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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1532 by ph10, Fri Mar 6 11:59:39 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  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    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
87        call within a group that also contained a recursive back reference caused
88        incorrect code to be compiled. This bug was reported as "heap overflow",
89        discovered by Kai Lu of Fortinet's FortiGuard Labs.
90    
91    24. Computing the size of the JIT read-only data in advance has been a source
92        of various issues, and new ones are still appear unfortunately. To fix
93        existing and future issues, size computation is eliminated from the code,
94        and replaced by on-demand memory allocation.
95    
96    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
97        adjacent to the end of the range, and the range contained characters with
98        more than one other case, caused incorrect behaviour when compiled in UTF
99        mode. In that example, the range a-j was left out of the class.
100    
101    
102    Version 8.36 26-September-2014
103    ------------------------------
104    
105    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
106        -Wmissing-field-initializers and -Wunused-parameter.
107    
108    2.  The tests for quantifiers being too big (greater than 65535) were being
109        applied after reading the number, and stupidly assuming that integer
110        overflow would give a negative number. The tests are now applied as the
111        numbers are read.
112    
113    3.  Tidy code in pcre_exec.c where two branches that used to be different are
114        now the same.
115    
116    4.  The JIT compiler did not generate match limit checks for certain
117        bracketed expressions with quantifiers. This may lead to exponential
118        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
119        issue should be resolved now.
120    
121    5.  Fixed an issue, which occures when nested alternatives are optimized
122        with table jumps.
123    
124    6.  Inserted two casts and changed some ints to size_t in the light of some
125        reported 64-bit compiler warnings (Bugzilla 1477).
126    
127    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
128        an empty string, which sometimes were behaving incorrectly in the
129        interpreter (though correctly in the JIT matcher). This pcretest input is
130        an example:
131    
132          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
133          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
134    
135        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
136        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
137        for an empty string was breaking the inner loop and carrying on at a lower
138        level, when possessive repeated groups should always return to a higher
139        level as they have no backtrack points in them. The empty string test now
140        occurs at the outer level.
141    
142    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
143        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
144    
145    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
146        doesn't).
147    
148    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
149        to hold a pattern to be incorrectly computed (too small) when there were
150        named back references to duplicated names. This could cause "internal
151        error: code overflow" or "double free or corruption" or other memory
152        handling errors.
153    
154    11. When named subpatterns had the same prefixes, back references could be
155        confused. For example, in this pattern:
156    
157          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
158    
159        the reference to 'Name' was incorrectly treated as a reference to a
160        duplicate name.
161    
162    12. A pattern such as /^s?c/mi8 where the optional character has more than
163        one "other case" was incorrectly compiled such that it would only try to
164        match starting at "c".
165    
166    13. When a pattern starting with \s was studied, VT was not included in the
167        list of possible starting characters; this should have been part of the
168        8.34/18 patch.
169    
170    14. If a character class started [\Qx]... where x is any character, the class
171        was incorrectly terminated at the ].
172    
173    15. If a pattern that started with a caseless match for a character with more
174        than one "other case" was studied, PCRE did not set up the starting code
175        unit bit map for the list of possible characters. Now it does. This is an
176        optimization improvement, not a bug fix.
177    
178    16. The Unicode data tables have been updated to Unicode 7.0.0.
179    
180    17. Fixed a number of memory leaks in pcregrep.
181    
182    18. Avoid a compiler warning (from some compilers) for a function call with
183        a cast that removes "const" from an lvalue by using an intermediate
184        variable (to which the compiler does not object).
185    
186    19. Incorrect code was compiled if a group that contained an internal recursive
187        back reference was optional (had quantifier with a minimum of zero). This
188        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
189        caused segmentation faults because of stack overflows at compile time.
190    
191    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
192        group that is quantified with an indefinite repeat, caused a compile-time
193        loop which used up all the system stack and provoked a segmentation fault.
194        This was not the same bug as 19 above.
195    
196    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
197        Patch by Mike Frysinger.
198    
199    
200    Version 8.35 04-April-2014
201    --------------------------
202    
203    1.  A new flag is set, when property checks are present in an XCLASS.
204        When this flag is not set, PCRE can perform certain optimizations
205        such as studying these XCLASS-es.
206    
207    2.  The auto-possessification of character sets were improved: a normal
208        and an extended character set can be compared now. Furthermore
209        the JIT compiler optimizes more character set checks.
210    
211    3.  Got rid of some compiler warnings for potentially uninitialized variables
212        that show up only when compiled with -O2.
213    
214    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
215        of a match later then the end of the match. The pcretest program was not
216        handling the case sensibly - it was outputting from the start to the next
217        binary zero. It now reports this situation in a message, and outputs the
218        text from the end to the start.
219    
220    5.  Fast forward search is improved in JIT. Instead of the first three
221        characters, any three characters with fixed position can be searched.
222        Search order: first, last, middle.
223    
224    6.  Improve character range checks in JIT. Characters are read by an inprecise
225        function now, which returns with an unknown value if the character code is
226        above a certain threshold (e.g: 256). The only limitation is that the value
227        must be bigger than the threshold as well. This function is useful when
228        the characters above the threshold are handled in the same way.
229    
230    7.  The macros whose names start with RAWUCHAR are placeholders for a future
231        mode in which only the bottom 21 bits of 32-bit data items are used. To
232        make this more memorable for those maintaining the code, the names have
233        been changed to start with UCHAR21, and an extensive comment has been added
234        to their definition.
235    
236    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
237        to the export list in Makefile.am (they were accidentally omitted from the
238        8.34 tarball).
239    
240    9.  The informational output from pcretest used the phrase "starting byte set"
241        which is inappropriate for the 16-bit and 32-bit libraries. As the output
242        for "first char" and "need char" really means "non-UTF-char", I've changed
243        "byte" to "char", and slightly reworded the output. The documentation about
244        these values has also been (I hope) clarified.
245    
246    10. Another JIT related optimization: use table jumps for selecting the correct
247        backtracking path, when more than four alternatives are present inside a
248        bracket.
249    
250    11. Empty match is not possible, when the minimum length is greater than zero,
251        and there is no \K in the pattern. JIT should avoid empty match checks in
252        such cases.
253    
254    12. In a caseless character class with UCP support, when a character with more
255        than one alternative case was not the first character of a range, not all
256        the alternative cases were added to the class. For example, s and \x{17f}
257        are both alternative cases for S: the class [RST] was handled correctly,
258        but [R-T] was not.
259    
260    13. The configure.ac file always checked for pthread support when JIT was
261        enabled. This is not used in Windows, so I have put this test inside a
262        check for the presence of windows.h (which was already tested for).
263    
264    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
265        The algorithm provides a way to skip certain starting offsets, and usually
266        faster than linear prefix searches.
267    
268    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
269        as for 'fr_FR' and 'french'. For some reason, however, it then used the
270        Windows-specific input and output files, which have 'french' screwed in.
271        So this could never have worked. One of the problems with locales is that
272        they aren't always the same. I have now updated RunTest so that it checks
273        the output of the locale test (test 3) against three different output
274        files, and it allows the test to pass if any one of them matches. With luck
275        this should make the test pass on some versions of Solaris where it was
276        failing. Because of the uncertainty, the script did not used to stop if
277        test 3 failed; it now does. If further versions of a French locale ever
278        come to light, they can now easily be added.
279    
280    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
281        there was a message during ./configure, but it did not stop. This now
282        provokes an error. The invalid example in README has been corrected.
283        If a value less than the minimum is given, the minimum value has always
284        been used, but now a warning is given.
285    
286    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
287        was a bug in the test system, which is now fixed. Also, the list of various
288        configurations that are tested for each release did not have one with both
289        16/32 bits and --enable-bar-anycrlf. It now does.
290    
291    18. pcretest was missing "-C bsr" for displaying the \R default setting.
292    
293    19. Little endian PowerPC systems are supported now by the JIT compiler.
294    
295    20. The fast forward newline mechanism could enter to an infinite loop on
296        certain invalid UTF-8 input. Although we don't support these cases
297        this issue can be fixed by a performance optimization.
298    
299    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
300        not take account if existing stack usage. There is now a new global
301        variable called pcre_stack_guard that can be set to point to an external
302        function to check stack availability. It is called at the start of
303        processing every parenthesized group.
304    
305    22. A typo in the code meant that in ungreedy mode the max/min qualifier
306        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
307        not match "ab".
308    
309    23. When UTF was disabled, the JIT program reported some incorrect compile
310        errors. These messages are silenced now.
311    
312    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
313        compiler.
314    
315    25. Change all the temporary files used in RunGrepTest to be different to those
316        used by RunTest so that the tests can be run simultaneously, for example by
317        "make -j check".
318    
319    
320    Version 8.34 15-December-2013
321    -----------------------------
322    
323    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
324        executable memory. Patch inspired by Carsten Klein.
325    
326    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
327        this macro is never tested and has no effect, because the work to support
328        coverage involves only compiling and linking options and special targets in
329        the Makefile. The comment in config.h implied that defining the macro would
330        enable coverage support, which is totally false. There was also support for
331        setting this macro in the CMake files (my fault, I just copied it from
332        configure). SUPPORT_GCOV has now been removed.
333    
334    3.  Make a small performance improvement in strlen16() and strlen32() in
335        pcretest.
336    
337    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
338        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
339        these cases). There was also one in pcretest.c.
340    
341    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
342    
343    6.  In UTF mode, the code for checking whether a group could match an empty
344        string (which is used for indefinitely repeated groups to allow for
345        breaking an infinite loop) was broken when the group contained a repeated
346        negated single-character class with a character that occupied more than one
347        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
348        UTF-8 mode). The effect was undefined: the group might or might not be
349        deemed as matching an empty string, or the program might have crashed.
350    
351    7.  The code for checking whether a group could match an empty string was not
352        recognizing that \h, \H, \v, \V, and \R must match a character.
353    
354    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
355        an empty string. If it can, pcretest shows this in its information output.
356    
357    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
358        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
359        matched by pcre_exec() without using JIT:
360    
361        (a) If the rest of the pattern did not match after a maximal run of
362            grapheme clusters, the code for backing up to try with fewer of them
363            did not always back up over a full grapheme when characters that do not
364            have the modifier quality were involved, e.g. Hangul syllables.
365    
366        (b) If the match point in a subject started with modifier character, and
367            there was no match, the code could incorrectly back up beyond the match
368            point, and potentially beyond the first character in the subject,
369            leading to a segfault or an incorrect match result.
370    
371    10. A conditional group with an assertion condition could lead to PCRE
372        recording an incorrect first data item for a match if no other first data
373        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
374        first data item, and therefore matched "ca" after "c" instead of at the
375        start.
376    
377    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
378        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
379    
380    12. The source of pcregrep now includes z/OS-specific code so that it can be
381        compiled for z/OS as part of the special z/OS distribution.
382    
383    13. Added the -T and -TM options to pcretest.
384    
385    14. The code in pcre_compile.c for creating the table of named capturing groups
386        has been refactored. Instead of creating the table dynamically during the
387        actual compiling pass, the information is remembered during the pre-compile
388        pass (on the stack unless there are more than 20 named groups, in which
389        case malloc() is used) and the whole table is created before the actual
390        compile happens. This has simplified the code (it is now nearly 150 lines
391        shorter) and prepared the way for better handling of references to groups
392        with duplicate names.
393    
394    15. A back reference to a named subpattern when there is more than one of the
395        same name now checks them in the order in which they appear in the pattern.
