each release did not have one with both
274        16/32 bits and --enable-bar-anycrlf. It now does.
275    
276    18. pcretest was missing "-C bsr" for displaying the \R default setting.
277    
278    19. Little endian PowerPC systems are supported now by the JIT compiler.
279    
280    20. The fast forward newline mechanism could enter to an infinite loop on
281        certain invalid UTF-8 input. Although we don't support these cases
282        this issue can be fixed by a performance optimization.
283    
284    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
285        not take account if existing stack usage. There is now a new global
286        variable called pcre_stack_guard that can be set to point to an external
287        function to check stack availability. It is called at the start of
288        processing every parenthesized group.
289    
290    22. A typo in the code meant that in ungreedy mode the max/min qualifier
291        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
292        not match "ab".
293    
294    23. When UTF was disabled, the JIT program reported some incorrect compile
295        errors. These messages are silenced now.
296    
297    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
298        compiler.
299    
300    25. Change all the temporary files used in RunGrepTest to be different to those
301        used by RunTest so that the tests can be run simultaneously, for example by
302        "make -j check".
303    
304    
305    Version 8.34 15-December-2013
306    -----------------------------
307    
308    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
309        executable memory. Patch inspired by Carsten Klein.
310    
311    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
312        this macro is never tested and has no effect, because the work to support
313        coverage involves only compiling and linking options and special targets in
314        the Makefile. The comment in config.h implied that defining the macro would
315        enable coverage support, which is totally false. There was also support for
316        setting this macro in the CMake files (my fault, I just copied it from
317        configure). SUPPORT_GCOV has now been removed.
318    
319    3.  Make a small performance improvement in strlen16() and strlen32() in
320        pcretest.
321    
322    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
323        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
324        these cases). There was also one in pcretest.c.
325    
326    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
327    
328    6.  In UTF mode, the code for checking whether a group could match an empty
329        string (which is used for indefinitely repeated groups to allow for
330        breaking an infinite loop) was broken when the group contained a repeated
331        negated single-character class with a character that occupied more than one
332        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
333        UTF-8 mode). The effect was undefined: the group might or might not be
334        deemed as matching an empty string, or the program might have crashed.
335    
336    7.  The code for checking whether a group could match an empty string was not
337        recognizing that \h, \H, \v, \V, and \R must match a character.
338    
339    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
340        an empty string. If it can, pcretest shows this in its information output.
341    
342    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
343        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
344        matched by pcre_exec() without using JIT:
345    
346        (a) If the rest of the pattern did not match after a maximal run of
347            grapheme clusters, the code for backing up to try with fewer of them
348            did not always back up over a full grapheme when characters that do not
349            have the modifier quality were involved, e.g. Hangul syllables.
350    
351        (b) If the match point in a subject started with modifier character, and
352            there was no match, the code could incorrectly back up beyond the match
353            point, and potentially beyond the first character in the subject,
354            leading to a segfault or an incorrect match result.
355    
356    10. A conditional group with an assertion condition could lead to PCRE
357        recording an incorrect first data item for a match if no other first data
358        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
359        first data item, and therefore matched "ca" after "c" instead of at the
360        start.
361    
362    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
363        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
364    
365    12. The source of pcregrep now includes z/OS-specific code so that it can be
366        compiled for z/OS as part of the special z/OS distribution.
367    
368    13. Added the -T and -TM options to pcretest.
369    
370    14. The code in pcre_compile.c for creating the table of named capturing groups
371        has been refactored. Instead of creating the table dynamically during the
372        actual compiling pass, the information is remembered during the pre-compile
373        pass (on the stack unless there are more than 20 named groups, in which
374        case malloc() is used) and the whole table is created before the actual
375        compile happens. This has simplified the code (it is now nearly 150 lines
376        shorter) and prepared the way for better handling of references to groups
377        with duplicate names.
378    
379    15. A back reference to a named subpattern when there is more than one of the
380        same name now checks them in the order in which they appear in the pattern.
381        The first one that is set is used for the reference. Previously only the
382        first one was inspected. This change makes PCRE more compatible with Perl.
383    
384    16. Unicode character properties were updated from Unicode 6.3.0.
385    
386    17. The compile-time code for auto-possessification has been refactored, based
387        on a patch by Zoltan Herczeg. It now happens after instead of during
388        compilation. The code is cleaner, and more cases are handled. The option
389        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
390        options in pcretest are provided to set it. It can also be set by
391        (*NO_AUTO_POSSESS) at the start of a pattern.
392    
393    18. The character VT has been added to the default ("C" locale) set of
394        characters that match \s and are generally treated as white space,
395        following this same change in Perl 5.18. There is now no difference between
396        "Perl space" and "POSIX space". Whether VT is treated as white space in
397        other locales depends on the locale.
398    
399    19. The code for checking named groups as conditions, either for being set or
400        for being recursed, has been refactored (this is related to 14 and 15
401        above). Processing unduplicated named groups should now be as fast at
402        numerical groups, and processing duplicated groups should be faster than
403        before.
404    
405    20. Two patches to the CMake build system, by Alexander Barkov:
406    
407          (1) Replace the "source" command by "." in CMakeLists.txt because
408              "source" is a bash-ism.
409    
410          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
411              without these the CMake build does not work on Solaris.
412    
413    21. Perl has changed its handling of \8 and \9. If there is no previously
414        encountered capturing group of those numbers, they are treated as the
415        literal characters 8 and 9 instead of a binary zero followed by the
416        literals. PCRE now does the same.
417    
418    22. Following Perl, added \o{} to specify codepoints in octal, making it
419        possible to specify values greater than 0777 and also making them
420        unambiguous.
421    
422    23. Perl now gives an error for missing closing braces after \x{... instead of
423        treating the string as literal. PCRE now does the same.
424    
425    24. RunTest used to grumble if an inappropriate test was selected explicitly,
426        but just skip it when running all tests. This make it awkward to run ranges
427        of tests when one of them was inappropriate. Now it just skips any
428        inappropriate tests, as it always did when running all tests.
429    
430    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
431        character types such as \d or \w, too many callouts were inserted, and the
432        data that they returned was rubbish.
433    
434    26. In UCP mode, \s was not matching two of the characters that Perl matches,
435        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
436        were matched by \h. The code has now been refactored so that the lists of
437        the horizontal and vertical whitespace characters used for \h and \v (which
438        are defined only in one place) are now also used for \s.
439    
440    27. Add JIT support for the 64 bit TileGX architecture.
441        Patch by Jiong Wang (Tilera Corporation).
442    
443    28. Possessive quantifiers for classes (both explicit and automatically
444        generated) now use special opcodes instead of wrapping in ONCE brackets.
445    
446    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
447        (because it's meaningless), this was not happening when PCRE_CASELESS was
448        set. Not wrong, but inefficient.
449    
450    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
451        properties for \w, \d, etc) is present in a test regex. Otherwise if the
452        test contains no characters greater than 255, Perl doesn't realise it
453        should be using Unicode semantics.
454    
455    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
456        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
457        does in Unicode mode.
458    
459    32. Added the "forbid" facility to pcretest so that putting tests into the
460        wrong test files can sometimes be quickly detected.
461    
462    33. There is now a limit (default 250) on the depth of nesting of parentheses.
463        This limit is imposed to control the amount of system stack used at compile
464        time. It can be changed at build time by --with-parens-nest-limit=xxx or
465        the equivalent in CMake.
466    
467    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
468        errors. Perl warns for these when in warning mode, but PCRE has no facility
469        for giving warnings.
470    
471    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
472        However, this was not working for (?!) because it is optimized to (*FAIL),
473        for which PCRE does not allow quantifiers. The optimization is now disabled
474        when a quantifier follows (?!). I can't see any use for this, but it makes
475        things uniform.
476    
477    36. Perl no longer allows group names to start with digits, so I have made this
478        change also in PCRE. It simplifies the code a bit.
479    
480    37. In extended mode, Perl ignores spaces before a + that indicates a
481        possessive quantifier. PCRE allowed a space before the quantifier, but not
482        before the possessive +. It now does.
483    
484    38. The use of \K (reset reported match start) within a repeated possessive
485        group such as (a\Kb)*+ was not working.
486    
487    40. Document that the same character tables must be used at compile time and
488        run time, and that the facility to pass tables to pcre_exec() and
489        pcre_dfa_exec() is for use only with saved/restored patterns.
490    
491    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
492        features for Builds with MSVC:
493    
494        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
495           builds). The .rc files can be used to set FileDescription and many other
496           attributes.
497    
498        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
499           This allows higher-level build scripts which want .pdb files to avoid
500           hard-coding the exact files needed."
501    
502    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
503        mean "start of word" and "end of word", respectively, as a transition aid.
504    
505    43. A minimizing repeat of a class containing codepoints greater than 255 in
506        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
507        compiled to use the heap for recursion.
508    
509    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
510        is configured.
511    
512    
513    Version 8.33 28-May-2013
514    ------------------------
515    
516    1.  Added 'U' to some constants that are compared to unsigned integers, to
517        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
518        variables that are added to signed variables, to ensure the result is
519        signed and can be negated.
520    
521    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
522        CMake config files.
523    
524    3.  Revise the creation of config.h.generic so that all boolean macros are
525        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
526        overriding via -D on the command line possible.
527    
528    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
529        to unsigned int is reported to make a quite noticeable speed difference in
530        a specific Windows environment. Testing on Linux did also appear to show
531        some benefit (and it is clearly not harmful). Also fixed the definition of
532        Xop which should be unsigned.
533    
534    5.  Related to (4), changing the definition of the intermediate variable cc
535        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
536        improvements.
537    
538    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
539        unnecessary spaces.
540    
541    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
542        and later.
543    
544    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
545    
546    9.  Optimizing fast_forward_start_bits in JIT.
547    
548    10. Adding support for callouts in JIT, and fixing some issues revealed
549        during this work. Namely:
550    
551        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
552    
553        (b) Minimum length was not checked before the matching is started.
554    
555    11. The value of capture_last that is passed to callouts was incorrect in some
556        cases when there was a capture on one path that was subsequently abandoned
557        after a backtrack. Also, the capture_last value is now reset after a
558        recursion, since all captures are also reset in this case.
559    
560    12. The interpreter no longer returns the "too many substrings" error in the
561        case when an overflowing capture is in a branch that is subsequently
562        abandoned after a backtrack.
563    
564    13. In the pathological case when an offset vector of size 2 is used, pcretest
565        now prints out the matched string after a yield of 0 or 1.
566    
567    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
568        Only supported by the JIT compiler at the moment.
569    
570    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
571    
572    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
573        offset of the starting point of the matching process, provided the offsets
574        vector is large enough.
575    
576    17. The \A escape now records a lookbehind value of 1, though its execution
577        does not actually inspect the previous character. This is to ensure that,
578        in partial multi-segment matching, at least one character from the old
579        segment is retained when a new segment is processed. Otherwise, if there
580        are no lookbehinds in the pattern, \A might match incorrectly at the start
581        of a new segment.
582    
583    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
584    
585    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
586        gives some modest performance improvement in 8-bit mode.
587    
588    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
589        be expressed in certain programming languages using Universal Character
590        Names.
591    
592    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
593        which points out that "non characters" are not "characters that may not
594        appear in Unicode strings" but rather "characters that are reserved for
595        internal use and have only local meaning".