396        The first one that is set is used for the reference. Previously only the
397        first one was inspected. This change makes PCRE more compatible with Perl.
398    
399    16. Unicode character properties were updated from Unicode 6.3.0.
400    
401    17. The compile-time code for auto-possessification has been refactored, based
402        on a patch by Zoltan Herczeg. It now happens after instead of during
403        compilation. The code is cleaner, and more cases are handled. The option
404        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
405        options in pcretest are provided to set it. It can also be set by
406        (*NO_AUTO_POSSESS) at the start of a pattern.
407    
408    18. The character VT has been added to the default ("C" locale) set of
409        characters that match \s and are generally treated as white space,
410        following this same change in Perl 5.18. There is now no difference between
411        "Perl space" and "POSIX space". Whether VT is treated as white space in
412        other locales depends on the locale.
413    
414    19. The code for checking named groups as conditions, either for being set or
415        for being recursed, has been refactored (this is related to 14 and 15
416        above). Processing unduplicated named groups should now be as fast at
417        numerical groups, and processing duplicated groups should be faster than
418        before.
419    
420    20. Two patches to the CMake build system, by Alexander Barkov:
421    
422          (1) Replace the "source" command by "." in CMakeLists.txt because
423              "source" is a bash-ism.
424    
425          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
426              without these the CMake build does not work on Solaris.
427    
428    21. Perl has changed its handling of \8 and \9. If there is no previously
429        encountered capturing group of those numbers, they are treated as the
430        literal characters 8 and 9 instead of a binary zero followed by the
431        literals. PCRE now does the same.
432    
433    22. Following Perl, added \o{} to specify codepoints in octal, making it
434        possible to specify values greater than 0777 and also making them
435        unambiguous.
436    
437    23. Perl now gives an error for missing closing braces after \x{... instead of
438        treating the string as literal. PCRE now does the same.
439    
440    24. RunTest used to grumble if an inappropriate test was selected explicitly,
441        but just skip it when running all tests. This make it awkward to run ranges
442        of tests when one of them was inappropriate. Now it just skips any
443        inappropriate tests, as it always did when running all tests.
444    
445    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
446        character types such as \d or \w, too many callouts were inserted, and the
447        data that they returned was rubbish.
448    
449    26. In UCP mode, \s was not matching two of the characters that Perl matches,
450        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
451        were matched by \h. The code has now been refactored so that the lists of
452        the horizontal and vertical whitespace characters used for \h and \v (which
453        are defined only in one place) are now also used for \s.
454    
455    27. Add JIT support for the 64 bit TileGX architecture.
456        Patch by Jiong Wang (Tilera Corporation).
457    
458    28. Possessive quantifiers for classes (both explicit and automatically
459        generated) now use special opcodes instead of wrapping in ONCE brackets.
460    
461    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
462        (because it's meaningless), this was not happening when PCRE_CASELESS was
463        set. Not wrong, but inefficient.
464    
465    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
466        properties for \w, \d, etc) is present in a test regex. Otherwise if the
467        test contains no characters greater than 255, Perl doesn't realise it
468        should be using Unicode semantics.
469    
470    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
471        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
472        does in Unicode mode.
473    
474    32. Added the "forbid" facility to pcretest so that putting tests into the
475        wrong test files can sometimes be quickly detected.
476    
477    33. There is now a limit (default 250) on the depth of nesting of parentheses.
478        This limit is imposed to control the amount of system stack used at compile
479        time. It can be changed at build time by --with-parens-nest-limit=xxx or
480        the equivalent in CMake.
481    
482    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
483        errors. Perl warns for these when in warning mode, but PCRE has no facility
484        for giving warnings.
485    
486    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
487        However, this was not working for (?!) because it is optimized to (*FAIL),
488        for which PCRE does not allow quantifiers. The optimization is now disabled
489        when a quantifier follows (?!). I can't see any use for this, but it makes
490        things uniform.
491    
492    36. Perl no longer allows group names to start with digits, so I have made this
493        change also in PCRE. It simplifies the code a bit.
494    
495    37. In extended mode, Perl ignores spaces before a + that indicates a
496        possessive quantifier. PCRE allowed a space before the quantifier, but not
497        before the possessive +. It now does.
498    
499    38. The use of \K (reset reported match start) within a repeated possessive
500        group such as (a\Kb)*+ was not working.
501    
502    40. Document that the same character tables must be used at compile time and
503        run time, and that the facility to pass tables to pcre_exec() and
504        pcre_dfa_exec() is for use only with saved/restored patterns.
505    
506    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
507        features for Builds with MSVC:
508    
509        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
510           builds). The .rc files can be used to set FileDescription and many other
511           attributes.
512    
513        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
514           This allows higher-level build scripts which want .pdb files to avoid
515           hard-coding the exact files needed."
516    
517    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
518        mean "start of word" and "end of word", respectively, as a transition aid.
519    
520    43. A minimizing repeat of a class containing codepoints greater than 255 in
521        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
522        compiled to use the heap for recursion.
523    
524    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
525        is configured.
526    
527    
528    Version 8.33 28-May-2013
529    ------------------------
530    
531    1.  Added 'U' to some constants that are compared to unsigned integers, to
532        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
533        variables that are added to signed variables, to ensure the result is
534        signed and can be negated.
535    
536    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
537        CMake config files.
538    
539    3.  Revise the creation of config.h.generic so that all boolean macros are
540        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
541        overriding via -D on the command line possible.
542    
543    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
544        to unsigned int is reported to make a quite noticeable speed difference in
545        a specific Windows environment. Testing on Linux did also appear to show
546        some benefit (and it is clearly not harmful). Also fixed the definition of
547        Xop which should be unsigned.
548    
549    5.  Related to (4), changing the definition of the intermediate variable cc
550        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
551        improvements.
552    
553    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
554        unnecessary spaces.
555    
556    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
557        and later.
558    
559    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
560    
561    9.  Optimizing fast_forward_start_bits in JIT.
562    
563    10. Adding support for callouts in JIT, and fixing some issues revealed
564        during this work. Namely:
565    
566        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
567    
568        (b) Minimum length was not checked before the matching is started.
569    
570    11. The value of capture_last that is passed to callouts was incorrect in some
571        cases when there was a capture on one path that was subsequently abandoned
572        after a backtrack. Also, the capture_last value is now reset after a
573        recursion, since all captures are also reset in this case.
574    
575    12. The interpreter no longer returns the "too many substrings" error in the
576        case when an overflowing capture is in a branch that is subsequently
577        abandoned after a backtrack.
578    
579    13. In the pathological case when an offset vector of size 2 is used, pcretest
580        now prints out the matched string after a yield of 0 or 1.
581    
582    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
583        Only supported by the JIT compiler at the moment.
584    
585    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
586    
587    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
588        offset of the starting point of the matching process, provided the offsets
589        vector is large enough.
590    
591    17. The \A escape now records a lookbehind value of 1, though its execution
592        does not actually inspect the previous character. This is to ensure that,
593        in partial multi-segment matching, at least one character from the old
594        segment is retained when a new segment is processed. Otherwise, if there
595        are no lookbehinds in the pattern, \A might match incorrectly at the start
596        of a new segment.
597    
598    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
599    
600    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
601        gives some modest performance improvement in 8-bit mode.
602    
603    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
604        be expressed in certain programming languages using Universal Character
605        Names.
606    
607    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
608        which points out that "non characters" are not "characters that may not
609        appear in Unicode strings" but rather "characters that are reserved for
610        internal use and have only local meaning".
611    
612    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
613        there was a conditional group that depended on an assertion, if the
614        assertion was false, the callout that immediately followed the alternation
615        in the condition was skipped when pcre_exec() was used for matching.
616    
617    23. Allow an explicit callout to be inserted before an assertion that is the
618        condition for a conditional group, for compatibility with automatic
619        callouts, which always insert a callout at this point.
620    
621    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
622        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
623    
624    25. (*PRUNE) is now supported by the JIT compiler.
625    
626    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
627    
628    27. Fix the case where there are two or more SKIPs with arguments that may be
629        ignored.
630    
631    28. (*SKIP) is now supported by the JIT compiler.
632    
633    29. (*THEN) is now supported by the JIT compiler.
634    
635    30. Update RunTest with additional test selector options.
636    
637    31. The way PCRE handles backtracking verbs has been changed in two ways.
638    
639        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
640        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
641        backtracking. In some cases this makes it more Perl-compatible, but Perl's
642        rather obscure rules do not always do the same thing.
643    
644        (2) Previously, backtracking verbs were confined within assertions. This is
645        no longer the case for positive assertions, except for (*ACCEPT). Again,
646        this sometimes improves Perl compatibility, and sometimes does not.
647    
648    32. A number of tests that were in test 2 because Perl did things differently
649        have been moved to test 1, because either Perl or PCRE has changed, and
650        these tests are now compatible.
651    
652    32. Backtracking control verbs are now handled in the same way in JIT and
653        interpreter.
654    
655    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
656        contained a forward subroutine reference caused a compile error.
657    
658    34. Auto-detect and optimize limited repetitions in JIT.
659    
660    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
661        blocking (*UTF) etc.
662    
663    36. In the interpreter, maximizing pattern repetitions for characters and
664        character types now use tail recursion, which reduces stack usage.
665    
666    37. The value of the max lookbehind was not correctly preserved if a compiled
667        and saved regex was reloaded on a host of different endianness.
668    
669    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
670        of the compiled pattern block, expand the flags field from 16 to 32 bits
671        because it was almost full.
672    
673    39. Try madvise first before posix_madvise.
674    
675    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
676        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
677        original change locked out all matches of empty strings. This has been
678        changed so that one match of an empty string per line is recognized.
679        Subsequent searches on the same line (for colouring or for --only-matching,
680        for example) do not recognize empty strings.
681    
682    41. Applied a user patch to fix a number of spelling mistakes in comments.
683    
684    42. Data lines longer than 65536 caused pcretest to crash.
685    
686    43. Clarified the data type for length and startoffset arguments for pcre_exec
687        and pcre_dfa_exec in the function-specific man pages, where they were
688        explicitly stated to be in bytes, never having been updated. I also added
689        some clarification to the pcreapi man page.
690    
691    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
692        a segmentation fault.
693    
694    
695    Version 8.32 30-November-2012
696    -----------------------------
697    
698    1.  Improved JIT compiler optimizations for first character search and single
699        character iterators.
700    
701    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
702        Patch by Daniel Richard G.
703    
704    3.  Single character iterator optimizations in the JIT compiler.
705    
706    4.  Improved JIT compiler optimizations for character ranges.
707    
708    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
709        Reported by Giuseppe D'Angelo.
710    
711    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
712        of a line, was being set incorrectly in cases where .* appeared inside
713        atomic brackets at the start of a pattern, or where there was a subsequent
714        *PRUNE or *SKIP.
715    
716    7.  Improved instruction cache flush for POWER/PowerPC.
717        Patch by Daniel Richard G.
718    
719    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
720        grep:
721    
722        (a) There is now no limit to the number of patterns to be matched.
723    
724        (b) An error is given if a pattern is too long.
725    
726        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
727            are now supported.
728    
729        (d) --exclude-from and --include-from (multiple use) have been added.
730    
731        (e) Exclusions and inclusions now apply to all files and directories, not
732            just to those obtained from scanning a directory recursively.
733    
734        (f) Multiple uses of -f and --file-list are now supported.
735    
736        (g) In a Windows environment, the default for -d has been changed from
737            "read" (the GNU grep default) to "skip", because otherwise the presence
738            of a directory in the file list provokes an error.