596    
597    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
598        there was a conditional group that depended on an assertion, if the
599        assertion was false, the callout that immediately followed the alternation
600        in the condition was skipped when pcre_exec() was used for matching.
601    
602    23. Allow an explicit callout to be inserted before an assertion that is the
603        condition for a conditional group, for compatibility with automatic
604        callouts, which always insert a callout at this point.
605    
606    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
607        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
608    
609    25. (*PRUNE) is now supported by the JIT compiler.
610    
611    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
612    
613    27. Fix the case where there are two or more SKIPs with arguments that may be
614        ignored.
615    
616    28. (*SKIP) is now supported by the JIT compiler.
617    
618    29. (*THEN) is now supported by the JIT compiler.
619    
620    30. Update RunTest with additional test selector options.
621    
622    31. The way PCRE handles backtracking verbs has been changed in two ways.
623    
624        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
625        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
626        backtracking. In some cases this makes it more Perl-compatible, but Perl's
627        rather obscure rules do not always do the same thing.
628    
629        (2) Previously, backtracking verbs were confined within assertions. This is
630        no longer the case for positive assertions, except for (*ACCEPT). Again,
631        this sometimes improves Perl compatibility, and sometimes does not.
632    
633    32. A number of tests that were in test 2 because Perl did things differently
634        have been moved to test 1, because either Perl or PCRE has changed, and
635        these tests are now compatible.
636    
637    32. Backtracking control verbs are now handled in the same way in JIT and
638        interpreter.
639    
640    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
641        contained a forward subroutine reference caused a compile error.
642    
643    34. Auto-detect and optimize limited repetitions in JIT.
644    
645    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
646        blocking (*UTF) etc.
647    
648    36. In the interpreter, maximizing pattern repetitions for characters and
649        character types now use tail recursion, which reduces stack usage.
650    
651    37. The value of the max lookbehind was not correctly preserved if a compiled
652        and saved regex was reloaded on a host of different endianness.
653    
654    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
655        of the compiled pattern block, expand the flags field from 16 to 32 bits
656        because it was almost full.
657    
658    39. Try madvise first before posix_madvise.
659    
660    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
661        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
662        original change locked out all matches of empty strings. This has been
663        changed so that one match of an empty string per line is recognized.
664        Subsequent searches on the same line (for colouring or for --only-matching,
665        for example) do not recognize empty strings.
666    
667    41. Applied a user patch to fix a number of spelling mistakes in comments.
668    
669    42. Data lines longer than 65536 caused pcretest to crash.
670    
671    43. Clarified the data type for length and startoffset arguments for pcre_exec
672        and pcre_dfa_exec in the function-specific man pages, where they were
673        explicitly stated to be in bytes, never having been updated. I also added
674        some clarification to the pcreapi man page.
675    
676    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
677        a segmentation fault.
678    
679    
680    Version 8.32 30-November-2012
681    -----------------------------
682    
683    1.  Improved JIT compiler optimizations for first character search and single
684        character iterators.
685    
686    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
687        Patch by Daniel Richard G.
688    
689    3.  Single character iterator optimizations in the JIT compiler.
690    
691    4.  Improved JIT compiler optimizations for character ranges.
692    
693    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
694        Reported by Giuseppe D'Angelo.
695    
696    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
697        of a line, was being set incorrectly in cases where .* appeared inside
698        atomic brackets at the start of a pattern, or where there was a subsequent
699        *PRUNE or *SKIP.
700    
701    7.  Improved instruction cache flush for POWER/PowerPC.
702        Patch by Daniel Richard G.
703    
704    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
705        grep:
706    
707        (a) There is now no limit to the number of patterns to be matched.
708    
709        (b) An error is given if a pattern is too long.
710    
711        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
712            are now supported.
713    
714        (d) --exclude-from and --include-from (multiple use) have been added.
715    
716        (e) Exclusions and inclusions now apply to all files and directories, not
717            just to those obtained from scanning a directory recursively.
718    
719        (f) Multiple uses of -f and --file-list are now supported.
720    
721        (g) In a Windows environment, the default for -d has been changed from
722            "read" (the GNU grep default) to "skip", because otherwise the presence
723            of a directory in the file list provokes an error.
724    
725        (h) The documentation has been revised and clarified in places.
726    
727    9.  Improve the matching speed of capturing brackets.
728    
729    10. Changed the meaning of \X so that it now matches a Unicode extended
730        grapheme cluster.
731    
732    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
733        out POSIX threads when JIT support is configured.
734    
735    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
736    
737    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
738        structure could go wrong in environments where size_t is not the same size
739        as int.
740    
741    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
742        set.
743    
744    15. The EBCDIC support had decayed; later updates to the code had included
745        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
746        general tidy up of EBCDIC-related issues, and the documentation was also
747        not quite right. There is now a test that can be run on ASCII systems to
748        check some of the EBCDIC-related things (but is it not a full test).
749    
750    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
751        in a small tidy to the code.
752    
753    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
754    
755    18. If the --only-matching (-o) option in pcregrep is specified multiple
756        times, each one causes appropriate output. For example, -o1 -o2 outputs the
757        substrings matched by the 1st and 2nd capturing parentheses. A separating
758        string can be specified by --om-separator (default empty).
759    
760    19. Improving the first n character searches.
761    
762    20. Turn case lists for horizontal and vertical white space into macros so that
763        they are defined only once.
764    
765    21. This set of changes together give more compatible Unicode case-folding
766        behaviour for characters that have more than one other case when UCP
767        support is available.
768    
769        (a) The Unicode property table now has offsets into a new table of sets of
770            three or more characters that are case-equivalent. The MultiStage2.py
771            script that generates these tables (the pcre_ucd.c file) now scans
772            CaseFolding.txt instead of UnicodeData.txt for character case
773            information.
774    
775        (b) The code for adding characters or ranges of characters to a character
776            class has been abstracted into a generalized function that also handles
777            case-independence. In UTF-mode with UCP support, this uses the new data
778            to handle characters with more than one other case.
779    
780        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
781            whose other case is greater than 256 are now correctly matched
782            caselessly. Previously, the high codepoint matched the low one, but not
783            vice versa.
784    
785        (d) The processing of \h, \H, \v, and \ in character classes now makes use
786            of the new class addition function, using character lists defined as
787            macros alongside the case definitions of 20 above.
788    
789        (e) Caseless back references now work with characters that have more than
790            one other case.
791    
792        (f) General caseless matching of characters with more than one other case
793            is supported.
794    
795    22. Unicode character properties were updated from Unicode 6.2.0
796    
797    23. Improved CMake support under Windows. Patch by Daniel Richard G.
798    
799    24. Add support for 32-bit character strings, and UTF-32
800    
801    25. Major JIT compiler update (code refactoring and bugfixing).
802        Experimental Sparc 32 support is added.
803    
804    26. Applied a modified version of Daniel Richard G's patch to create
805        pcre.h.generic and config.h.generic by "make" instead of in the
806        PrepareRelease script.
807    
808    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
809        pcre_compile.c when checking for a zero character.
810    
811    28. Introducing a native interface for JIT. Through this interface, the compiled
812        machine code can be directly executed. The purpose of this interface is to
813        provide fast pattern matching, so several sanity checks are not performed.
814        However, feature tests are still performed. The new interface provides
815        1.4x speedup compared to the old one.
816    
817    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
818        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
819        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
820    
821    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
822        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
823        the "old" RFC 2279). Instead, it ended up passing a negative length to
824        pcre_exec().
825    
826    31. Add support for GCC's visibility feature to hide internal functions.
827    
828    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
829        "unknown -C option" after outputting 0 or 1.
830    
831    33. There is now support for generating a code coverage report for the test
832        suite in environments where gcc is the compiler and lcov is installed. This
833        is mainly for the benefit of the developers.
834    
835    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
836        unaddressable using valgrind annotations, allowing valgrind to detect
837        invalid memory accesses. This is mainly for the benefit of the developers.
838    
839    25. (*UTF) can now be used to start a pattern in any of the three libraries.
840    
841    26. Give configure error if --enable-cpp but no C++ compiler found.
842    
843    
844    Version 8.31 06-July-2012
845    -------------------------
846    
847    1.  Fixing a wrong JIT test case and some compiler warnings.
848    
849    2.  Removed a bashism from the RunTest script.
850    
851    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
852        to unsigned type, result still unsigned" that was given by an MS compiler
853        on encountering the code "-sizeof(xxx)".
854    
855    4.  Partial matching support is added to the JIT compiler.
856    
857    5.  Fixed several bugs concerned with partial matching of items that consist
858        of more than one character:
859    
860        (a) /^(..)\1/ did not partially match "aba" because checking references was
861            done on an "all or nothing" basis. This also applied to repeated
862            references.
863    
864        (b) \R did not give a hard partial match if \r was found at the end of the
865            subject.
866    
867        (c) \X did not give a hard partial match after matching one or more
868            characters at the end of the subject.
869    
870        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
871            a partial match for the string "\r".
872    
873        (e) When newline was set to CRLF, the metacharacter "." did not recognize
874            a partial match for a CR character at the end of the subject string.
875    
876    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
877        running pcretest, the text "(JIT)" added to the output whenever JIT is
878        actually used to run the match.
879    
880    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
881        or /S+[+] with a digit between 1 and 7.
882    
883    8.  OP_NOT now supports any UTF character not just single-byte ones.
884    
885    9.  (*MARK) control verb is now supported by the JIT compiler.
886    
887    10. The command "./RunTest list" lists the available tests without actually
888        running any of them. (Because I keep forgetting what they all are.)
889    
890    11. Add PCRE_INFO_MAXLOOKBEHIND.
891    
892    12. Applied a (slightly modified) user-supplied patch that improves performance
893        when the heap is used for recursion (compiled with --disable-stack-for-
894        recursion). Instead of malloc and free for each heap frame each time a
895        logical recursion happens, frames are retained on a chain and re-used where
896        possible. This sometimes gives as much as 30% improvement.
897    
898    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
899        call.
900    
901    14. As documented, (*COMMIT) is now confined to within a positive assertion.
902    
903    15. It is now possible to link pcretest with libedit as an alternative to
904        libreadline.
905    
906    16. (*COMMIT) control verb is now supported by the JIT compiler.
907    
908    17. The Unicode data tables have been updated to Unicode 6.1.0.
909    
910    18. Added --file-list option to pcregrep.
911    
912    19. Added binary file support to pcregrep, including the -a, --binary-files,
913        -I, and --text options.
914    
915    20. The madvise function is renamed for posix_madvise for QNX compatibility
916        reasons. Fixed by Giuseppe D'Angelo.
917    
918    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
919    
920    22. Changed the option for creating symbolic links for 16-bit man pages from
921        -s to -sf so that re-installing does not cause issues.
922    
923    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
924    
925    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
926        matches in certain environments (the workspace was not being correctly
927        retained). Also added to pcre_dfa_exec() a simple plausibility check on
928        some of the workspace data at the beginning of a restart.
929    
930    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
931        was not doing so when it should - probably a typo introduced by SVN 528
932        (change 8.10/14).
933    
934    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
935        \w+ when the character tables indicated that \x{c4} was a word character.
936        There were several related cases, all because the tests for doing a table
937        lookup were testing for characters less than 127 instead of 255.