739    
740        (h) The documentation has been revised and clarified in places.
741    
742    9.  Improve the matching speed of capturing brackets.
743    
744    10. Changed the meaning of \X so that it now matches a Unicode extended
745        grapheme cluster.
746    
747    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
748        out POSIX threads when JIT support is configured.
749    
750    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
751    
752    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
753        structure could go wrong in environments where size_t is not the same size
754        as int.
755    
756    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
757        set.
758    
759    15. The EBCDIC support had decayed; later updates to the code had included
760        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
761        general tidy up of EBCDIC-related issues, and the documentation was also
762        not quite right. There is now a test that can be run on ASCII systems to
763        check some of the EBCDIC-related things (but is it not a full test).
764    
765    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
766        in a small tidy to the code.
767    
768    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
769    
770    18. If the --only-matching (-o) option in pcregrep is specified multiple
771        times, each one causes appropriate output. For example, -o1 -o2 outputs the
772        substrings matched by the 1st and 2nd capturing parentheses. A separating
773        string can be specified by --om-separator (default empty).
774    
775    19. Improving the first n character searches.
776    
777    20. Turn case lists for horizontal and vertical white space into macros so that
778        they are defined only once.
779    
780    21. This set of changes together give more compatible Unicode case-folding
781        behaviour for characters that have more than one other case when UCP
782        support is available.
783    
784        (a) The Unicode property table now has offsets into a new table of sets of
785            three or more characters that are case-equivalent. The MultiStage2.py
786            script that generates these tables (the pcre_ucd.c file) now scans
787            CaseFolding.txt instead of UnicodeData.txt for character case
788            information.
789    
790        (b) The code for adding characters or ranges of characters to a character
791            class has been abstracted into a generalized function that also handles
792            case-independence. In UTF-mode with UCP support, this uses the new data
793            to handle characters with more than one other case.
794    
795        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
796            whose other case is greater than 256 are now correctly matched
797            caselessly. Previously, the high codepoint matched the low one, but not
798            vice versa.
799    
800        (d) The processing of \h, \H, \v, and \ in character classes now makes use
801            of the new class addition function, using character lists defined as
802            macros alongside the case definitions of 20 above.
803    
804        (e) Caseless back references now work with characters that have more than
805            one other case.
806    
807        (f) General caseless matching of characters with more than one other case
808            is supported.
809    
810    22. Unicode character properties were updated from Unicode 6.2.0
811    
812    23. Improved CMake support under Windows. Patch by Daniel Richard G.
813    
814    24. Add support for 32-bit character strings, and UTF-32
815    
816    25. Major JIT compiler update (code refactoring and bugfixing).
817        Experimental Sparc 32 support is added.
818    
819    26. Applied a modified version of Daniel Richard G's patch to create
820        pcre.h.generic and config.h.generic by "make" instead of in the
821        PrepareRelease script.
822    
823    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
824        pcre_compile.c when checking for a zero character.
825    
826    28. Introducing a native interface for JIT. Through this interface, the compiled
827        machine code can be directly executed. The purpose of this interface is to
828        provide fast pattern matching, so several sanity checks are not performed.
829        However, feature tests are still performed. The new interface provides
830        1.4x speedup compared to the old one.
831    
832    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
833        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
834        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
835    
836    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
837        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
838        the "old" RFC 2279). Instead, it ended up passing a negative length to
839        pcre_exec().
840    
841    31. Add support for GCC's visibility feature to hide internal functions.
842    
843    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
844        "unknown -C option" after outputting 0 or 1.
845    
846    33. There is now support for generating a code coverage report for the test
847        suite in environments where gcc is the compiler and lcov is installed. This
848        is mainly for the benefit of the developers.
849    
850    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
851        unaddressable using valgrind annotations, allowing valgrind to detect
852        invalid memory accesses. This is mainly for the benefit of the developers.
853    
854    25. (*UTF) can now be used to start a pattern in any of the three libraries.
855    
856    26. Give configure error if --enable-cpp but no C++ compiler found.
857    
858    
859    Version 8.31 06-July-2012
860    -------------------------
861    
862    1.  Fixing a wrong JIT test case and some compiler warnings.
863    
864    2.  Removed a bashism from the RunTest script.
865    
866    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
867        to unsigned type, result still unsigned" that was given by an MS compiler
868        on encountering the code "-sizeof(xxx)".
869    
870    4.  Partial matching support is added to the JIT compiler.
871    
872    5.  Fixed several bugs concerned with partial matching of items that consist
873        of more than one character:
874    
875        (a) /^(..)\1/ did not partially match "aba" because checking references was
876            done on an "all or nothing" basis. This also applied to repeated
877            references.
878    
879        (b) \R did not give a hard partial match if \r was found at the end of the
880            subject.
881    
882        (c) \X did not give a hard partial match after matching one or more
883            characters at the end of the subject.
884    
885        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
886            a partial match for the string "\r".
887    
888        (e) When newline was set to CRLF, the metacharacter "." did not recognize
889            a partial match for a CR character at the end of the subject string.
890    
891    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
892        running pcretest, the text "(JIT)" added to the output whenever JIT is
893        actually used to run the match.
894    
895    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
896        or /S+[+] with a digit between 1 and 7.
897    
898    8.  OP_NOT now supports any UTF character not just single-byte ones.
899    
900    9.  (*MARK) control verb is now supported by the JIT compiler.
901    
902    10. The command "./RunTest list" lists the available tests without actually
903        running any of them. (Because I keep forgetting what they all are.)
904    
905    11. Add PCRE_INFO_MAXLOOKBEHIND.
906    
907    12. Applied a (slightly modified) user-supplied patch that improves performance
908        when the heap is used for recursion (compiled with --disable-stack-for-
909        recursion). Instead of malloc and free for each heap frame each time a
910        logical recursion happens, frames are retained on a chain and re-used where
911        possible. This sometimes gives as much as 30% improvement.
912    
913    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
914        call.
915    
916    14. As documented, (*COMMIT) is now confined to within a positive assertion.
917    
918    15. It is now possible to link pcretest with libedit as an alternative to
919        libreadline.
920    
921    16. (*COMMIT) control verb is now supported by the JIT compiler.
922    
923    17. The Unicode data tables have been updated to Unicode 6.1.0.
924    
925    18. Added --file-list option to pcregrep.
926    
927    19. Added binary file support to pcregrep, including the -a, --binary-files,
928        -I, and --text options.
929    
930    20. The madvise function is renamed for posix_madvise for QNX compatibility
931        reasons. Fixed by Giuseppe D'Angelo.
932    
933    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
934    
935    22. Changed the option for creating symbolic links for 16-bit man pages from
936        -s to -sf so that re-installing does not cause issues.
937    
938    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
939    
940    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
941        matches in certain environments (the workspace was not being correctly
942        retained). Also added to pcre_dfa_exec() a simple plausibility check on
943        some of the workspace data at the beginning of a restart.
944    
945    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
946        was not doing so when it should - probably a typo introduced by SVN 528
947        (change 8.10/14).
948    
949    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
950        \w+ when the character tables indicated that \x{c4} was a word character.
951        There were several related cases, all because the tests for doing a table
952        lookup were testing for characters less than 127 instead of 255.
953    
954    27. If a pattern contains capturing parentheses that are not used in a match,
955        their slots in the ovector are set to -1. For those that are higher than
956        any matched groups, this happens at the end of processing. In the case when
957        there were back references that the ovector was too small to contain
958        (causing temporary malloc'd memory to be used during matching), and the
959        highest capturing number was not used, memory off the end of the ovector
960        was incorrectly being set to -1. (It was using the size of the temporary
961        memory instead of the true size.)
962    
963    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
964        ovector size, it uses memory at the end of the block that it has got.
965    
966    29. Check for an overlong MARK name and give an error at compile time. The
967        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
968    
969    30. JIT compiler update.
970    
971    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
972        Rill for the patch.
973    
974    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
975    
976    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
977    
978    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
979        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
980        was enabled.
981    
982    35. Improve JIT code generation for greedy plus quantifier.
983    
984    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
985        1 to "aa" instead of to an empty string. The bug affected repeated groups
986        that could potentially match an empty string.
987    
988    37. Optimizing single character iterators in JIT.
989    
990    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
991        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
992        codepoints that are too big for the mode are faulted, and in a UTF mode,
993        disallowed codepoints are also faulted.
994    
995    39. If PCRE was compiled with UTF support, in three places in the DFA
996        matcher there was code that should only have been obeyed in UTF mode, but
997        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
998        processing when bytes with values greater than 127 were present. In 16-bit
999        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1000        both cases the values are those that cannot be the first data item in a UTF
1001        character. The three items that might have provoked this were recursions,
1002        possessively repeated groups, and atomic groups.
1003    
1004    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1005        and pcregrep, because some OS require shared objects to be explicitly
1006        passed to ld, causing the link step to fail if they are not.
1007    
1008    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1009        mode, patterns that started with \h* or \R* might be incorrectly matched.
1010    
1011    
1012    Version 8.30 04-February-2012
1013    -----------------------------
1014    
1015    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1016        name is defined in ctype.h.
1017    
1018    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1019        only in quite long subpatterns.
1020    
1021    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1022        since it was replaced by pcre_fullinfo() in February 2000.
1023    
1024    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1025        match a (*MARK), and the match failed at the start of the subject, a
1026        reference to memory before the start of the subject could occur. This bug
1027        was introduced by fix 17 of release 8.21.
1028    
1029    5.  A reference to an unset group with zero minimum repetition was giving
1030        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1031        /(another)?(\1?)test/ matched against "hello world test". This bug was
1032        introduced in release 8.13.
1033    
1034    6.  Add support for 16-bit character strings (a large amount of work involving
1035        many changes and refactorings).
1036    
1037    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1038        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1039        from a file.
1040    
1041    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1042        rounding is not applied in this particular case).
1043    
1044    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1045        if they appear, or are escaped, in patterns.
1046    
1047    10. Get rid of a number of -Wunused-but-set-variable warnings.
1048    
1049    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1050        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1051        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1052        also returns the mark "x". This bug applied to capturing parentheses,
1053        non-capturing parentheses, and atomic parentheses. It also applied to some
1054        assertions.
1055    
1056    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1057        information out of configure.ac instead of relying on pcre.h.generic, which
1058        is not stored in the repository.
1059    
1060    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1061        -lreadline.
1062    
1063    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1064    
1065    15. Applied Graycode's patch to put the top-level frame on the stack rather
1066        than the heap when not using the stack for recursion. This gives a
1067        performance improvement in many cases when recursion is not deep.
1068    
1069    16. Experimental code added to "pcretest -C" to output the stack frame size.
1070    
1071    
1072    Version 8.21 12-Dec-2011
1073    ------------------------
1074    
1075    1.  Updating the JIT compiler.
1076    
1077    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1078        are added as well.
1079    
1080    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1081        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1082        calling _pcre_jit_exec. Some extra comments are added.
1083    
1084    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1085        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1086        was introduced by change 18 for 8.20.
1087    
1088    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1089        ECMA-262 standard.
1090    
1091    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1092        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1093        This bug was probably introduced by change 9 of 8.13.
1094    
1095    7.  While fixing 6 above, I noticed that a number of other items were being
1096        incorrectly rejected as "not fixed length". This arose partly because newer
1097        opcodes had not been added to the fixed-length checking code. I have (a)
1098        corrected the bug and added tests for these items, and (b) arranged for an
1099        error to occur if an unknown opcode is encountered while checking for fixed
1100        length instead of just assuming "not fixed length". The items that were
1101        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1102        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1103        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1104    
1105    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1106        being incorrectly compiled and would have given unpredicatble results.
1107    
1108    9.  A possessively repeated subpattern with minimum repeat count greater than
1109        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1110        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1111        the first (A) could occur when it should not.
1112    
1113    10. Add a cast and remove a redundant test from the code.
1114    
1115    11. JIT should use pcre_malloc/pcre_free for allocation.
1116    
1117    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1118        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1119        is anything other than /usr, -L is still shown).
1120    
1121    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1122    
1123    14. Perl does not support \N without a following name in a [] class; PCRE now
1124        also gives an error.
1125    
1126    15. If a forward reference was repeated with an upper limit of around 2000,
1127        it caused the error "internal error: overran compiling workspace". The
1128        maximum number of forward references (including repeats) was limited by the
1129        internal workspace, and dependent on the LINK_SIZE. The code has been
1130        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1131        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1132        of around 200,000 forward references. While doing this, I also speeded up
1133        the filling in of repeated forward references.
1134    
1135    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1136        incorrectly expecting the subject to contain another "a" after the start.
1137    
1138    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1139        in the match, the SKIP should be ignored. This was not happening; instead
1140        the SKIP was being treated as NOMATCH. For patterns such as
1141        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1142        tested.
1143    
1144    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1145        now much more compatible with Perl, in particular in cases where the result
1146        is a non-match for a non-anchored pattern. For example, if
1147        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1148        "m", where previously it did not return a name. A side effect of this
1149        change is that for partial matches, the last encountered mark name is
1150        returned, as for non matches. A number of tests that were previously not
1151        Perl-compatible have been moved into the Perl-compatible test files. The
1152        refactoring has had the pleasing side effect of removing one argument from
1153        the match() function, thus reducing its stack requirements.
1154    
1155    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1156        subsequent uses of /S (without +) incorrectly behaved like /S+.
1157    
1158    21. Retrieve executable code size support for the JIT compiler and fixing
1159        some warnings.
1160    
1161    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1162        not work when the shorter character appeared right at the end of the
1163        subject string.
1164    
1165    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1166        systems.
1167    
1168    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1169        output it when the /M option is used in pcretest.
1170    
1171    25. The CheckMan script was not being included in the distribution. Also, added
1172        an explicit "perl" to run Perl scripts from the PrepareRelease script
1173        because this is reportedly needed in Windows.
1174    
1175    26. If study data was being save in a file and studying had not found a set of
1176        "starts with" bytes for the pattern, the data written to the file (though
1177        never used) was taken from uninitialized memory and so caused valgrind to
1178        complain.
1179    
1180    27. Updated RunTest.bat as provided by Sheri Pierce.
1181    
1182    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1183    
1184    29. Computation of memory usage for the table of capturing group names was
1185        giving an unnecessarily large value.
1186    
1187    
1188    Version 8.20 21-Oct-2011
1189    ------------------------
1190    
1191    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1192        a POSIX class. After further experiments with Perl, which convinced me that
1193        Perl has bugs and confusions, a closing square bracket is no longer allowed
1194        in a POSIX name. This bug also affected patterns with classes that started
1195        with full stops.
1196    
1197    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1198        captured substring, but while checking the failing first alternative,
1199        substring 1 is temporarily captured. If the output vector supplied to
1200        pcre_exec() was not big enough for this capture, the yield of the function
1201        was still zero ("insufficient space for captured substrings"). This cannot
1202        be totally fixed without adding another stack variable, which seems a lot
1203        of expense for a edge case. However, I have improved the situation in cases
1204        such as /(a)(b)x|abc/ matched against "abc", where the return code
1205        indicates that fewer than the maximum number of slots in the ovector have
1206        been set.
1207    
1208    3.  Related to (2) above: when there are more back references in a pattern than
1209        slots in the output vector, pcre_exec() uses temporary memory during
1210        matching, and copies in the captures as far as possible afterwards. It was
1211        using the entire output vector, but this conflicts with the specification
1212        that only 2/3 is used for passing back captured substrings. Now it uses
1213        only the first 2/3, for compatibility. This is, of course, another edge
1214        case.
1215    
1216    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1217        main code base, and can be used by building with --enable-jit. When this is
1218        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1219        runtime --no-jit option is given.
1220    
1221    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1222        ovector, the return from the function was zero, implying that there were
1223        other matches that did not fit. The correct "exactly full" value is now
1224        returned.
1225    
1226    6.  If a subpattern that was called recursively or as a subroutine contained
1227        (*PRUNE) or any other control that caused it to give a non-standard return,
1228        invalid errors such as "Error -26 (nested recursion at the same subject
1229        position)" or even infinite loops could occur.
1230    
1231    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1232        computing the minimum length on reaching *ACCEPT, and so ended up with the
1233        wrong value of 1 rather than 0. Further investigation indicates that
1234        computing a minimum subject length in the presence of *ACCEPT is difficult
1235        (think back references, subroutine calls), and so I have changed the code
1236        so that no minimum is registered for a pattern that contains *ACCEPT.
1237    
1238    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1239        it was not handled as intended. [But see 16 below.]
1240    
1241    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1242        Sheri Pierce.
1243    
1244    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1245        the first byte in a match must be "a".
1246    
1247    11. Change 17 for 8.13 increased the recursion depth for patterns like
1248        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1249        pattern contains any instances of (*THEN). If it does not, the old
1250        optimizations are restored. It would be nice to do this on a per-group
1251        basis, but at the moment that is not feasible.
1252    
1253    12. In some environments, the output of pcretest -C is CRLF terminated. This
1254        broke RunTest's code that checks for the link size. A single white space
1255        character after the value is now allowed for.
1256    
1257    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1258        For "fr", it uses the Windows-specific input and output files.
1259    
1260    14. If (*THEN) appeared in a group that was called recursively or as a
1261        subroutine, it did not work as intended. [But see next item.]
1262    
1263    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1264        pattern fragments (but not containing any | characters). If A and B are
1265        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1266        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1267        D. In other words, Perl considers parentheses that do not contain any |
1268        characters to be part of a surrounding alternative, whereas PCRE was
1269        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1270        differently. PCRE now behaves in the same way as Perl, except in the case
1271        of subroutine/recursion calls such as (?1) which have in any case always
1272        been different (but PCRE had them first :-).
1273    
1274    16. Related to 15 above: Perl does not treat the | in a conditional group as
1275        creating alternatives. Such a group is treated in the same way as an
1276        ordinary group without any | characters when processing (*THEN). PCRE has
1277        been changed to match Perl's behaviour.
1278    
1279    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1280        RunGrepTest script failed.
1281    
1282    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1283        inevitable for groups that contain captures, but it can lead to a lot of
1284        stack use in large patterns. The old behaviour has been restored for atomic
1285        groups that do not contain any capturing parentheses.
1286    
1287    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1288        suppress the check for a minimum subject length at run time. (If it was
1289        given to pcre_exec() or pcre_dfa_exec() it did work.)
1290    
1291    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1292        fail to work when decoding hex characters in data strings in EBCDIC
1293        environments.
1294    
1295    21. It appears that in at least one Mac OS environment, the isxdigit() function
1296        is implemented as a macro that evaluates to its argument more than once,
1297        contravening the C 90 Standard (I haven't checked a later standard). There
1298        was an instance in pcretest which caused it to go wrong when processing
1299        \x{...} escapes in subject strings. The has been rewritten to avoid using
1300        things like p++ in the argument of isxdigit().
1301    
1302    
1303    Version 8.13 16-Aug-2011
1304    ------------------------
1305    
1306    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1307    
1308    2.  Two minor typos in pcre_internal.h have been fixed.
1309    
1310    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1311        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1312        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1313    
1314    4.  There were a number of related bugs in the code for matching backrefences
1315        caselessly in UTF-8 mode when codes for the characters concerned were
1316        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1317        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1318        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1319        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1320        2-byte code at the end of the subject (it thought there wasn't enough data
1321        left).
1322    
1323    5.  Comprehensive information about what went wrong is now returned by
1324        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1325        as the output vector has at least 2 elements. The offset of the start of
1326        the failing character and a reason code are placed in the vector.
1327    
1328    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1329        now returned is for the first byte of the failing character, instead of the
1330        last byte inspected. This is an incompatible change, but I hope it is small
1331        enough not to be a problem. It makes the returned offset consistent with
1332        pcre_exec() and pcre_dfa_exec().
1333    
1334    7.  pcretest now gives a text phrase as well as the error number when
1335        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1336        failure, the offset and reason code are output.
1337    
1338    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1339        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1340        back over a single character (\n). This seems wrong (because it treated the
1341        two characters as a single entity when going forwards), conflicts with the
1342        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1343        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1344        has been changed.
1345    
1346    9.  Some internal refactoring has changed the processing so that the handling
1347        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1348        time (the PCRE_DOTALL option was changed this way some time ago: version
1349        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1350        which was always a bit of a fudge. It also means that there is one less
1351        argument for the match() function, which reduces its stack requirements
1352        slightly. This change also fixes an incompatibility with Perl: the pattern
1353        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1354    
1355    10. More internal refactoring has drastically reduced the number of recursive
1356        calls to match() for possessively repeated groups such as (abc)++ when
1357        using pcre_exec().
1358    
1359    11. While implementing 10, a number of bugs in the handling of groups were
1360        discovered and fixed:
1361    
1362        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1363        (a|)*(?1) gave a compile-time internal error.
1364        ((a|)+)+  did not notice that the outer group could match an empty string.
1365        (^a|^)+   was not marked as anchored.
1366        (.*a|.*)+ was not marked as matching at start or after a newline.
1367    
1368    12. Yet more internal refactoring has removed another argument from the match()
1369        function. Special calls to this function are now indicated by setting a
1370        value in a variable in the "match data" data block.
1371    
1372    13. Be more explicit in pcre_study() instead of relying on "default" for
1373        opcodes that mean there is no starting character; this means that when new
1374        ones are added and accidentally left out of pcre_study(), testing should
1375        pick them up.
1376    
1377    14. The -s option of pcretest has been documented for ages as being an old
1378        synonym of -m (show memory usage). I have changed it to mean "force study
1379        for every regex", that is, assume /S for every regex. This is similar to -i
1380        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1381        using it. It makes it easier to run collections of tests with and without
1382        study enabled, and thereby test pcre_study() more easily. All the standard
1383        tests are now run with and without -s (but some patterns can be marked as
1384        "never study" - see 20 below).
1385    
1386    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1387        restoration of the capturing data to the outer values was not happening
1388        correctly.
1389    
1390    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1391        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1392        pattern had matched an empty string, and so incorrectly returned a no
1393        match.
1394    
1395    17. There was optimizing code for the last branch of non-capturing parentheses,
1396        and also for the obeyed branch of a conditional subexpression, which used
1397        tail recursion to cut down on stack usage. Unfortunately, now that there is
1398        the possibility of (*THEN) occurring in these branches, tail recursion is
1399        no longer possible because the return has to be checked for (*THEN). These
1400        two optimizations have therefore been removed. [But see 8.20/11 above.]
1401    
1402    18. If a pattern containing \R was studied, it was assumed that \R always
1403        matched two bytes, thus causing the minimum subject length to be
1404        incorrectly computed because \R can also match just one byte.
1405    
1406    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1407        was incorrectly computed.
1408    
1409    20. If /S is present twice on a test pattern in pcretest input, it now
1410        *disables* studying, thereby overriding the use of -s on the command line
1411        (see 14 above). This is necessary for one or two tests to keep the output
1412        identical in both cases.