938    
939    27. If a pattern contains capturing parentheses that are not used in a match,
940        their slots in the ovector are set to -1. For those that are higher than
941        any matched groups, this happens at the end of processing. In the case when
942        there were back references that the ovector was too small to contain
943        (causing temporary malloc'd memory to be used during matching), and the
944        highest capturing number was not used, memory off the end of the ovector
945        was incorrectly being set to -1. (It was using the size of the temporary
946        memory instead of the true size.)
947    
948    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
949        ovector size, it uses memory at the end of the block that it has got.
950    
951    29. Check for an overlong MARK name and give an error at compile time. The
952        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
953    
954    30. JIT compiler update.
955    
956    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
957        Rill for the patch.
958    
959    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
960    
961    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
962    
963    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
964        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
965        was enabled.
966    
967    35. Improve JIT code generation for greedy plus quantifier.
968    
969    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
970        1 to "aa" instead of to an empty string. The bug affected repeated groups
971        that could potentially match an empty string.
972    
973    37. Optimizing single character iterators in JIT.
974    
975    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
976        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
977        codepoints that are too big for the mode are faulted, and in a UTF mode,
978        disallowed codepoints are also faulted.
979    
980    39. If PCRE was compiled with UTF support, in three places in the DFA
981        matcher there was code that should only have been obeyed in UTF mode, but
982        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
983        processing when bytes with values greater than 127 were present. In 16-bit
984        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
985        both cases the values are those that cannot be the first data item in a UTF
986        character. The three items that might have provoked this were recursions,
987        possessively repeated groups, and atomic groups.
988    
989    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
990        and pcregrep, because some OS require shared objects to be explicitly
991        passed to ld, causing the link step to fail if they are not.
992    
993    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
994        mode, patterns that started with \h* or \R* might be incorrectly matched.
995    
996    
997    Version 8.30 04-February-2012
998    -----------------------------
999    
1000    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1001        name is defined in ctype.h.
1002    
1003    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1004        only in quite long subpatterns.
1005    
1006    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1007        since it was replaced by pcre_fullinfo() in February 2000.
1008    
1009    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1010        match a (*MARK), and the match failed at the start of the subject, a
1011        reference to memory before the start of the subject could occur. This bug
1012        was introduced by fix 17 of release 8.21.
1013    
1014    5.  A reference to an unset group with zero minimum repetition was giving
1015        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1016        /(another)?(\1?)test/ matched against "hello world test". This bug was
1017        introduced in release 8.13.
1018    
1019    6.  Add support for 16-bit character strings (a large amount of work involving
1020        many changes and refactorings).
1021    
1022    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1023        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1024        from a file.
1025    
1026    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1027        rounding is not applied in this particular case).
1028    
1029    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1030        if they appear, or are escaped, in patterns.
1031    
1032    10. Get rid of a number of -Wunused-but-set-variable warnings.
1033    
1034    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1035        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1036        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1037        also returns the mark "x". This bug applied to capturing parentheses,
1038        non-capturing parentheses, and atomic parentheses. It also applied to some
1039        assertions.
1040    
1041    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1042        information out of configure.ac instead of relying on pcre.h.generic, which
1043        is not stored in the repository.
1044    
1045    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1046        -lreadline.
1047    
1048    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1049    
1050    15. Applied Graycode's patch to put the top-level frame on the stack rather
1051        than the heap when not using the stack for recursion. This gives a
1052        performance improvement in many cases when recursion is not deep.
1053    
1054    16. Experimental code added to "pcretest -C" to output the stack frame size.
1055    
1056    
1057    Version 8.21 12-Dec-2011
1058    ------------------------
1059    
1060    1.  Updating the JIT compiler.
1061    
1062    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1063        are added as well.
1064    
1065    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1066        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1067        calling _pcre_jit_exec. Some extra comments are added.
1068    
1069    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1070        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1071        was introduced by change 18 for 8.20.
1072    
1073    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1074        ECMA-262 standard.
1075    
1076    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1077        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1078        This bug was probably introduced by change 9 of 8.13.
1079    
1080    7.  While fixing 6 above, I noticed that a number of other items were being
1081        incorrectly rejected as "not fixed length". This arose partly because newer
1082        opcodes had not been added to the fixed-length checking code. I have (a)
1083        corrected the bug and added tests for these items, and (b) arranged for an
1084        error to occur if an unknown opcode is encountered while checking for fixed
1085        length instead of just assuming "not fixed length". The items that were
1086        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1087        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1088        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1089    
1090    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1091        being incorrectly compiled and would have given unpredicatble results.
1092    
1093    9.  A possessively repeated subpattern with minimum repeat count greater than
1094        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1095        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1096        the first (A) could occur when it should not.
1097    
1098    10. Add a cast and remove a redundant test from the code.
1099    
1100    11. JIT should use pcre_malloc/pcre_free for allocation.
1101    
1102    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1103        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1104        is anything other than /usr, -L is still shown).
1105    
1106    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1107    
1108    14. Perl does not support \N without a following name in a [] class; PCRE now
1109        also gives an error.
1110    
1111    15. If a forward reference was repeated with an upper limit of around 2000,
1112        it caused the error "internal error: overran compiling workspace". The
1113        maximum number of forward references (including repeats) was limited by the
1114        internal workspace, and dependent on the LINK_SIZE. The code has been
1115        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1116        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1117        of around 200,000 forward references. While doing this, I also speeded up
1118        the filling in of repeated forward references.
1119    
1120    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1121        incorrectly expecting the subject to contain another "a" after the start.
1122    
1123    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1124        in the match, the SKIP should be ignored. This was not happening; instead
1125        the SKIP was being treated as NOMATCH. For patterns such as
1126        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1127        tested.
1128    
1129    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1130        now much more compatible with Perl, in particular in cases where the result
1131        is a non-match for a non-anchored pattern. For example, if
1132        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1133        "m", where previously it did not return a name. A side effect of this
1134        change is that for partial matches, the last encountered mark name is
1135        returned, as for non matches. A number of tests that were previously not
1136        Perl-compatible have been moved into the Perl-compatible test files. The
1137        refactoring has had the pleasing side effect of removing one argument from
1138        the match() function, thus reducing its stack requirements.
1139    
1140    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1141        subsequent uses of /S (without +) incorrectly behaved like /S+.
1142    
1143    21. Retrieve executable code size support for the JIT compiler and fixing
1144        some warnings.
1145    
1146    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1147        not work when the shorter character appeared right at the end of the
1148        subject string.
1149    
1150    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1151        systems.
1152    
1153    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1154        output it when the /M option is used in pcretest.
1155    
1156    25. The CheckMan script was not being included in the distribution. Also, added
1157        an explicit "perl" to run Perl scripts from the PrepareRelease script
1158        because this is reportedly needed in Windows.
1159    
1160    26. If study data was being save in a file and studying had not found a set of
1161        "starts with" bytes for the pattern, the data written to the file (though
1162        never used) was taken from uninitialized memory and so caused valgrind to
1163        complain.
1164    
1165    27. Updated RunTest.bat as provided by Sheri Pierce.
1166    
1167    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1168    
1169    29. Computation of memory usage for the table of capturing group names was
1170        giving an unnecessarily large value.
1171    
1172    
1173    Version 8.20 21-Oct-2011
1174    ------------------------
1175    
1176    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1177        a POSIX class. After further experiments with Perl, which convinced me that
1178        Perl has bugs and confusions, a closing square bracket is no longer allowed
1179        in a POSIX name. This bug also affected patterns with classes that started
1180        with full stops.
1181    
1182    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1183        captured substring, but while checking the failing first alternative,
1184        substring 1 is temporarily captured. If the output vector supplied to
1185        pcre_exec() was not big enough for this capture, the yield of the function
1186        was still zero ("insufficient space for captured substrings"). This cannot
1187        be totally fixed without adding another stack variable, which seems a lot
1188        of expense for a edge case. However, I have improved the situation in cases
1189        such as /(a)(b)x|abc/ matched against "abc", where the return code
1190        indicates that fewer than the maximum number of slots in the ovector have
1191        been set.
1192    
1193    3.  Related to (2) above: when there are more back references in a pattern than
1194        slots in the output vector, pcre_exec() uses temporary memory during
1195        matching, and copies in the captures as far as possible afterwards. It was
1196        using the entire output vector, but this conflicts with the specification
1197        that only 2/3 is used for passing back captured substrings. Now it uses
1198        only the first 2/3, for compatibility. This is, of course, another edge
1199        case.
1200    
1201    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1202        main code base, and can be used by building with --enable-jit. When this is
1203        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1204        runtime --no-jit option is given.
1205    
1206    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1207        ovector, the return from the function was zero, implying that there were
1208        other matches that did not fit. The correct "exactly full" value is now
1209        returned.
1210    
1211    6.  If a subpattern that was called recursively or as a subroutine contained
1212        (*PRUNE) or any other control that caused it to give a non-standard return,
1213        invalid errors such as "Error -26 (nested recursion at the same subject
1214        position)" or even infinite loops could occur.
1215    
1216    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1217        computing the minimum length on reaching *ACCEPT, and so ended up with the
1218        wrong value of 1 rather than 0. Further investigation indicates that
1219        computing a minimum subject length in the presence of *ACCEPT is difficult
1220        (think back references, subroutine calls), and so I have changed the code
1221        so that no minimum is registered for a pattern that contains *ACCEPT.
1222    
1223    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1224        it was not handled as intended. [But see 16 below.]
1225    
1226    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1227        Sheri Pierce.
1228    
1229    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1230        the first byte in a match must be "a".
1231    
1232    11. Change 17 for 8.13 increased the recursion depth for patterns like
1233        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1234        pattern contains any instances of (*THEN). If it does not, the old
1235        optimizations are restored. It would be nice to do this on a per-group
1236        basis, but at the moment that is not feasible.
1237    
1238    12. In some environments, the output of pcretest -C is CRLF terminated. This
1239        broke RunTest's code that checks for the link size. A single white space
1240        character after the value is now allowed for.
1241    
1242    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1243        For "fr", it uses the Windows-specific input and output files.
1244    
1245    14. If (*THEN) appeared in a group that was called recursively or as a
1246        subroutine, it did not work as intended. [But see next item.]
1247    
1248    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1249        pattern fragments (but not containing any | characters). If A and B are
1250        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1251        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1252        D. In other words, Perl considers parentheses that do not contain any |
1253        characters to be part of a surrounding alternative, whereas PCRE was
1254        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1255        differently. PCRE now behaves in the same way as Perl, except in the case
1256        of subroutine/recursion calls such as (?1) which have in any case always
1257        been different (but PCRE had them first :-).
1258    
1259    16. Related to 15 above: Perl does not treat the | in a conditional group as
1260        creating alternatives. Such a group is treated in the same way as an
1261        ordinary group without any | characters when processing (*THEN). PCRE has
1262        been changed to match Perl's behaviour.
1263    
1264    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1265        RunGrepTest script failed.
1266    
1267    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1268        inevitable for groups that contain captures, but it can lead to a lot of
1269        stack use in large patterns. The old behaviour has been restored for atomic
1270        groups that do not contain any capturing parentheses.
1271    
1272    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1273        suppress the check for a minimum subject length at run time. (If it was
1274        given to pcre_exec() or pcre_dfa_exec() it did work.)