1413    
1414    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1415        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1416    
1417    22. When an atomic group that contained a capturing parenthesis was
1418        successfully matched, but the branch in which it appeared failed, the
1419        capturing was not being forgotten if a higher numbered group was later
1420        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1421        group 1 to "a", when in fact it should be unset. This applied to multi-
1422        branched capturing and non-capturing groups, repeated or not, and also to
1423        positive assertions (capturing in negative assertions does not happen
1424        in PCRE) and also to nested atomic groups.
1425    
1426    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1427        subject after a captured substring, to make it easier to tell which of a
1428        number of identical substrings has been captured.
1429    
1430    24. The way atomic groups are processed by pcre_exec() has been changed so that
1431        if they are repeated, backtracking one repetition now resets captured
1432        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1433        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1434        "aaa". Previously, it would have been "a". As part of this code
1435        refactoring, the way recursive calls are handled has also been changed.
1436    
1437    25. If an assertion condition captured any substrings, they were not passed
1438        back unless some other capturing happened later. For example, if
1439        (?(?=(a))a) was matched against "a", no capturing was returned.
1440    
1441    26. When studying a pattern that contained subroutine calls or assertions,
1442        the code for finding the minimum length of a possible match was handling
1443        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1444        group 1 called group 2 while simultaneously a separate group 2 called group
1445        1). A stack overflow occurred in this case. I have fixed this by limiting
1446        the recursion depth to 10.
1447    
1448    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1449        Fortmann. This supports explicit test numbers on the command line, and has
1450        argument validation and error reporting.
1451    
1452    28. An instance of \X with an unlimited repeat could fail if at any point the
1453        first character it looked at was a mark character.
1454    
1455    29. Some minor code refactoring concerning Unicode properties and scripts
1456        should reduce the stack requirement of match() slightly.
1457    
1458    30. Added the '=' option to pcretest to check the setting of unused capturing
1459        slots at the end of the pattern, which are documented as being -1, but are
1460        not included in the return count.
1461    
1462    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1463        compiled something random. Now it gives a compile-time error (as does
1464        Perl).
1465    
1466    32. A *MARK encountered during the processing of a positive assertion is now
1467        recorded and passed back (compatible with Perl).
1468    
1469    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1470        had alternative anchored branches, the search for a second match in a line
1471        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1472        matched the line "0102" twice. The same bug affected patterns that started
1473        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1474        twice.
1475    
1476    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1477        does, and because of capturing effects, quantifying parenthesized
1478        assertions may at times be useful. Quantifiers are now allowed for
1479        parenthesized assertions.
1480    
1481    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1482    
1483    36. \g was being checked for fancy things in a character class, when it should
1484        just be a literal "g".
1485    
1486    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1487        appearance of a nested POSIX class supersedes an apparent external class.
1488        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1489        unescaped square brackets may also appear as part of class names. For
1490        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1491        more like Perl. (But see 8.20/1 above.)
1492    
1493    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1494        was because it thought it was \N{name}, which is not supported).
1495    
1496    39. Add minix to OS list not supporting the -S option in pcretest.
1497    
1498    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1499        cannot analyze patterns in sufficient detail to catch mutual recursions
1500        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1501        subgroup is called recursively as a subpattern for a second time at the
1502        same position in the subject string. In previous releases this might have
1503        been caught by the recursion limit, or it might have run out of stack.
1504    
1505    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1506        happen only once. PCRE was, however incorrectly giving a compile time error
1507        "recursive call could loop indefinitely" because it cannot analyze the
1508        pattern in sufficient detail. The compile time test no longer happens when
1509        PCRE is compiling a conditional subpattern, but actual runaway loops are
1510        now caught at runtime (see 40 above).
1511    
1512    42. It seems that Perl allows any characters other than a closing parenthesis
1513        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1514        has been changed to be the same.
1515    
1516    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1517        as not to get warnings when autogen.sh is called. Also changed
1518        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1519    
1520    44. To help people who use pcregrep to scan files containing exceedingly long
1521        lines, the following changes have been made:
1522    
1523        (a) The default value of the buffer size parameter has been increased from
1524            8K to 20K. (The actual buffer used is three times this size.)
1525    
1526        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1527            PCRE is built.
1528    
1529        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1530            to be set at run time.
1531    
1532        (d) Numerical values in pcregrep options can be followed by K or M, for
1533            example --buffer-size=50K.
1534    
1535        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1536            given and the return code is set to 2.
1537    
1538    45. Add a pointer to the latest mark to the callout data block.
1539    
1540    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1541        partial match of an empty string instead of no match. This was specific to
1542        the use of ".".
1543    
1544    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1545        complete match instead of a partial match. This bug was dependent on both
1546        the PCRE_UTF8 and PCRE_DOTALL options being set.
1547    
1548    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1549        starting byte set, because \b was not being ignored.
1550    
1551    
1552    Version 8.12 15-Jan-2011
1553    ------------------------
1554    
1555    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1556        checks for such things as part of the documentation building process.
1557    
1558    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1559        --match-limit and --recursion-limit options (added for 8.11). In
1560        particular, this made one of the standard tests fail. (The integer value
1561        went into the wrong half of a long int.)
1562    
1563    3.  If the --colour option was given to pcregrep with -v (invert match), it
1564        did strange things, either producing crazy output, or crashing. It should,
1565        of course, ignore a request for colour when reporting lines that do not
1566        match.
1567    
1568    4.  Another pcregrep bug caused similar problems if --colour was specified with
1569        -M (multiline) and the pattern match finished with a line ending.
1570    
1571    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1572        matched in multiline mode, the following line was shown as part of the
1573        match. This seems wrong, so I have changed it.
1574    
1575    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1576        the check for further matches in the same line (so they could be coloured)
1577        to overrun the end of the current line. If another match was found, it was
1578        incorrectly shown (and then shown again when found in the next line).
1579    
1580    7.  If pcregrep was compiled under Windows, there was a reference to the
1581        function pcregrep_exit() before it was defined. I am assuming this was
1582        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1583        reported by a user. I've moved the definition above the reference.
1584    
1585    
1586    Version 8.11 10-Dec-2010
1587    ------------------------
1588    
1589    1.  (*THEN) was not working properly if there were untried alternatives prior
1590        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1591        backtracked to try for "b" instead of moving to the next alternative branch
1592        at the same level (in this case, to look for "c"). The Perl documentation
1593        is clear that when (*THEN) is backtracked onto, it goes to the "next
1594        alternative in the innermost enclosing group".
1595    
1596    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1597        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1598        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1599        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1600        (*THEN).
1601    
1602    3.  If \s appeared in a character class, it removed the VT character from
1603        the class, even if it had been included by some previous item, for example
1604        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1605        of \s, but is part of the POSIX "space" class.)
1606    
1607    4.  A partial match never returns an empty string (because you can always
1608        match an empty string at the end of the subject); however the checking for
1609        an empty string was starting at the "start of match" point. This has been
1610        changed to the "earliest inspected character" point, because the returned
1611        data for a partial match starts at this character. This means that, for
1612        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1613        (previously it gave "no match").
1614    
1615    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1616        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1617        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1618        has an implication that the given string is incomplete (because a partial
1619        match is preferred over a full match). For this reason, these items now
1620        give a partial match in this situation. [Aside: previously, the one case
1621        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1622        match rather than a full match, which was wrong by the old rules, but is
1623        now correct.]
1624    
1625    6.  There was a bug in the handling of #-introduced comments, recognized when
1626        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1627        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1628        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1629        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1630        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1631        places in pcre_compile().
1632    
1633    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1634        comments when looking ahead for named forward references to subpatterns,
1635        the only newline sequence it recognized was NL. It now handles newlines
1636        according to the set newline convention.
1637    
1638    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1639        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1640        cater for a lack of strerror(). These oversights have been fixed.
1641    
1642    9.  Added --match-limit and --recursion-limit to pcregrep.
1643    
1644    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1645    
1646    11. When the -o option was used, pcregrep was setting a return code of 1, even
1647        when matches were found, and --line-buffered was not being honoured.
1648    
1649    12. Added an optional parentheses number to the -o and --only-matching options
1650        of pcregrep.
1651    
1652    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1653        can match an empty string. The code to do it in pcretest and pcredemo
1654        needed fixing:
1655    
1656        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1657            only one byte after an empty string match just before CRLF (this case
1658            just got forgotten; "any" and "anycrlf" were OK).
1659    
1660        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1661            mode when an empty string match preceded an ASCII character followed by
1662            a non-ASCII character. (The code for advancing by one character rather
1663            than one byte was nonsense.)
1664    
1665        (c) The pcredemo.c sample program did not have any code at all to handle
1666            the cases when CRLF is a valid newline sequence.
1667    
1668    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1669        as a starting offset was within the subject string. There is now a new
1670        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1671        negative or greater than the length of the string. In order to test this,
1672        pcretest is extended to allow the setting of negative starting offsets.
1673    
1674    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1675        starting offset points to the beginning of a UTF-8 character was
1676        unnecessarily clumsy. I tidied it up.
1677    
1678    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1679        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1680    
1681    17. Nobody had reported that the --include_dir option, which was added in
1682        release 7.7 should have been called --include-dir (hyphen, not underscore)
1683        for compatibility with GNU grep. I have changed it to --include-dir, but
1684        left --include_dir as an undocumented synonym, and the same for
1685        --exclude-dir, though that is not available in GNU grep, at least as of
1686        release 2.5.4.
1687    
1688    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1689        characters from a string of bytes) have been redefined so as not to use
1690        loops, in order to improve performance in some environments. At the same
1691        time, I abstracted some of the common code into auxiliary macros to save
1692        repetition (this should not affect the compiled code).
1693    
1694    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1695        compile-time error is now given if \c is not followed by an ASCII
1696        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1697        different, and any byte value is allowed.)
1698    
1699    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1700        START_OPTIMIZE option, which is now allowed at compile time - but just
1701        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1702        to pcregrep and other applications that have no direct access to PCRE
1703        options. The new /Y option in pcretest sets this option when calling
1704        pcre_compile().
1705    
1706    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1707        back references. Groups containing recursive back references were forced to
1708        be atomic by that change, but in the case of named groups, the amount of
1709        memory required was incorrectly computed, leading to "Failed: internal
1710        error: code overflow". This has been fixed.
1711    
1712    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1713        pcretest.c, to avoid build problems in some Borland environments.
1714    
1715    
1716    Version 8.10 25-Jun-2010
1717    ------------------------
1718    
1719    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1720        THEN.
1721    
1722    2.  (*ACCEPT) was not working when inside an atomic group.
1723    
1724    3.  Inside a character class, \B is treated as a literal by default, but
1725        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1726        causes the error). The code is unchanged, but I tidied the documentation.
1727    
1728    4.  Inside a character class, PCRE always treated \R and \X as literals,
1729        whereas Perl faults them if its -w option is set. I have changed PCRE so
1730        that it faults them when PCRE_EXTRA is set.
1731    
1732    5.  Added support for \N, which always matches any character other than
1733        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1734    
1735    6.  When compiling pcregrep with newer versions of gcc which may have
1736        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1737        declared with attribute warn_unused_result" were given. Just casting the
1738        result to (void) does not stop the warnings; a more elaborate fudge is
1739        needed. I've used a macro to implement this.
1740    
1741    7.  Minor change to pcretest.c to avoid a compiler warning.
1742    
1743    8.  Added four artifical Unicode properties to help with an option to make
1744        \s etc use properties (see next item). The new properties are: Xan
1745        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1746    
1747    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1748        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1749        this option. Modified pcretest to add /W to test this facility. Added
1750        REG_UCP to make it available via the POSIX interface.
1751    
1752    10. Added --line-buffered to pcregrep.
1753    
1754    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1755        studied, and the match started with a letter with a code point greater than
1756        127 whose first byte was different to the first byte of the other case of
1757        the letter, the other case of this starting letter was not recognized
1758        (#976).
1759    
1760    12. If a pattern that was studied started with a repeated Unicode property
1761        test, for example, \p{Nd}+, there was the theoretical possibility of
1762        setting up an incorrect bitmap of starting bytes, but fortunately it could
1763        not have actually happened in practice until change 8 above was made (it
1764        added property types that matched character-matching opcodes).
1765    
1766    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1767        possible starting bytes for non-anchored patterns.
1768    
1769    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1770        \R, and also a number of cases that involve Unicode properties, both
1771        explicit and implicit when PCRE_UCP is set.
1772    
1773    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1774        input, it could crash or give wrong results if characters with values
1775        greater than 0xc0 were present in the subject string. (Detail: it assumed
1776        UTF-8 input when processing these items.)
1777    
1778    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1779        size_t is 64-bit (#991).
1780    
1781    17. Added a check for running out of memory when PCRE is compiled with
1782        --disable-stack-for-recursion (#990).
1783    
1784    18. If the last data line in a file for pcretest does not have a newline on
1785        the end, a newline was missing in the output.
1786    
1787    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1788        less than 128) in its various bitmaps. However, there is a facility for
1789        generating tables according to the current locale when PCRE is compiled. It
1790        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1791        characters, are recognized by isspace() and therefore were getting set in
1792        these tables, and indeed these tables seem to approximate to ISO 8859. This
1793        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1794        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1795        which of course cannot start UTF-8 characters. I have changed the code so
1796        that only real ASCII characters (less than 128) and the correct starting
1797        bytes for UTF-8 encodings are set for characters greater than 127 when in
1798        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1799        altogether.)
1800    
1801    20. Added the /T option to pcretest so as to be able to run tests with non-
1802        standard character tables, thus making it possible to include the tests
1803        used for 19 above in the standard set of tests.
1804    
1805    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1806        reference to a subpattern the other side of a comment that contains an
1807        opening parenthesis caused either an internal compiling error, or a
1808        reference to the wrong subpattern.
1809    
1810    
1811    Version 8.02 19-Mar-2010
1812    ------------------------
1813    
1814    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1815    
1816    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1817        configured.
1818    
1819    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1820        original author of that file, following a query about its status.
1821    
1822    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1823        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1824    
1825    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1826        quantifier applied to a forward-referencing subroutine call, could compile
1827        incorrect code or give the error "internal error: previously-checked
1828        referenced subpattern not found".
1829    
1830    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1831        variables to point to external functions. For these systems, therefore,
1832        pcre_malloc etc. are now initialized to local functions that call the
1833        relevant global functions.
1834    
1835    7.  There were two entries missing in the vectors called coptable and poptable
1836        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1837        I've fixed the data, and added a kludgy way of testing at compile time that
1838        the lengths are correct (equal to the number of opcodes).
1839    
1840    8.  Following on from 7, I added a similar kludge to check the length of the
1841        eint vector in pcreposix.c.
1842    
1843    9.  Error texts for pcre_compile() are held as one long string to avoid too
1844        much relocation at load time. To find a text, the string is searched,
1845        counting zeros. There was no check for running off the end of the string,
1846        which could happen if a new error number was added without updating the
1847        string.
1848    
1849    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1850    
1851    11. \K was not working if it appeared in an atomic group or in a group that
1852        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1853        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1854        the assertion is positive, but not if it is negative.
1855    
1856    12. Change 11 fortuitously reduced the size of the stack frame used in the
1857        "match()" function of pcre_exec.c by one pointer. Forthcoming
1858        implementation of support for (*MARK) will need an extra pointer on the
1859        stack; I have reserved it now, so that the stack frame size does not
1860        decrease.
1861    
1862    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1863        item in branch that calls a recursion is a subroutine call - as in the
1864        second branch in the above example - was incorrectly given the compile-
1865        time error "recursive call could loop indefinitely" because pcre_compile()
1866        was not correctly checking the subroutine for matching a non-empty string.
1867    
1868    14. The checks for overrunning compiling workspace could trigger after an
1869        overrun had occurred. This is a "should never occur" error, but it can be
1870        triggered by pathological patterns such as hundreds of nested parentheses.
1871        The checks now trigger 100 bytes before the end of the workspace.
1872    
1873    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1874    
1875    
1876    Version 8.01 19-Jan-2010
1877    ------------------------
1878    
1879    1.  If a pattern contained a conditional subpattern with only one branch (in
1880        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1881        computed the wrong minimum data length (which is of course zero for such
1882        subpatterns). This could cause incorrect "no match" results.
1883    
1884    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1885        the pattern is reset in the first branch, pcre_compile() failed with
1886        "internal error: code overflow at offset...". This happened only when
1887        the reset was to the original external option setting. (An optimization
1888        abstracts leading options settings into an external setting, which was the
1889        cause of this.)
1890    
1891    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1892        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1893        assertion pattern did not match (meaning that the assertion was true), it
1894        was incorrectly treated as false if the SKIP had been reached during the
1895        matching. This also applied to assertions used as conditions.
1896    
1897    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1898        assertion subpattern, including such a pattern used as a condition,
1899        unpredictable results occurred, instead of the error return
1900        PCRE_ERROR_DFA_UITEM.
1901    
1902    5.  The C++ GlobalReplace function was not working like Perl for the special
1903        situation when an empty string is matched. It now does the fancy magic
1904        stuff that is necessary.
1905    
1906    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1907        removed. (These were left over from very, very early versions of PCRE.)
1908    
1909    7.  Some cosmetic changes to the code to make life easier when compiling it
1910        as part of something else:
1911    
1912        (a) Change DEBUG to PCRE_DEBUG.
1913    
1914        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1915            called "current" as "current_branch", to prevent a collision with the
1916            Linux macro when compiled as a kernel module.
1917    
1918        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1919            prevent a collision with the Linux macro when compiled as a kernel
1920            module.
1921    
1922    8.  In pcre_compile() there are some checks for integer overflows that used to
1923        cast potentially large values to (double). This has been changed to that
1924        when building, a check for int64_t is made, and if it is found, it is used
1925        instead, thus avoiding the use of floating point arithmetic. (There is no
1926        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1927        double.
1928    
1929    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1930        2005 (difference between two addresses compared to an unsigned value).
1931    
1932    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1933        custom one, because of the following reported problem in Windows:
1934    
1935          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1936              under Win32.
1937          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1938              therefore missing the function definition.
1939          - The compiler thus generates a "C" signature for the test function.
1940          - The linker fails to find the "C" function.
1941          - PCRE fails to configure if asked to do so against libbz2.
1942    
1943    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1944        messages were output:
1945    
1946          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1947          rerunning libtoolize, to keep the correct libtool macros in-tree.
1948          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1949    
1950        I have done both of these things.
1951    
1952    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1953        most of the time, it *can* run out if it is given a pattern that contains a
1954        runaway infinite recursion. I updated the discussion in the pcrestack man
1955        page.
1956    
1957    13. Now that we have gone to the x.xx style of version numbers, the minor
1958        version may start with zero. Using 08 or 09 is a bad idea because users
1959        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1960        interpreted as invalid octal numbers. I've updated the previous comment in
1961        configure.ac, and also added a check that gives an error if 08 or 09 are
1962        used.
1963    
1964    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1965        causing partial matching to fail when the end of the subject matched \W
1966        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1967    
1968    15. There were some discrepancies between the declarations in pcre_internal.h
1969        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1970        their definitions. The declarations used "const uschar *" and the
1971        definitions used USPTR. Even though USPTR is normally defined as "const
1972        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1973        reported that: "This difference in casting confuses some C++ compilers, for
1974        example, SunCC recognizes above declarations as different functions and
1975        generates broken code for hbpcre." I have changed the declarations to use
1976        USPTR.
1977    
1978    16. GNU libtool is named differently on some systems. The autogen.sh script now
1979        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1980        (FreeBSD).
1981    
1982    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1983        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1984        comment: "Figure out how to create a longlong from a string: strtoll and
1985        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1986        instance, but it only takes 2 args instead of 3!"
1987    
1988    18. A subtle bug concerned with back references has been fixed by a change of
1989        specification, with a corresponding code fix. A pattern such as
1990        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1991        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1992        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1993        same bug. Such groups have to be quantified to be useful, or contained
1994        inside another quantified group. (If there's no repetition, the reference
1995        can never match.) The problem arises because, having left the group and
1996        moved on to the rest of the pattern, a later failure that backtracks into
1997        the group uses the captured value from the final iteration of the group
1998        rather than the correct earlier one. I have fixed this in PCRE by forcing
1999        any group that contains a reference to itself to be an atomic group; that
2000        is, there cannot be any backtracking into it once it has completed. This is
2001        similar to recursive and subroutine calls.
2002    
2003    
2004    Version 8.00 19-Oct-09
2005    ----------------------
2006    
2007    1.  The table for translating pcre_compile() error codes into POSIX error codes
2008        was out-of-date, and there was no check on the pcre_compile() error code
2009        being within the table. This could lead to an OK return being given in
2010        error.
2011    
2012    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2013        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2014        in a Windows environment.
2015    
2016    3.  The pcregrep --count option prints the count for each file even when it is
2017        zero, as does GNU grep. However, pcregrep was also printing all files when
2018        --files-with-matches was added. Now, when both options are given, it prints
2019        counts only for those files that have at least one match. (GNU grep just
2020        prints the file name in this circumstance, but including the count seems
2021        more useful - otherwise, why use --count?) Also ensured that the
2022        combination -clh just lists non-zero counts, with no names.
2023    
2024    4.  The long form of the pcregrep -F option was incorrectly implemented as
2025        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2026        but it seems right to fix it, and I didn't think it was worth preserving
2027        the old behaviour.
2028    
2029    5.  The command line items --regex=pattern and --regexp=pattern were not
2030        recognized by pcregrep, which required --regex pattern or --regexp pattern
2031        (with a space rather than an '='). The man page documented the '=' forms,
2032        which are compatible with GNU grep; these now work.
2033    
2034    6.  No libpcreposix.pc file was created for pkg-config; there was just
2035        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2036    
2037    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2038        when UCP support is not needed, by modifying the Python script that
2039        generates it from Unicode data files. This should not matter if the module
2040        is correctly used as a library, but I received one complaint about 50K of
2041        unwanted data. My guess is that the person linked everything into his
2042        program rather than using a library. Anyway, it does no harm.
2043    
2044    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2045        was a minimum greater than 1 for a wide character in a possessive
2046        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2047        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2048        character. Chaos in the form of incorrect output or a compiling loop could
2049        result.
2050    
2051    9.  The restrictions on what a pattern can contain when partial matching is
2052        requested for pcre_exec() have been removed. All patterns can now be
2053        partially matched by this function. In addition, if there are at least two
2054        slots in the offset vector, the offset of the earliest inspected character
2055        for the match and the offset of the end of the subject are set in them when
2056        PCRE_ERROR_PARTIAL is returned.
2057    
2058    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2059        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2060        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2061        and may be more useful for multi-segment matching.
2062    
2063    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2064        used to be given if ever the end of the subject was reached; now it is
2065        given only if matching could not proceed because another character was
2066        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2067        string "Z", which now yields "no match" instead of "partial match". In the
2068        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2069        final character ended with (*FAIL).