1275    
1276    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1277        fail to work when decoding hex characters in data strings in EBCDIC
1278        environments.
1279    
1280    21. It appears that in at least one Mac OS environment, the isxdigit() function
1281        is implemented as a macro that evaluates to its argument more than once,
1282        contravening the C 90 Standard (I haven't checked a later standard). There
1283        was an instance in pcretest which caused it to go wrong when processing
1284        \x{...} escapes in subject strings. The has been rewritten to avoid using
1285        things like p++ in the argument of isxdigit().
1286    
1287    
1288    Version 8.13 16-Aug-2011
1289    ------------------------
1290    
1291    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1292    
1293    2.  Two minor typos in pcre_internal.h have been fixed.
1294    
1295    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1296        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1297        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1298    
1299    4.  There were a number of related bugs in the code for matching backrefences
1300        caselessly in UTF-8 mode when codes for the characters concerned were
1301        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1302        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1303        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1304        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1305        2-byte code at the end of the subject (it thought there wasn't enough data
1306        left).
1307    
1308    5.  Comprehensive information about what went wrong is now returned by
1309        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1310        as the output vector has at least 2 elements. The offset of the start of
1311        the failing character and a reason code are placed in the vector.
1312    
1313    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1314        now returned is for the first byte of the failing character, instead of the
1315        last byte inspected. This is an incompatible change, but I hope it is small
1316        enough not to be a problem. It makes the returned offset consistent with
1317        pcre_exec() and pcre_dfa_exec().
1318    
1319    7.  pcretest now gives a text phrase as well as the error number when
1320        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1321        failure, the offset and reason code are output.
1322    
1323    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1324        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1325        back over a single character (\n). This seems wrong (because it treated the
1326        two characters as a single entity when going forwards), conflicts with the
1327        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1328        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1329        has been changed.
1330    
1331    9.  Some internal refactoring has changed the processing so that the handling
1332        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1333        time (the PCRE_DOTALL option was changed this way some time ago: version
1334        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1335        which was always a bit of a fudge. It also means that there is one less
1336        argument for the match() function, which reduces its stack requirements
1337        slightly. This change also fixes an incompatibility with Perl: the pattern
1338        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1339    
1340    10. More internal refactoring has drastically reduced the number of recursive
1341        calls to match() for possessively repeated groups such as (abc)++ when
1342        using pcre_exec().
1343    
1344    11. While implementing 10, a number of bugs in the handling of groups were
1345        discovered and fixed:
1346    
1347        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1348        (a|)*(?1) gave a compile-time internal error.
1349        ((a|)+)+  did not notice that the outer group could match an empty string.
1350        (^a|^)+   was not marked as anchored.
1351        (.*a|.*)+ was not marked as matching at start or after a newline.
1352    
1353    12. Yet more internal refactoring has removed another argument from the match()
1354        function. Special calls to this function are now indicated by setting a
1355        value in a variable in the "match data" data block.
1356    
1357    13. Be more explicit in pcre_study() instead of relying on "default" for
1358        opcodes that mean there is no starting character; this means that when new
1359        ones are added and accidentally left out of pcre_study(), testing should
1360        pick them up.
1361    
1362    14. The -s option of pcretest has been documented for ages as being an old
1363        synonym of -m (show memory usage). I have changed it to mean "force study
1364        for every regex", that is, assume /S for every regex. This is similar to -i
1365        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1366        using it. It makes it easier to run collections of tests with and without
1367        study enabled, and thereby test pcre_study() more easily. All the standard
1368        tests are now run with and without -s (but some patterns can be marked as
1369        "never study" - see 20 below).
1370    
1371    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1372        restoration of the capturing data to the outer values was not happening
1373        correctly.
1374    
1375    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1376        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1377        pattern had matched an empty string, and so incorrectly returned a no
1378        match.
1379    
1380    17. There was optimizing code for the last branch of non-capturing parentheses,
1381        and also for the obeyed branch of a conditional subexpression, which used
1382        tail recursion to cut down on stack usage. Unfortunately, now that there is
1383        the possibility of (*THEN) occurring in these branches, tail recursion is
1384        no longer possible because the return has to be checked for (*THEN). These
1385        two optimizations have therefore been removed. [But see 8.20/11 above.]
1386    
1387    18. If a pattern containing \R was studied, it was assumed that \R always
1388        matched two bytes, thus causing the minimum subject length to be
1389        incorrectly computed because \R can also match just one byte.
1390    
1391    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1392        was incorrectly computed.
1393    
1394    20. If /S is present twice on a test pattern in pcretest input, it now
1395        *disables* studying, thereby overriding the use of -s on the command line
1396        (see 14 above). This is necessary for one or two tests to keep the output
1397        identical in both cases.
1398    
1399    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1400        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1401    
1402    22. When an atomic group that contained a capturing parenthesis was
1403        successfully matched, but the branch in which it appeared failed, the
1404        capturing was not being forgotten if a higher numbered group was later
1405        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1406        group 1 to "a", when in fact it should be unset. This applied to multi-
1407        branched capturing and non-capturing groups, repeated or not, and also to
1408        positive assertions (capturing in negative assertions does not happen
1409        in PCRE) and also to nested atomic groups.
1410    
1411    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1412        subject after a captured substring, to make it easier to tell which of a
1413        number of identical substrings has been captured.
1414    
1415    24. The way atomic groups are processed by pcre_exec() has been changed so that
1416        if they are repeated, backtracking one repetition now resets captured
1417        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1418        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1419        "aaa". Previously, it would have been "a". As part of this code
1420        refactoring, the way recursive calls are handled has also been changed.
1421    
1422    25. If an assertion condition captured any substrings, they were not passed
1423        back unless some other capturing happened later. For example, if
1424        (?(?=(a))a) was matched against "a", no capturing was returned.
1425    
1426    26. When studying a pattern that contained subroutine calls or assertions,
1427        the code for finding the minimum length of a possible match was handling
1428        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1429        group 1 called group 2 while simultaneously a separate group 2 called group
1430        1). A stack overflow occurred in this case. I have fixed this by limiting
1431        the recursion depth to 10.
1432    
1433    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1434        Fortmann. This supports explicit test numbers on the command line, and has
1435        argument validation and error reporting.
1436    
1437    28. An instance of \X with an unlimited repeat could fail if at any point the
1438        first character it looked at was a mark character.
1439    
1440    29. Some minor code refactoring concerning Unicode properties and scripts
1441        should reduce the stack requirement of match() slightly.
1442    
1443    30. Added the '=' option to pcretest to check the setting of unused capturing
1444        slots at the end of the pattern, which are documented as being -1, but are
1445        not included in the return count.
1446    
1447    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1448        compiled something random. Now it gives a compile-time error (as does
1449        Perl).
1450    
1451    32. A *MARK encountered during the processing of a positive assertion is now
1452        recorded and passed back (compatible with Perl).
1453    
1454    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1455        had alternative anchored branches, the search for a second match in a line
1456        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1457        matched the line "0102" twice. The same bug affected patterns that started
1458        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1459        twice.
1460    
1461    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1462        does, and because of capturing effects, quantifying parenthesized
1463        assertions may at times be useful. Quantifiers are now allowed for
1464        parenthesized assertions.
1465    
1466    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1467    
1468    36. \g was being checked for fancy things in a character class, when it should
1469        just be a literal "g".
1470    
1471    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1472        appearance of a nested POSIX class supersedes an apparent external class.
1473        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1474        unescaped square brackets may also appear as part of class names. For
1475        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1476        more like Perl. (But see 8.20/1 above.)
1477    
1478    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1479        was because it thought it was \N{name}, which is not supported).
1480    
1481    39. Add minix to OS list not supporting the -S option in pcretest.
1482    
1483    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1484        cannot analyze patterns in sufficient detail to catch mutual recursions
1485        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1486        subgroup is called recursively as a subpattern for a second time at the
1487        same position in the subject string. In previous releases this might have
1488        been caught by the recursion limit, or it might have run out of stack.
1489    
1490    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1491        happen only once. PCRE was, however incorrectly giving a compile time error
1492        "recursive call could loop indefinitely" because it cannot analyze the
1493        pattern in sufficient detail. The compile time test no longer happens when
1494        PCRE is compiling a conditional subpattern, but actual runaway loops are
1495        now caught at runtime (see 40 above).
1496    
1497    42. It seems that Perl allows any characters other than a closing parenthesis
1498        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1499        has been changed to be the same.
1500    
1501    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1502        as not to get warnings when autogen.sh is called. Also changed
1503        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1504    
1505    44. To help people who use pcregrep to scan files containing exceedingly long
1506        lines, the following changes have been made:
1507    
1508        (a) The default value of the buffer size parameter has been increased from
1509            8K to 20K. (The actual buffer used is three times this size.)
1510    
1511        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1512            PCRE is built.
1513    
1514        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1515            to be set at run time.
1516    
1517        (d) Numerical values in pcregrep options can be followed by K or M, for
1518            example --buffer-size=50K.
1519    
1520        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1521            given and the return code is set to 2.
1522    
1523    45. Add a pointer to the latest mark to the callout data block.
1524    
1525    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1526        partial match of an empty string instead of no match. This was specific to
1527        the use of ".".
1528    
1529    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1530        complete match instead of a partial match. This bug was dependent on both
1531        the PCRE_UTF8 and PCRE_DOTALL options being set.
1532    
1533    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1534        starting byte set, because \b was not being ignored.
1535    
1536    
1537    Version 8.12 15-Jan-2011
1538    ------------------------
1539    
1540    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1541        checks for such things as part of the documentation building process.
1542    
1543    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1544        --match-limit and --recursion-limit options (added for 8.11). In
1545        particular, this made one of the standard tests fail. (The integer value
1546        went into the wrong half of a long int.)
1547    
1548    3.  If the --colour option was given to pcregrep with -v (invert match), it
1549        did strange things, either producing crazy output, or crashing. It should,
1550        of course, ignore a request for colour when reporting lines that do not
1551        match.
1552    
1553    4.  Another pcregrep bug caused similar problems if --colour was specified with
1554        -M (multiline) and the pattern match finished with a line ending.
1555    
1556    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1557        matched in multiline mode, the following line was shown as part of the
1558        match. This seems wrong, so I have changed it.
1559    
1560    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1561        the check for further matches in the same line (so they could be coloured)
1562        to overrun the end of the current line. If another match was found, it was
1563        incorrectly shown (and then shown again when found in the next line).
1564    
1565    7.  If pcregrep was compiled under Windows, there was a reference to the
1566        function pcregrep_exit() before it was defined. I am assuming this was
1567        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1568        reported by a user. I've moved the definition above the reference.
1569    
1570    
1571    Version 8.11 10-Dec-2010
1572    ------------------------
1573    
1574    1.  (*THEN) was not working properly if there were untried alternatives prior
1575        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1576        backtracked to try for "b" instead of moving to the next alternative branch
1577        at the same level (in this case, to look for "c"). The Perl documentation
1578        is clear that when (*THEN) is backtracked onto, it goes to the "next
1579        alternative in the innermost enclosing group".
1580    
1581    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1582        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1583        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1584        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1585        (*THEN).
1586    
1587    3.  If \s appeared in a character class, it removed the VT character from
1588        the class, even if it had been included by some previous item, for example
1589        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1590        of \s, but is part of the POSIX "space" class.)
1591    
1592    4.  A partial match never returns an empty string (because you can always
1593        match an empty string at the end of the subject); however the checking for
1594        an empty string was starting at the "start of match" point. This has been
1595        changed to the "earliest inspected character" point, because the returned
1596        data for a partial match starts at this character. This means that, for
1597        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1598        (previously it gave "no match").
1599    
1600    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1601        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1602        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1603        has an implication that the given string is incomplete (because a partial
1604        match is preferred over a full match). For this reason, these items now
1605        give a partial match in this situation. [Aside: previously, the one case
1606        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1607        match rather than a full match, which was wrong by the old rules, but is
1608        now correct.]
1609    
1610    6.  There was a bug in the handling of #-introduced comments, recognized when
1611        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1612        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1613        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1614        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1615        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1616        places in pcre_compile().
1617    
1618    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1619        comments when looking ahead for named forward references to subpatterns,
1620        the only newline sequence it recognized was NL. It now handles newlines
1621        according to the set newline convention.
1622    
1623    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1624        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1625        cater for a lack of strerror(). These oversights have been fixed.
1626    
1627    9.  Added --match-limit and --recursion-limit to pcregrep.
1628    
1629    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1630    
1631    11. When the -o option was used, pcregrep was setting a return code of 1, even
1632        when matches were found, and --line-buffered was not being honoured.
1633    
1634    12. Added an optional parentheses number to the -o and --only-matching options
1635        of pcregrep.
1636    
1637    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1638        can match an empty string. The code to do it in pcretest and pcredemo
1639        needed fixing:
1640    
1641        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1642            only one byte after an empty string match just before CRLF (this case
1643            just got forgotten; "any" and "anycrlf" were OK).
1644    
1645        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1646            mode when an empty string match preceded an ASCII character followed by
1647            a non-ASCII character. (The code for advancing by one character rather
1648            than one byte was nonsense.)
1649    
1650        (c) The pcredemo.c sample program did not have any code at all to handle
1651            the cases when CRLF is a valid newline sequence.
1652    
1653    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1654        as a starting offset was within the subject string. There is now a new
1655        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1656        negative or greater than the length of the string. In order to test this,
1657        pcretest is extended to allow the setting of negative starting offsets.
1658    
1659    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1660        starting offset points to the beginning of a UTF-8 character was
1661        unnecessarily clumsy. I tidied it up.
1662    
1663    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1664        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1665    
1666    17. Nobody had reported that the --include_dir option, which was added in
1667        release 7.7 should have been called --include-dir (hyphen, not underscore)
1668        for compatibility with GNU grep. I have changed it to --include-dir, but
1669        left --include_dir as an undocumented synonym, and the same for
1670        --exclude-dir, though that is not available in GNU grep, at least as of
1671        release 2.5.4.
1672    
1673    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1674        characters from a string of bytes) have been redefined so as not to use
1675        loops, in order to improve performance in some environments. At the same
1676        time, I abstracted some of the common code into auxiliary macros to save
1677        repetition (this should not affect the compiled code).
1678    
1679    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1680        compile-time error is now given if \c is not followed by an ASCII
1681        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1682        different, and any byte value is allowed.)
1683    
1684    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1685        START_OPTIMIZE option, which is now allowed at compile time - but just
1686        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1687        to pcregrep and other applications that have no direct access to PCRE
1688        options. The new /Y option in pcretest sets this option when calling
1689        pcre_compile().
1690    
1691    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1692        back references. Groups containing recursive back references were forced to
1693        be atomic by that change, but in the case of named groups, the amount of
1694        memory required was incorrectly computed, leading to "Failed: internal
1695        error: code overflow". This has been fixed.
1696    
1697    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1698        pcretest.c, to avoid build problems in some Borland environments.
1699    
1700    
1701    Version 8.10 25-Jun-2010
1702    ------------------------
1703    
1704    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1705        THEN.
1706    
1707    2.  (*ACCEPT) was not working when inside an atomic group.
1708    
1709    3.  Inside a character class, \B is treated as a literal by default, but
1710        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1711        causes the error). The code is unchanged, but I tidied the documentation.
1712    
1713    4.  Inside a character class, PCRE always treated \R and \X as literals,
1714        whereas Perl faults them if its -w option is set. I have changed PCRE so
1715        that it faults them when PCRE_EXTRA is set.
1716    
1717    5.  Added support for \N, which always matches any character other than
1718        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1719    
1720    6.  When compiling pcregrep with newer versions of gcc which may have
1721        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1722        declared with attribute warn_unused_result" were given. Just casting the
1723        result to (void) does not stop the warnings; a more elaborate fudge is
1724        needed. I've used a macro to implement this.
1725    
1726    7.  Minor change to pcretest.c to avoid a compiler warning.
1727    
1728    8.  Added four artifical Unicode properties to help with an option to make
1729        \s etc use properties (see next item). The new properties are: Xan
1730        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1731    
1732    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1733        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1734        this option. Modified pcretest to add /W to test this facility. Added
1735        REG_UCP to make it available via the POSIX interface.
1736    
1737    10. Added --line-buffered to pcregrep.
1738    
1739    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1740        studied, and the match started with a letter with a code point greater than
1741        127 whose first byte was different to the first byte of the other case of
1742        the letter, the other case of this starting letter was not recognized
1743        (#976).
1744    
1745    12. If a pattern that was studied started with a repeated Unicode property
1746        test, for example, \p{Nd}+, there was the theoretical possibility of
1747        setting up an incorrect bitmap of starting bytes, but fortunately it could
1748        not have actually happened in practice until change 8 above was made (it
1749        added property types that matched character-matching opcodes).
1750    
1751    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1752        possible starting bytes for non-anchored patterns.
1753    
1754    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1755        \R, and also a number of cases that involve Unicode properties, both
1756        explicit and implicit when PCRE_UCP is set.
1757    
1758    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1759        input, it could crash or give wrong results if characters with values
1760        greater than 0xc0 were present in the subject string. (Detail: it assumed
1761        UTF-8 input when processing these items.)
1762    
1763    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1764        size_t is 64-bit (#991).
1765    
1766    17. Added a check for running out of memory when PCRE is compiled with
1767        --disable-stack-for-recursion (#990).
1768    
1769    18. If the last data line in a file for pcretest does not have a newline on
1770        the end, a newline was missing in the output.
1771    
1772    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1773        less than 128) in its various bitmaps. However, there is a facility for
1774        generating tables according to the current locale when PCRE is compiled. It
1775        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1776        characters, are recognized by isspace() and therefore were getting set in
1777        these tables, and indeed these tables seem to approximate to ISO 8859. This
1778        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1779        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1780        which of course cannot start UTF-8 characters. I have changed the code so
1781        that only real ASCII characters (less than 128) and the correct starting
1782        bytes for UTF-8 encodings are set for characters greater than 127 when in
1783        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1784        altogether.)
1785    
1786    20. Added the /T option to pcretest so as to be able to run tests with non-
1787        standard character tables, thus making it possible to include the tests
1788        used for 19 above in the standard set of tests.
1789    
1790    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1791        reference to a subpattern the other side of a comment that contains an
1792        opening parenthesis caused either an internal compiling error, or a
1793        reference to the wrong subpattern.
1794    
1795    
1796    Version 8.02 19-Mar-2010
1797    ------------------------
1798    
1799    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1800    
1801    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1802        configured.
1803    
1804    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1805        original author of that file, following a query about its status.
1806    
1807    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1808        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1809    
1810    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1811        quantifier applied to a forward-referencing subroutine call, could compile
1812        incorrect code or give the error "internal error: previously-checked
1813        referenced subpattern not found".
1814    
1815    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1816        variables to point to external functions. For these systems, therefore,
1817        pcre_malloc etc. are now initialized to local functions that call the
1818        relevant global functions.
1819    
1820    7.  There were two entries missing in the vectors called coptable and poptable
1821        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1822        I've fixed the data, and added a kludgy way of testing at compile time that
1823        the lengths are correct (equal to the number of opcodes).
1824    
1825    8.  Following on from 7, I added a similar kludge to check the length of the
1826        eint vector in pcreposix.c.
1827    
1828    9.  Error texts for pcre_compile() are held as one long string to avoid too
1829        much relocation at load time. To find a text, the string is searched,
1830        counting zeros. There was no check for running off the end of the string,
1831        which could happen if a new error number was added without updating the
1832        string.
1833    
1834    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1835    
1836    11. \K was not working if it appeared in an atomic group or in a group that
1837        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1838        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1839        the assertion is positive, but not if it is negative.
1840    
1841    12. Change 11 fortuitously reduced the size of the stack frame used in the
1842        "match()" function of pcre_exec.c by one pointer. Forthcoming
1843        implementation of support for (*MARK) will need an extra pointer on the
1844        stack; I have reserved it now, so that the stack frame size does not
1845        decrease.
1846    
1847    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1848        item in branch that calls a recursion is a subroutine call - as in the
1849        second branch in the above example - was incorrectly given the compile-
1850        time error "recursive call could loop indefinitely" because pcre_compile()
1851        was not correctly checking the subroutine for matching a non-empty string.
1852    
1853    14. The checks for overrunning compiling workspace could trigger after an
1854        overrun had occurred. This is a "should never occur" error, but it can be
1855        triggered by pathological patterns such as hundreds of nested parentheses.
1856        The checks now trigger 100 bytes before the end of the workspace.
1857    
1858    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1859    
1860    
1861    Version 8.01 19-Jan-2010
1862    ------------------------
1863    
1864    1.  If a pattern contained a conditional subpattern with only one branch (in
1865        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1866        computed the wrong minimum data length (which is of course zero for such
1867        subpatterns). This could cause incorrect "no match" results.
1868    
1869    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1870        the pattern is reset in the first branch, pcre_compile() failed with
1871        "internal error: code overflow at offset...". This happened only when
1872        the reset was to the original external option setting. (An optimization
1873        abstracts leading options settings into an external setting, which was the
1874        cause of this.)
1875    
1876    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1877        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1878        assertion pattern did not match (meaning that the assertion was true), it
1879        was incorrectly treated as false if the SKIP had been reached during the
1880        matching. This also applied to assertions used as conditions.
1881    
1882    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1883        assertion subpattern, including such a pattern used as a condition,
1884        unpredictable results occurred, instead of the error return
1885        PCRE_ERROR_DFA_UITEM.
1886    
1887    5.  The C++ GlobalReplace function was not working like Perl for the special
1888        situation when an empty string is matched. It now does the fancy magic
1889        stuff that is necessary.
1890    
1891    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1892        removed. (These were left over from very, very early versions of PCRE.)
1893    
1894    7.  Some cosmetic changes to the code to make life easier when compiling it
1895        as part of something else:
1896    
1897        (a) Change DEBUG to PCRE_DEBUG.
1898    
1899        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1900            called "current" as "current_branch", to prevent a collision with the
1901            Linux macro when compiled as a kernel module.
1902    
1903        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1904            prevent a collision with the Linux macro when compiled as a kernel
1905            module.
1906    
1907    8.  In pcre_compile() there are some checks for integer overflows that used to
1908        cast potentially large values to (double). This has been changed to that
1909        when building, a check for int64_t is made, and if it is found, it is used
1910        instead, thus avoiding the use of floating point arithmetic. (There is no
1911        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1912        double.