2070    
2071    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2072        if the pattern had a "must contain" character that was already found in the
2073        earlier partial match, unless partial matching was again requested. For
2074        example, with the pattern /dog.(body)?/, the "must contain" character is
2075        "g". If the first part-match was for the string "dog", restarting with
2076        "sbody" failed. This bug has been fixed.
2077    
2078    13. The string returned by pcre_dfa_exec() after a partial match has been
2079        changed so that it starts at the first inspected character rather than the
2080        first character of the match. This makes a difference only if the pattern
2081        starts with a lookbehind assertion or \b or \B (\K is not supported by
2082        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2083        matching functions compatible, and I think it's the right thing to do.
2084    
2085    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2086        so that the demonstration program is easily available in environments where
2087        PCRE has not been installed from source.
2088    
2089    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2090        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2091        library.
2092    
2093    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2094        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2095        is not the first non-POSIX option to be added. Clearly some people find
2096        these options useful.
2097    
2098    17. If a caller to the POSIX matching function regexec() passes a non-zero
2099        value for nmatch with a NULL value for pmatch, the value of
2100        nmatch is forced to zero.
2101    
2102    18. RunGrepTest did not have a test for the availability of the -u option of
2103        the diff command, as RunTest does. It now checks in the same way as
2104        RunTest, and also checks for the -b option.
2105    
2106    19. If an odd number of negated classes containing just a single character
2107        interposed, within parentheses, between a forward reference to a named
2108        subpattern and the definition of the subpattern, compilation crashed with
2109        an internal error, complaining that it could not find the referenced
2110        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2111        [The bug was that it was starting one character too far in when skipping
2112        over the character class, thus treating the ] as data rather than
2113        terminating the class. This meant it could skip too much.]
2114    
2115    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2116        /g option in pcretest when the pattern contains \K, which makes it possible
2117        to have an empty string match not at the start, even when the pattern is
2118        anchored. Updated pcretest and pcredemo to use this option.
2119    
2120    21. If the maximum number of capturing subpatterns in a recursion was greater
2121        than the maximum at the outer level, the higher number was returned, but
2122        with unset values at the outer level. The correct (outer level) value is
2123        now given.
2124    
2125    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2126        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2127        make it do so. The string so far is captured, making this feature
2128        compatible with Perl.
2129    
2130    23. The tests have been re-organized, adding tests 11 and 12, to make it
2131        possible to check the Perl 5.10 features against Perl 5.10.
2132    
2133    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2134        pattern matches a fixed length string. PCRE did not allow this; now it
2135        does. Neither allows recursion.
2136    
2137    25. I finally figured out how to implement a request to provide the minimum
2138        length of subject string that was needed in order to match a given pattern.
2139        (It was back references and recursion that I had previously got hung up
2140        on.) This code has now been added to pcre_study(); it finds a lower bound
2141        to the length of subject needed. It is not necessarily the greatest lower
2142        bound, but using it to avoid searching strings that are too short does give
2143        some useful speed-ups. The value is available to calling programs via
2144        pcre_fullinfo().
2145    
2146    26. While implementing 25, I discovered to my embarrassment that pcretest had
2147        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2148        study optimizations had never been tested with that matching function.
2149        Oops. What is worse, even when it was passed study data, there was a bug in
2150        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2151        were also very few tests of studied patterns with pcre_dfa_exec().
2152    
2153    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2154        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2155        on the other side of the coin, they are no longer allowed to have different
2156        names, because these cannot be distinguished in PCRE, and this has caused
2157        confusion. (This is a difference from Perl.)
2158    
2159    28. When duplicate subpattern names are present (necessarily with different
2160        numbers, as required by 27 above), and a test is made by name in a
2161        conditional pattern, either for a subpattern having been matched, or for
2162        recursion in such a pattern, all the associated numbered subpatterns are
2163        tested, and the overall condition is true if the condition is true for any
2164        one of them. This is the way Perl works, and is also more like the way
2165        testing by number works.
2166    
2167    
2168    Version 7.9 11-Apr-09
2169    ---------------------
2170    
2171    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2172        (pcretest), all targets were linked against these libraries. This included
2173        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2174        libraries. This caused unwanted dependencies to be created. This problem
2175        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2176        pcretest is linked with readline.
2177    
2178    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2179        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2180        moved outside it again, because FALSE and TRUE are already defined in AIX,
2181        but BOOL is not.
2182    
2183    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2184        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2185    
2186    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2187        hyphens) following file names and line numbers when outputting matching
2188        lines. This is not true; no spaces are inserted. I have also clarified the
2189        wording for the --colour (or --color) option.
2190    
2191    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2192        was not coloured; this is different to GNU grep, so I have changed it to be
2193        the same.
2194    
2195    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2196        each matching line was coloured. Now it goes on to look for further matches
2197        of any of the test patterns, which is the same behaviour as GNU grep.
2198    
2199    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2200        doesn't make sense to accept an empty string match in pcregrep, so I have
2201        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2202        seems to be how GNU grep behaves. [But see later change 40 for release
2203        8.33.]
2204    
2205    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2206        start or after a newline", because the conditional assertion was not being
2207        correctly handled. The rule now is that both the assertion and what follows
2208        in the first alternative must satisfy the test.
2209    
2210    9.  If auto-callout was enabled in a pattern with a conditional group whose
2211        condition was an assertion, PCRE could crash during matching, both with
2212        pcre_exec() and pcre_dfa_exec().
2213    
2214    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2215        used for matching.
2216    
2217    11. Unicode property support in character classes was not working for
2218        characters (bytes) greater than 127 when not in UTF-8 mode.
2219    
2220    12. Added the -M command line option to pcretest.
2221    
2222    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2223    
2224    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2225    
2226    16. Added comments and documentation about mis-use of no_arg in the C++
2227        wrapper.
2228    
2229    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2230        from Martin Jerabek that uses macro names for all relevant character and
2231        string constants.
2232    
2233    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2234        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2235        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2236        these, but not everybody uses configure.
2237    
2238    19. A conditional group that had only one branch was not being correctly
2239        recognized as an item that could match an empty string. This meant that an
2240        enclosing group might also not be so recognized, causing infinite looping
2241        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2242        with the subject "ab", where knowledge that the repeated group can match
2243        nothing is needed in order to break the loop.
2244    
2245    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2246        exec(), but without supplying a callout function, matching went wrong.
2247    
2248    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2249        leak if the size of the offset vector was greater than 30. When the vector
2250        is smaller, the saved offsets during recursion go onto a local stack
2251        vector, but for larger vectors malloc() is used. It was failing to free
2252        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2253        error, in fact).
2254    
2255    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2256        heapframe that is used only when UTF-8 support is enabled. This caused no
2257        problem, but was untidy.
2258    
2259    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2260        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2261        included within another project.
2262    
2263    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2264        slightly modified by me:
2265    
2266          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2267              not building pcregrep.
2268    
2269          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2270              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2271    
2272    25. Forward references, both numeric and by name, in patterns that made use of
2273        duplicate group numbers, could behave incorrectly or give incorrect errors,
2274        because when scanning forward to find the reference group, PCRE was not
2275        taking into account the duplicate group numbers. A pattern such as
2276        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2277    
2278    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2279        the feature of a custom pointer more persuasive (as requested by a user).
2280    
2281    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2282        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2283        pre-defined.
2284    
2285    28. Added support for (*UTF8) at the start of a pattern.
2286    
2287    29. Arrange for flags added by the "release type" setting in CMake to be shown
2288        in the configuration summary.
2289    
2290    
2291    Version 7.8 05-Sep-08
2292    ---------------------
2293    
2294    1.  Replaced UCP searching code with optimized version as implemented for Ad
2295        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2296        stage table and inline lookup instead of a function, giving speed ups of 2
2297        to 5 times on some simple patterns that I tested. Permission was given to
2298        distribute the MultiStage2.py script that generates the tables (it's not in
2299        the tarball, but is in the Subversion repository).
2300    
2301    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2302        scripts.
2303    
2304    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2305        a group with a zero qualifier. The result of the study could be incorrect,
2306        or the function might crash, depending on the pattern.
2307    
2308    4.  Caseless matching was not working for non-ASCII characters in back
2309        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2310        It now works when Unicode Property Support is available.
2311    
2312    5.  In pcretest, an escape such as \x{de} in the data was always generating
2313        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2314        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2315        truncation.
2316    
2317    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2318    
2319    7.  Added two (int) casts to pcregrep when printing the difference of two
2320        pointers, in case they are 64-bit values.
2321    
2322    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2323        test 2 if it fails.
2324    
2325    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2326        and a #define of that name to empty if it is not externally set. This is to
2327        allow users of MSVC to set it if necessary.
2328    
2329    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2330        the convenience functions in the pcre_get.c source file.
2331    
2332    11. An option change at the start of a pattern that had top-level alternatives
2333        could cause overwriting and/or a crash. This command provoked a crash in
2334        some environments:
2335    
2336          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2337    
2338        This potential security problem was recorded as CVE-2008-2371.
2339    
2340    12. For a pattern where the match had to start at the beginning or immediately
2341        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2342        pcre_dfa_exec() could read past the end of the passed subject if there was
2343        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2344        pcretest so that it places the subject at the end of its malloc-ed buffer.
2345    
2346    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2347        exec() might read past the end of the data buffer in UTF-8 mode.
2348    
2349    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2350        the data contained the byte 0x85 as part of a UTF-8 character within its
2351        first line. This applied both to normal and DFA matching.
2352    
2353    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2354        /^[^d]*?$/8 failed to match "abc".
2355    
2356    16. Added a missing copyright notice to pcrecpp_internal.h.
2357    
2358    17. Make it more clear in the documentation that values returned from
2359        pcre_exec() in ovector are byte offsets, not character counts.
2360    
2361    18. Tidied a few places to stop certain compilers from issuing warnings.
2362    
2363    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2364        supplied by Stefan Weber. I made a further small update for 7.8 because
2365        there is a change of source arrangements: the pcre_searchfuncs.c module is
2366        replaced by pcre_ucd.c.
2367    
2368    
2369    Version 7.7 07-May-08
2370    ---------------------
2371    
2372    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2373        a string to a long long, pretend we don't even have a long long." This is
2374        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2375    
2376    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2377        pre-7.6 versions, which defined a global no_arg variable instead of putting
2378        it in the RE class. (See also #8 below.)
2379    
2380    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2381        Lopes.
2382    
2383    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2384    
2385        (1) The include/exclude patterns were being applied to the whole pathnames
2386            of files, instead of just to the final components.
2387    
2388        (2) If there was more than one level of directory, the subdirectories were
2389            skipped unless they satisfied the include/exclude conditions. This is
2390            inconsistent with GNU grep (and could even be seen as contrary to the
2391            pcregrep specification - which I improved to make it absolutely clear).
2392            The action now is always to scan all levels of directory, and just
2393            apply the include/exclude patterns to regular files.
2394    
2395    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2396        --exclude_dir in the tests to avoid scanning .svn directories.
2397    
2398    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2399        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2400        doesn't support NULs in patterns.
2401    
2402    7.  Added some missing "const"s to declarations of static tables in
2403        pcre_compile.c and pcre_dfa_exec.c.
2404    
2405    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2406        caused by fix #2  above. (Subsequently also a second patch to fix the
2407        first patch. And a third patch - this was a messy problem.)
2408    
2409    9.  Applied Craig's patch to remove the use of push_back().