1913    
1914    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1915        2005 (difference between two addresses compared to an unsigned value).
1916    
1917    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1918        custom one, because of the following reported problem in Windows:
1919    
1920          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1921              under Win32.
1922          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1923              therefore missing the function definition.
1924          - The compiler thus generates a "C" signature for the test function.
1925          - The linker fails to find the "C" function.
1926          - PCRE fails to configure if asked to do so against libbz2.
1927    
1928    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1929        messages were output:
1930    
1931          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1932          rerunning libtoolize, to keep the correct libtool macros in-tree.
1933          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1934    
1935        I have done both of these things.
1936    
1937    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1938        most of the time, it *can* run out if it is given a pattern that contains a
1939        runaway infinite recursion. I updated the discussion in the pcrestack man
1940        page.
1941    
1942    13. Now that we have gone to the x.xx style of version numbers, the minor
1943        version may start with zero. Using 08 or 09 is a bad idea because users
1944        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1945        interpreted as invalid octal numbers. I've updated the previous comment in
1946        configure.ac, and also added a check that gives an error if 08 or 09 are
1947        used.
1948    
1949    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1950        causing partial matching to fail when the end of the subject matched \W
1951        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1952    
1953    15. There were some discrepancies between the declarations in pcre_internal.h
1954        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1955        their definitions. The declarations used "const uschar *" and the
1956        definitions used USPTR. Even though USPTR is normally defined as "const
1957        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1958        reported that: "This difference in casting confuses some C++ compilers, for
1959        example, SunCC recognizes above declarations as different functions and
1960        generates broken code for hbpcre." I have changed the declarations to use
1961        USPTR.
1962    
1963    16. GNU libtool is named differently on some systems. The autogen.sh script now
1964        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1965        (FreeBSD).
1966    
1967    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1968        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1969        comment: "Figure out how to create a longlong from a string: strtoll and
1970        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1971        instance, but it only takes 2 args instead of 3!"
1972    
1973    18. A subtle bug concerned with back references has been fixed by a change of
1974        specification, with a corresponding code fix. A pattern such as
1975        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1976        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1977        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1978        same bug. Such groups have to be quantified to be useful, or contained
1979        inside another quantified group. (If there's no repetition, the reference
1980        can never match.) The problem arises because, having left the group and
1981        moved on to the rest of the pattern, a later failure that backtracks into
1982        the group uses the captured value from the final iteration of the group
1983        rather than the correct earlier one. I have fixed this in PCRE by forcing
1984        any group that contains a reference to itself to be an atomic group; that
1985        is, there cannot be any backtracking into it once it has completed. This is
1986        similar to recursive and subroutine calls.
1987    
1988    
1989    Version 8.00 19-Oct-09
1990    ----------------------
1991    
1992    1.  The table for translating pcre_compile() error codes into POSIX error codes
1993        was out-of-date, and there was no check on the pcre_compile() error code
1994        being within the table. This could lead to an OK return being given in
1995        error.
1996    
1997    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1998        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1999        in a Windows environment.
2000    
2001    3.  The pcregrep --count option prints the count for each file even when it is
2002        zero, as does GNU grep. However, pcregrep was also printing all files when
2003        --files-with-matches was added. Now, when both options are given, it prints
2004        counts only for those files that have at least one match. (GNU grep just
2005        prints the file name in this circumstance, but including the count seems
2006        more useful - otherwise, why use --count?) Also ensured that the
2007        combination -clh just lists non-zero counts, with no names.
2008    
2009    4.  The long form of the pcregrep -F option was incorrectly implemented as
2010        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2011        but it seems right to fix it, and I didn't think it was worth preserving
2012        the old behaviour.
2013    
2014    5.  The command line items --regex=pattern and --regexp=pattern were not
2015        recognized by pcregrep, which required --regex pattern or --regexp pattern
2016        (with a space rather than an '='). The man page documented the '=' forms,
2017        which are compatible with GNU grep; these now work.
2018    
2019    6.  No libpcreposix.pc file was created for pkg-config; there was just
2020        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2021    
2022    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2023        when UCP support is not needed, by modifying the Python script that
2024        generates it from Unicode data files. This should not matter if the module
2025        is correctly used as a library, but I received one complaint about 50K of
2026        unwanted data. My guess is that the person linked everything into his
2027        program rather than using a library. Anyway, it does no harm.
2028    
2029    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2030        was a minimum greater than 1 for a wide character in a possessive
2031        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2032        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2033        character. Chaos in the form of incorrect output or a compiling loop could
2034        result.
2035    
2036    9.  The restrictions on what a pattern can contain when partial matching is
2037        requested for pcre_exec() have been removed. All patterns can now be
2038        partially matched by this function. In addition, if there are at least two
2039        slots in the offset vector, the offset of the earliest inspected character
2040        for the match and the offset of the end of the subject are set in them when
2041        PCRE_ERROR_PARTIAL is returned.
2042    
2043    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2044        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2045        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2046        and may be more useful for multi-segment matching.
2047    
2048    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2049        used to be given if ever the end of the subject was reached; now it is
2050        given only if matching could not proceed because another character was
2051        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2052        string "Z", which now yields "no match" instead of "partial match". In the
2053        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2054        final character ended with (*FAIL).
2055    
2056    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2057        if the pattern had a "must contain" character that was already found in the
2058        earlier partial match, unless partial matching was again requested. For
2059        example, with the pattern /dog.(body)?/, the "must contain" character is
2060        "g". If the first part-match was for the string "dog", restarting with
2061        "sbody" failed. This bug has been fixed.
2062    
2063    13. The string returned by pcre_dfa_exec() after a partial match has been
2064        changed so that it starts at the first inspected character rather than the
2065        first character of the match. This makes a difference only if the pattern
2066        starts with a lookbehind assertion or \b or \B (\K is not supported by
2067        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2068        matching functions compatible, and I think it's the right thing to do.
2069    
2070    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2071        so that the demonstration program is easily available in environments where
2072        PCRE has not been installed from source.
2073    
2074    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2075        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2076        library.
2077    
2078    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2079        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2080        is not the first non-POSIX option to be added. Clearly some people find
2081        these options useful.
2082    
2083    17. If a caller to the POSIX matching function regexec() passes a non-zero
2084        value for nmatch with a NULL value for pmatch, the value of
2085        nmatch is forced to zero.
2086    
2087    18. RunGrepTest did not have a test for the availability of the -u option of
2088        the diff command, as RunTest does. It now checks in the same way as
2089        RunTest, and also checks for the -b option.
2090    
2091    19. If an odd number of negated classes containing just a single character
2092        interposed, within parentheses, between a forward reference to a named
2093        subpattern and the definition of the subpattern, compilation crashed with
2094        an internal error, complaining that it could not find the referenced
2095        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2096        [The bug was that it was starting one character too far in when skipping
2097        over the character class, thus treating the ] as data rather than
2098        terminating the class. This meant it could skip too much.]
2099    
2100    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2101        /g option in pcretest when the pattern contains \K, which makes it possible
2102        to have an empty string match not at the start, even when the pattern is
2103        anchored. Updated pcretest and pcredemo to use this option.
2104    
2105    21. If the maximum number of capturing subpatterns in a recursion was greater
2106        than the maximum at the outer level, the higher number was returned, but
2107        with unset values at the outer level. The correct (outer level) value is
2108        now given.
2109    
2110    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2111        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2112        make it do so. The string so far is captured, making this feature
2113        compatible with Perl.
2114    
2115    23. The tests have been re-organized, adding tests 11 and 12, to make it
2116        possible to check the Perl 5.10 features against Perl 5.10.
2117    
2118    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2119        pattern matches a fixed length string. PCRE did not allow this; now it
2120        does. Neither allows recursion.
2121    
2122    25. I finally figured out how to implement a request to provide the minimum
2123        length of subject string that was needed in order to match a given pattern.
2124        (It was back references and recursion that I had previously got hung up
2125        on.) This code has now been added to pcre_study(); it finds a lower bound
2126        to the length of subject needed. It is not necessarily the greatest lower
2127        bound, but using it to avoid searching strings that are too short does give
2128        some useful speed-ups. The value is available to calling programs via
2129        pcre_fullinfo().
2130    
2131    26. While implementing 25, I discovered to my embarrassment that pcretest had
2132        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2133        study optimizations had never been tested with that matching function.
2134        Oops. What is worse, even when it was passed study data, there was a bug in
2135        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2136        were also very few tests of studied patterns with pcre_dfa_exec().
2137    
2138    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2139        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2140        on the other side of the coin, they are no longer allowed to have different
2141        names, because these cannot be distinguished in PCRE, and this has caused
2142        confusion. (This is a difference from Perl.)
2143    
2144    28. When duplicate subpattern names are present (necessarily with different
2145        numbers, as required by 27 above), and a test is made by name in a
2146        conditional pattern, either for a subpattern having been matched, or for
2147        recursion in such a pattern, all the associated numbered subpatterns are
2148        tested, and the overall condition is true if the condition is true for any
2149        one of them. This is the way Perl works, and is also more like the way
2150        testing by number works.
2151    
2152    
2153    Version 7.9 11-Apr-09
2154    ---------------------
2155    
2156    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2157        (pcretest), all targets were linked against these libraries. This included
2158        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2159        libraries. This caused unwanted dependencies to be created. This problem
2160        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2161        pcretest is linked with readline.
2162    
2163    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2164        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2165        moved outside it again, because FALSE and TRUE are already defined in AIX,
2166        but BOOL is not.
2167    
2168    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2169        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2170    
2171    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2172        hyphens) following file names and line numbers when outputting matching
2173        lines. This is not true; no spaces are inserted. I have also clarified the
2174        wording for the --colour (or --color) option.
2175    
2176    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2177        was not coloured; this is different to GNU grep, so I have changed it to be
2178        the same.
2179    
2180    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2181        each matching line was coloured. Now it goes on to look for further matches
2182        of any of the test patterns, which is the same behaviour as GNU grep.
2183    
2184    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2185        doesn't make sense to accept an empty string match in pcregrep, so I have
2186        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2187        seems to be how GNU grep behaves. [But see later change 40 for release
2188        8.33.]
2189    
2190    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2191        start or after a newline", because the conditional assertion was not being
2192        correctly handled. The rule now is that both the assertion and what follows
2193        in the first alternative must satisfy the test.
2194    
2195    9.  If auto-callout was enabled in a pattern with a conditional group whose
2196        condition was an assertion, PCRE could crash during matching, both with
2197        pcre_exec() and pcre_dfa_exec().
2198    
2199    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2200        used for matching.
2201    
2202    11. Unicode property support in character classes was not working for
2203        characters (bytes) greater than 127 when not in UTF-8 mode.
2204    
2205    12. Added the -M command line option to pcretest.
2206    
2207    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2208    
2209    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2210    
2211    16. Added comments and documentation about mis-use of no_arg in the C++
2212        wrapper.
2213    
2214    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2215        from Martin Jerabek that uses macro names for all relevant character and
2216        string constants.
2217    
2218    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2219        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2220        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2221        these, but not everybody uses configure.
2222    
2223    19. A conditional group that had only one branch was not being correctly
2224        recognized as an item that could match an empty string. This meant that an
2225        enclosing group might also not be so recognized, causing infinite looping
2226        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2227        with the subject "ab", where knowledge that the repeated group can match
2228        nothing is needed in order to break the loop.
2229    
2230    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2231        exec(), but without supplying a callout function, matching went wrong.
2232    
2233    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2234        leak if the size of the offset vector was greater than 30. When the vector
2235        is smaller, the saved offsets during recursion go onto a local stack
2236        vector, but for larger vectors malloc() is used. It was failing to free
2237        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2238        error, in fact).
2239    
2240    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2241        heapframe that is used only when UTF-8 support is enabled. This caused no
2242        problem, but was untidy.
2243    
2244    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2245        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2246        included within another project.
2247    
2248    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2249        slightly modified by me:
2250    
2251          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2252              not building pcregrep.
2253    
2254          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2255              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2256    
2257    25. Forward references, both numeric and by name, in patterns that made use of
2258        duplicate group numbers, could behave incorrectly or give incorrect errors,
2259        because when scanning forward to find the reference group, PCRE was not
2260        taking into account the duplicate group numbers. A pattern such as
2261        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2262    
2263    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2264        the feature of a custom pointer more persuasive (as requested by a user).
2265    
2266    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2267        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2268        pre-defined.
2269    
2270    28. Added support for (*UTF8) at the start of a pattern.
2271    
2272    29. Arrange for flags added by the "release type" setting in CMake to be shown
2273        in the configuration summary.
2274    
2275    
2276    Version 7.8 05-Sep-08
2277    ---------------------
2278    
2279    1.  Replaced UCP searching code with optimized version as implemented for Ad
2280        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2281        stage table and inline lookup instead of a function, giving speed ups of 2
2282        to 5 times on some simple patterns that I tested. Permission was given to
2283        distribute the MultiStage2.py script that generates the tables (it's not in
2284        the tarball, but is in the Subversion repository).
2285    
2286    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2287        scripts.
2288    
2289    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2290        a group with a zero qualifier. The result of the study could be incorrect,
2291        or the function might crash, depending on the pattern.
2292    
2293    4.  Caseless matching was not working for non-ASCII characters in back
2294        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2295        It now works when Unicode Property Support is available.
2296    
2297    5.  In pcretest, an escape such as \x{de} in the data was always generating
2298        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2299        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2300        truncation.
2301    
2302    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2303    
2304    7.  Added two (int) casts to pcregrep when printing the difference of two
2305        pointers, in case they are 64-bit values.
2306    
2307    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2308        test 2 if it fails.
2309    
2310    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2311        and a #define of that name to empty if it is not externally set. This is to
2312        allow users of MSVC to set it if necessary.
2313    
2314    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2315        the convenience functions in the pcre_get.c source file.
2316    
2317    11. An option change at the start of a pattern that had top-level alternatives
2318        could cause overwriting and/or a crash. This command provoked a crash in
2319        some environments:
2320    
2321          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2322    
2323        This potential security problem was recorded as CVE-2008-2371.
2324    
2325    12. For a pattern where the match had to start at the beginning or immediately
2326        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2327        pcre_dfa_exec() could read past the end of the passed subject if there was
2328        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2329        pcretest so that it places the subject at the end of its malloc-ed buffer.
2330    
2331    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2332        exec() might read past the end of the data buffer in UTF-8 mode.
2333    
2334    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2335        the data contained the byte 0x85 as part of a UTF-8 character within its
2336        first line. This applied both to normal and DFA matching.
2337    
2338    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2339        /^[^d]*?$/8 failed to match "abc".
2340    
2341    16. Added a missing copyright notice to pcrecpp_internal.h.
2342    
2343    17. Make it more clear in the documentation that values returned from
2344        pcre_exec() in ovector are byte offsets, not character counts.
2345    
2346    18. Tidied a few places to stop certain compilers from issuing warnings.
2347    
2348    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2349        supplied by Stefan Weber. I made a further small update for 7.8 because
2350        there is a change of source arrangements: the pcre_searchfuncs.c module is
2351        replaced by pcre_ucd.c.
2352    
2353    
2354    Version 7.7 07-May-08
2355    ---------------------
2356    
2357    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2358        a string to a long long, pretend we don't even have a long long." This is
2359        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2360    
2361    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2362        pre-7.6 versions, which defined a global no_arg variable instead of putting
2363        it in the RE class. (See also #8 below.)
2364    
2365    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2366        Lopes.
2367    
2368    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2369    
2370        (1) The include/exclude patterns were being applied to the whole pathnames
2371            of files, instead of just to the final components.
2372    
2373        (2) If there was more than one level of directory, the subdirectories were
2374            skipped unless they satisfied the include/exclude conditions. This is
2375            inconsistent with GNU grep (and could even be seen as contrary to the
2376            pcregrep specification - which I improved to make it absolutely clear).
2377            The action now is always to scan all levels of directory, and just
2378            apply the include/exclude patterns to regular files.
2379    
2380    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2381        --exclude_dir in the tests to avoid scanning .svn directories.
2382    
2383    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2384        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2385        doesn't support NULs in patterns.
2386    
2387    7.  Added some missing "const"s to declarations of static tables in
2388        pcre_compile.c and pcre_dfa_exec.c.
2389    
2390    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2391        caused by fix #2  above. (Subsequently also a second patch to fix the
2392        first patch. And a third patch - this was a messy problem.)
2393    
2394    9.  Applied Craig's patch to remove the use of push_back().
2395    
2396    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2397        matching function regexec().
2398    
2399    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2400        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2401        references. PCRE supports relative numbers with this syntax (I don't think
2402        Oniguruma does).
2403    
2404    12. Previously, a group with a zero repeat such as (...){0} was completely
2405        omitted from the compiled regex. However, this means that if the group
2406        was called as a subroutine from elsewhere in the pattern, things went wrong
2407        (an internal error was given). Such groups are now left in the compiled
2408        pattern, with a new opcode that causes them to be skipped at execution
2409        time.
2410    
2411    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2412        to the way PCRE behaves:
2413    
2414        (a) A lone ] character is dis-allowed (Perl treats it as data).
2415    
2416        (b) A back reference to an unmatched subpattern matches an empty string
2417            (Perl fails the current match path).
2418    
2419        (c) A data ] in a character class must be notated as \] because if the
2420            first data character in a class is ], it defines an empty class. (In
2421            Perl it is not possible to have an empty class.) The empty class []
2422            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2423            The negative empty class [^] matches any one character, independently
2424            of the DOTALL setting.
2425    
2426    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2427        non-existent subpattern following a character class starting with ']' and
2428        containing () gave an internal compiling error instead of "reference to
2429        non-existent subpattern". Fortunately, when the pattern did exist, the
2430        compiled code was correct. (When scanning forwards to check for the
2431        existence of the subpattern, it was treating the data ']' as terminating
2432        the class, so got the count wrong. When actually compiling, the reference
2433        was subsequently set up correctly.)
2434    
2435    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2436        it was being rejected as not supported by pcre_dfa_exec(), even though
2437        other assertions are supported. I have made pcre_dfa_exec() support
2438        (*FAIL).
2439    
2440    16. The implementation of 13c above involved the invention of a new opcode,
2441        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2442        cannot be changed at match time, I realized I could make a small
2443        improvement to matching performance by compiling OP_ALLANY instead of
2444        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2445        on the OP_ANY path.
2446    
2447    17. Compiling pcretest on Windows with readline support failed without the
2448        following two fixes: (1) Make the unistd.h include conditional on
2449        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2450    
2451    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2452        ncurses library to be included for pcretest when ReadLine support is
2453        requested, but also to allow for it to be overridden. This patch came from
2454        Daniel Bergström.
2455    
2456    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2457        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2458        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2459        spotting this.
2460    
2461    
2462    Version 7.6 28-Jan-08
2463    ---------------------
2464    
2465    1.  A character class containing a very large number of characters with
2466        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2467        overflow.
2468    
2469    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2470        HAVE_LONG_LONG is not defined.
2471    
2472    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2473        bring them up to date and include new features. This patch includes:
2474    
2475        - Fixed PH's badly added libz and libbz2 support.
2476        - Fixed a problem with static linking.
2477        - Added pcredemo. [But later removed - see 7 below.]
2478        - Fixed dftables problem and added an option.
2479        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2480            HAVE_LONG_LONG.
2481        - Added readline support for pcretest.
2482        - Added an listing of the option settings after cmake has run.
2483    
2484    4.  A user submitted a patch to Makefile that makes it easy to create
2485        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2486        Makefile.am that cause it to include this special target, without
2487        affecting anything else. Note that the same mingw target plus all
2488        the other distribution libraries and programs are now supported
2489        when configuring with CMake (see 6 below) instead of with
2490        Configure/Make.
2491    
2492    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2493        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2494        exported in the Windows port". It has not yet been confirmed that the patch
2495        solves the problem, but it does no harm.
2496    
2497    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2498        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2499        with CMake, and also correct the comment about stack recursion.
2500    
2501    7.  Remove the automatic building of pcredemo from the ./configure system and
2502        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2503        of a program that users should build themselves after PCRE is installed, so
2504        building it automatically is not really right. What is more, it gave
2505        trouble in some build environments.
2506    
2507    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2508    
2509    
2510    Version 7.5 10-Jan-08
2511    ---------------------
2512    
2513    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2514        values in parens when parsing an RE using the C++ wrapper."
2515    
2516    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2517        Characters greater than 255 were excluded from the class instead of being
2518        included.
2519    
2520    3.  The same bug as (2) above applied to negated POSIX classes such as
2521        [:^space:].
2522    
2523    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2524        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2525        I have changed it.
2526    
2527    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2528        first named subpattern) and a construct such as (?&a) would reference the
2529        first named subpattern whose name started with "a" (in other words, the
2530        length check was missing). Both these problems are fixed. "Subpattern name
2531        expected" is now given for (?&) (a zero-length name), and this patch also
2532        makes it give the same error for \k'' (previously it complained that that
2533        was a reference to a non-existent subpattern).
2534    
2535    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2536        this is right because (?- can be followed by option settings as well as by
2537        digits. I have, however, made the messages clearer.
2538    
2539    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2540        than the number used in the conditional) now cause a compile-time error.
2541        This is actually not compatible with Perl, which accepts such patterns, but
2542        treats the conditional as always being FALSE (as PCRE used to), but it
2543        seems to me that giving a diagnostic is better.
2544    
2545    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2546        and messages.
2547    
2548    9.  Fix two occurrences of "backslash" in comments that should have been
2549        "backspace".
2550    
2551    10. Remove two redundant lines of code that can never be obeyed (their function
2552        was moved elsewhere).
2553    
2554    11. The program that makes PCRE's Unicode character property table had a bug
2555        which caused it to generate incorrect table entries for sequences of
2556        characters that have the same character type, but are in different scripts.
2557        It amalgamated them into a single range, with the script of the first of
2558        them. In other words, some characters were in the wrong script. There were
2559        thirteen such cases, affecting characters in the following ranges:
2560    
2561          U+002b0 - U+002c1
2562          U+0060c - U+0060d
2563          U+0061e - U+00612
2564          U+0064b - U+0065e
2565          U+0074d - U+0076d
2566          U+01800 - U+01805
2567          U+01d00 - U+01d77
2568          U+01d9b - U+01dbf
2569          U+0200b - U+0200f
2570          U+030fc - U+030fe
2571          U+03260 - U+0327f
2572          U+0fb46 - U+0fbb1
2573          U+10450 - U+1049d
2574    
2575    12. The -o option (show only the matching part of a line) for pcregrep was not
2576        compatible with GNU grep in that, if there was more than one match in a
2577        line, it showed only the first of them. It now behaves in the same way as
2578        GNU grep.
2579    
2580    13. If the -o and -v options were combined for pcregrep, it printed a blank
2581        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2582        does the same. The return code can be used to tell if there were any
2583        non-matching lines.
2584    
2585    14. Added --file-offsets and --line-offsets to pcregrep.
2586    
2587    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2588        infinitely looping recursion. The bug was that positive lookaheads were not
2589        being skipped when checking for a possible empty match (negative lookaheads
2590        and both kinds of lookbehind were skipped).
2591    
2592    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2593        inclusion of <windows.h> to before rather than after the definition of
2594        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2595    
2596    17. Specifying a possessive quantifier with a specific limit for a Unicode
2597        character property caused pcre_compile() to compile bad code, which led at
2598        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2599        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2600        caused the error; without that there was no problem.
2601    
2602    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2603    
2604    19. Added --enable-pcretest-libreadline.
2605    
2606    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2607        RE::GlobalReplace(). As a result, the number of replacements returned was
2608        double what it should be. I removed one of the increments, but Craig sent a
2609        later patch that removed the other one (the right fix) and added unit tests
2610        that check the return values (which was not done before).
2611    
2612    21. Several CMake things:
2613    
2614        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2615            the names libpcre and libpcreposix, not just pcre and pcreposix.
2616    
2617        (2) The above change means that pcretest and pcregrep are now correctly
2618            linked with the newly-built libraries, not previously installed ones.
2619    
2620        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2621    
2622    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2623        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2624        UTF-8 newline character). The key issue is that the pattern starts .*;
2625        this means that the match must be either at the beginning, or after a
2626        newline. The bug was in the code for advancing after a failed match and
2627        checking that the new position followed a newline. It was not taking
2628        account of UTF-8 characters correctly.
2629    
2630    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2631        character classes. PCRE was not treating the sequence [:...:] as a
2632        character class unless the ... were all letters. Perl, however, seems to
2633        allow any characters between [: and :], though of course it rejects as
2634        unknown any "names" that contain non-letters, because all the known class
2635        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2636        for example, whereas PCRE did not - it did not recognize a POSIX character
2637        class. This seemed a bit dangerous, so the code has been changed to be
2638        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2639        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2640        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2641        Perl does, and where it didn't before.
2642    
2643    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2644        Windows environments %n is disabled by default.
2645    
2646    
2647    Version 7.4 21-Sep-07
2648    ---------------------
2649    
2650    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2651        means that a class such as [\s] counted as "explicit reference to CR or
2652        LF". That isn't really right - the whole point of the change was to try to
2653        help when there was an actual mention of one of the two characters. So now
2654        the change happens only if \r or \n (or a literal CR or LF) character is
2655        encountered.
2656    
2657    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2658        of both had grown to the point where there were only 3 bits left.
2659        Fortunately, there was spare space in the data structure, and so I have
2660        moved the internal flags into a new 16-bit field to free up more option
2661        bits.
2662    
2663    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2664        but did not set the internal JCHANGED flag - either of these is enough to
2665        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2666        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2667        start sets both bits.
2668    
2669    4.  Added options (at build time, compile time, exec time) to change \R from
2670        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2671    
2672    5.  doc/pcresyntax.html was missing from the distribution.
2673    
2674    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2675        compatibility, even though it is no longer used.
2676    
2677    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2678        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2679        windows.h file is present (where different names are used). [This was
2680        reversed later after testing - see 16 below.]
2681    
2682    8.  Changed all #include <config.h> to #include "config.h". There were also
2683        some further <pcre.h> cases that I changed to "pcre.h".
2684    
2685    9.  When pcregrep was used with the --colour option, it missed the line ending
2686        sequence off the lines that it output.
2687    
2688    10. It was pointed out to me that arrays of string pointers cause lots of
2689        relocations when a shared library is dynamically loaded. A technique of
2690        using a single long string with a table of offsets can drastically reduce
2691        these. I have refactored PCRE in four places to do this. The result is
2692        dramatic:
2693    
2694          Originally:                          290
2695          After changing UCP table:            187
2696          After changing error message table:   43
2697          After changing table of "verbs"       36
2698          After changing table of Posix names   22
2699    
2700        Thanks to the folks working on Gregex for glib for this insight.
2701    
2702    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2703        unicode-properties was also set.
2704    
2705    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2706    
2707    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2708        checked only for CRLF.
2709    
2710    14. Added casts to pcretest.c to avoid compiler warnings.
2711    
2712    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2713    
2714    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2715        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2716        entirely. This removes changes made in 7 above.
2717    
2718    17. The CMake files have been updated, and there is now more information about
2719        building with CMake in the NON-UNIX-USE document.
2720    
2721    
2722    Version 7.3 28-Aug-07
2723    ---------------------
2724    
2725     1. In the rejigging of the build system that eventually resulted in 7.1, the
2726        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2727        brackets there is not right, since it causes compilers to look for an
2728        installed pcre.h, not the version that is in the source that is being
2729        compiled (which of course may be different). I have changed it back to:
2730    
2731          #include "pcre.h"
2732    
2733        I have a vague recollection that the change was concerned with compiling in
2734        different directories, but in the new build system, that is taken care of
2735        by the VPATH setting the Makefile.
2736    
2737     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2738        when the subject happened to end in the byte 0x85 (e.g. if the last
2739        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2740        characters but of course it shouldn't be taken as a newline when it is part
2741        of another character. The bug was that, for an unlimited repeat of . in
2742        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2743        characters when looking for a newline.
2744    
2745     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2746    
2747     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2748        in debug output.
2749    
2750     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2751        long printing in the pcrecpp unittest when running under MinGW.
2752    
2753     6. ESC_K was left out of the EBCDIC table.
2754    
2755     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2756        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2757        limit also applies to "virtual nesting" when a pattern is recursive, and in
2758        this case 1000 isn't so big. I have been able to remove this limit at the
2759        expense of backing off one optimization in certain circumstances. Normally,
2760        when pcre_exec() would call its internal match() function recursively and
2761        immediately return the result unconditionally, it uses a "tail recursion"
2762        feature to save stack. However, when a subpattern that can match an empty
2763        string has an unlimited repetition quantifier, it no longer makes this
2764        optimization. That gives it a stack frame in which to save the data for
2765        checking that an empty string has been matched. Previously this was taken
2766        from the 1000-entry workspace that had been reserved. So now there is no
2767        explicit limit, but more stack is used.
2768    
2769     8. Applied Daniel's patches to solve problems with the import/export magic
2770        syntax that is required for Windows, and which was going wrong for the
2771        pcreposix and pcrecpp parts of the library. These were overlooked when this
2772        problem was solved for the main library.
2773    
2774     9. There were some crude static tests to avoid integer overflow when computing
2775        the size of patterns that contain repeated groups with explicit upper
2776        limits. As the maximum quantifier is 65535, the maximum group length was
2777        set at 30,000 so that the product of these two numbers did not overflow a
2778        32-bit integer. However, it turns out that people want to use groups that
2779        are longer than 30,000 bytes (though not repeat them that many times).
2780        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2781        made it possible to implement the integer overflow checks in a much more
2782        dynamic way, which I have now done. The artificial limitation on group
2783        length has been removed - we now have only the limit on the total length of
2784        the compiled pattern, which depends on the LINK_SIZE setting.
2785    
2786    10. Fixed a bug in the documentation for get/copy named substring when
2787        duplicate names are permitted. If none of the named substrings are set, the
2788        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2789        empty string.
2790    
2791    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2792        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2793        because the ] is interpreted as the first data character and the
2794        terminating ] is not found. PCRE has been made compatible with Perl in this
2795        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2796        cause memory overwriting.
2797    
2798    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2799        string has been matched (to stop an infinite loop). It was not recognizing
2800        a conditional subpattern that could match an empty string if that
2801        subpattern was within another subpattern. For example, it looped when
2802        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2803        condition was not nested. This bug has been fixed.
2804    
2805    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2806        past the start of the subject in the presence of bytes with the top bit
2807        set, for example "\x8aBCD".
2808    
2809    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2810        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2811    
2812    14. Optimized (?!) to (*FAIL).
2813    
2814    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2815        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2816        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2817        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2818        does: it's just the validity check that is more restrictive.
2819    
2820    16. Inserted checks for integer overflows during escape sequence (backslash)
2821        processing, and also fixed erroneous offset values for syntax errors during
2822        backslash processing.
2823    
2824    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2825        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2826    
2827    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2828        caused an overrun.
2829    
2830    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2831        something other than just ASCII characters) inside a group that had an
2832        unlimited repeat caused a loop at compile time (while checking to see
2833        whether the group could match an empty string).
2834    
2835    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2836        [\P{Any}] did so. (Error in the code for printing property names.)
2837    
2838    21. An orphan \E inside a character class could cause a crash.
2839    
2840    22. A repeated capturing bracket such as (A)? could cause a wild memory
2841        reference during compilation.
2842    
2843    23. There are several functions in pcre_compile() that scan along a compiled
2844        expression for various reasons (e.g. to see if it's fixed length for look
2845        behind). There were bugs in these functions when a repeated \p or \P was
2846        present in the pattern. These operators have additional parameters compared
2847        with \d, etc, and these were not being taken into account when moving along
2848        the compiled data. Specifically:
2849    
2850        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2851            length.
2852    
2853        (b) An item such as \pL+ within a repeated group could cause crashes or
2854            loops.
2855    
2856        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2857            "reference to non-existent subpattern" error.
2858    
2859        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2860    
2861    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2862        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2863    
2864    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2865    
2866    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2867        character were causing crashes (broken optimization).
2868    
2869    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2870        \p or \P) caused a compile-time loop.
2871    
2872    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2873        break. For example, the unstudied pattern [\r\n]A does not match the string
2874        "\r\nA" because change 7.0/46 below moves the current point on by two
2875        characters after failing to match at the start. However, the pattern \nA
2876        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2877        the same is true. There doesn't seem any very clean way out of this, but
2878        what I have chosen to do makes the common cases work: PCRE now takes note
2879        of whether there can be an explicit match for \r or \n anywhere in the
2880        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2881        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2882        pattern has explicit CR or LF references.
2883    
2884    29. Added (*CR) etc for changing newline setting at start of pattern.
2885    
2886    
2887    Version 7.2 19-Jun-07
2888  ---------------------  ---------------------
2889    
2890   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 69  Version 7.2 13-June-07 Line 2952  Version 7.2 13-June-07
2952        pcrecpp::RE("a*?").FullMatch("aaa") does not, and        pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2953        pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.        pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2954    
2955    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2956        it matched the wrong number of bytes.
2957    
2958    
2959  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2960  ---------------------  ---------------------

Legend:
Removed from v.182  
changed lines
  Added in v.1528

  ViewVC Help
Powered by ViewVC 1.1.5