2410    
2411    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2412        matching function regexec().
2413    
2414    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2415        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2416        references. PCRE supports relative numbers with this syntax (I don't think
2417        Oniguruma does).
2418    
2419    12. Previously, a group with a zero repeat such as (...){0} was completely
2420        omitted from the compiled regex. However, this means that if the group
2421        was called as a subroutine from elsewhere in the pattern, things went wrong
2422        (an internal error was given). Such groups are now left in the compiled
2423        pattern, with a new opcode that causes them to be skipped at execution
2424        time.
2425    
2426    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2427        to the way PCRE behaves:
2428    
2429        (a) A lone ] character is dis-allowed (Perl treats it as data).
2430    
2431        (b) A back reference to an unmatched subpattern matches an empty string
2432            (Perl fails the current match path).
2433    
2434        (c) A data ] in a character class must be notated as \] because if the
2435            first data character in a class is ], it defines an empty class. (In
2436            Perl it is not possible to have an empty class.) The empty class []
2437            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2438            The negative empty class [^] matches any one character, independently
2439            of the DOTALL setting.
2440    
2441    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2442        non-existent subpattern following a character class starting with ']' and
2443        containing () gave an internal compiling error instead of "reference to
2444        non-existent subpattern". Fortunately, when the pattern did exist, the
2445        compiled code was correct. (When scanning forwards to check for the
2446        existence of the subpattern, it was treating the data ']' as terminating
2447        the class, so got the count wrong. When actually compiling, the reference
2448        was subsequently set up correctly.)
2449    
2450    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2451        it was being rejected as not supported by pcre_dfa_exec(), even though
2452        other assertions are supported. I have made pcre_dfa_exec() support
2453        (*FAIL).
2454    
2455    16. The implementation of 13c above involved the invention of a new opcode,
2456        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2457        cannot be changed at match time, I realized I could make a small
2458        improvement to matching performance by compiling OP_ALLANY instead of
2459        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2460        on the OP_ANY path.
2461    
2462    17. Compiling pcretest on Windows with readline support failed without the
2463        following two fixes: (1) Make the unistd.h include conditional on
2464        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2465    
2466    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2467        ncurses library to be included for pcretest when ReadLine support is
2468        requested, but also to allow for it to be overridden. This patch came from
2469        Daniel Bergström.
2470    
2471    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2472        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2473        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2474        spotting this.
2475    
2476    
2477    Version 7.6 28-Jan-08
2478    ---------------------
2479    
2480    1.  A character class containing a very large number of characters with
2481        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2482        overflow.
2483    
2484    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2485        HAVE_LONG_LONG is not defined.
2486    
2487    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2488        bring them up to date and include new features. This patch includes:
2489    
2490        - Fixed PH's badly added libz and libbz2 support.
2491        - Fixed a problem with static linking.
2492        - Added pcredemo. [But later removed - see 7 below.]
2493        - Fixed dftables problem and added an option.
2494        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2495            HAVE_LONG_LONG.
2496        - Added readline support for pcretest.
2497        - Added an listing of the option settings after cmake has run.
2498    
2499    4.  A user submitted a patch to Makefile that makes it easy to create
2500        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2501        Makefile.am that cause it to include this special target, without
2502        affecting anything else. Note that the same mingw target plus all
2503        the other distribution libraries and programs are now supported
2504        when configuring with CMake (see 6 below) instead of with
2505        Configure/Make.
2506    
2507    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2508        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2509        exported in the Windows port". It has not yet been confirmed that the patch
2510        solves the problem, but it does no harm.
2511    
2512    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2513        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2514        with CMake, and also correct the comment about stack recursion.
2515    
2516    7.  Remove the automatic building of pcredemo from the ./configure system and
2517        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2518        of a program that users should build themselves after PCRE is installed, so
2519        building it automatically is not really right. What is more, it gave
2520        trouble in some build environments.
2521    
2522    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2523    
2524    
2525    Version 7.5 10-Jan-08
2526    ---------------------
2527    
2528    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2529        values in parens when parsing an RE using the C++ wrapper."
2530    
2531    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2532        Characters greater than 255 were excluded from the class instead of being
2533        included.
2534    
2535    3.  The same bug as (2) above applied to negated POSIX classes such as
2536        [:^space:].
2537    
2538    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2539        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2540        I have changed it.
2541    
2542    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2543        first named subpattern) and a construct such as (?&a) would reference the
2544        first named subpattern whose name started with "a" (in other words, the
2545        length check was missing). Both these problems are fixed. "Subpattern name
2546        expected" is now given for (?&) (a zero-length name), and this patch also
2547        makes it give the same error for \k'' (previously it complained that that
2548        was a reference to a non-existent subpattern).
2549    
2550    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2551        this is right because (?- can be followed by option settings as well as by
2552        digits. I have, however, made the messages clearer.
2553    
2554    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2555        than the number used in the conditional) now cause a compile-time error.
2556        This is actually not compatible with Perl, which accepts such patterns, but
2557        treats the conditional as always being FALSE (as PCRE used to), but it
2558        seems to me that giving a diagnostic is better.
2559    
2560    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2561        and messages.
2562    
2563    9.  Fix two occurrences of "backslash" in comments that should have been
2564        "backspace".
2565    
2566    10. Remove two redundant lines of code that can never be obeyed (their function
2567        was moved elsewhere).
2568    
2569    11. The program that makes PCRE's Unicode character property table had a bug
2570        which caused it to generate incorrect table entries for sequences of
2571        characters that have the same character type, but are in different scripts.
2572        It amalgamated them into a single range, with the script of the first of
2573        them. In other words, some characters were in the wrong script. There were
2574        thirteen such cases, affecting characters in the following ranges:
2575    
2576          U+002b0 - U+002c1
2577          U+0060c - U+0060d
2578          U+0061e - U+00612
2579          U+0064b - U+0065e
2580          U+0074d - U+0076d
2581          U+01800 - U+01805
2582          U+01d00 - U+01d77
2583          U+01d9b - U+01dbf
2584          U+0200b - U+0200f
2585          U+030fc - U+030fe
2586          U+03260 - U+0327f
2587          U+0fb46 - U+0fbb1
2588          U+10450 - U+1049d
2589    
2590    12. The -o option (show only the matching part of a line) for pcregrep was not
2591        compatible with GNU grep in that, if there was more than one match in a
2592        line, it showed only the first of them. It now behaves in the same way as
2593        GNU grep.
2594    
2595    13. If the -o and -v options were combined for pcregrep, it printed a blank
2596        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2597        does the same. The return code can be used to tell if there were any
2598        non-matching lines.
2599    
2600    14. Added --file-offsets and --line-offsets to pcregrep.
2601    
2602    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2603        infinitely looping recursion. The bug was that positive lookaheads were not
2604        being skipped when checking for a possible empty match (negative lookaheads
2605        and both kinds of lookbehind were skipped).
2606    
2607    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2608        inclusion of <windows.h> to before rather than after the definition of
2609        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2610    
2611    17. Specifying a possessive quantifier with a specific limit for a Unicode
2612        character property caused pcre_compile() to compile bad code, which led at
2613        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2614        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2615        caused the error; without that there was no problem.
2616    
2617    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2618    
2619    19. Added --enable-pcretest-libreadline.
2620    
2621    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2622        RE::GlobalReplace(). As a result, the number of replacements returned was
2623        double what it should be. I removed one of the increments, but Craig sent a
2624        later patch that removed the other one (the right fix) and added unit tests
2625        that check the return values (which was not done before).
2626    
2627    21. Several CMake things:
2628    
2629        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2630            the names libpcre and libpcreposix, not just pcre and pcreposix.
2631    
2632        (2) The above change means that pcretest and pcregrep are now correctly
2633            linked with the newly-built libraries, not previously installed ones.
2634    
2635        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2636    
2637    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2638        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2639        UTF-8 newline character). The key issue is that the pattern starts .*;
2640        this means that the match must be either at the beginning, or after a
2641        newline. The bug was in the code for advancing after a failed match and
2642        checking that the new position followed a newline. It was not taking
2643        account of UTF-8 characters correctly.
2644    
2645    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2646        character classes. PCRE was not treating the sequence [:...:] as a
2647        character class unless the ... were all letters. Perl, however, seems to
2648        allow any characters between [: and :], though of course it rejects as
2649        unknown any "names" that contain non-letters, because all the known class
2650        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2651        for example, whereas PCRE did not - it did not recognize a POSIX character
2652        class. This seemed a bit dangerous, so the code has been changed to be
2653        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2654        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2655        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2656        Perl does, and where it didn't before.
2657    
2658    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2659        Windows environments %n is disabled by default.
2660    
2661    
2662    Version 7.4 21-Sep-07
2663    ---------------------
2664    
2665    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2666        means that a class such as [\s] counted as "explicit reference to CR or
2667        LF". That isn't really right - the whole point of the change was to try to
2668        help when there was an actual mention of one of the two characters. So now
2669        the change happens only if \r or \n (or a literal CR or LF) character is
2670        encountered.
2671    
2672    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2673        of both had grown to the point where there were only 3 bits left.
2674        Fortunately, there was spare space in the data structure, and so I have
2675        moved the internal flags into a new 16-bit field to free up more option
2676        bits.
2677    
2678    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2679        but did not set the internal JCHANGED flag - either of these is enough to
2680        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2681        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2682        start sets both bits.
2683    
2684    4.  Added options (at build time, compile time, exec time) to change \R from
2685        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2686    
2687    5.  doc/pcresyntax.html was missing from the distribution.
2688    
2689    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2690        compatibility, even though it is no longer used.
2691    
2692    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2693        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2694        windows.h file is present (where different names are used). [This was
2695        reversed later after testing - see 16 below.]
2696    
2697    8.  Changed all #include <config.h> to #include "config.h". There were also
2698        some further <pcre.h> cases that I changed to "pcre.h".
2699    
2700    9.  When pcregrep was used with the --colour option, it missed the line ending
2701        sequence off the lines that it output.
2702    
2703    10. It was pointed out to me that arrays of string pointers cause lots of
2704        relocations when a shared library is dynamically loaded. A technique of
2705        using a single long string with a table of offsets can drastically reduce
2706        these. I have refactored PCRE in four places to do this. The result is
2707        dramatic:
2708    
2709          Originally:                          290
2710          After changing UCP table:            187
2711          After changing error message table:   43
2712          After changing table of "verbs"       36
2713          After changing table of Posix names   22
2714    
2715        Thanks to the folks working on Gregex for glib for this insight.
2716    
2717    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2718        unicode-properties was also set.
2719    
2720    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2721    
2722    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2723        checked only for CRLF.
2724    
2725    14. Added casts to pcretest.c to avoid compiler warnings.
2726    
2727    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2728    
2729    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2730        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2731        entirely. This removes changes made in 7 above.
2732    
2733    17. The CMake files have been updated, and there is now more information about
2734        building with CMake in the NON-UNIX-USE document.
2735    
2736    
2737    Version 7.3 28-Aug-07
2738    ---------------------
2739    
2740     1. In the rejigging of the build system that eventually resulted in 7.1, the
2741        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2742        brackets there is not right, since it causes compilers to look for an
2743        installed pcre.h, not the version that is in the source that is being
2744        compiled (which of course may be different). I have changed it back to:
2745    
2746          #include "pcre.h"
2747    
2748        I have a vague recollection that the change was concerned with compiling in
2749        different directories, but in the new build system, that is taken care of
2750        by the VPATH setting the Makefile.
2751    
2752     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2753        when the subject happened to end in the byte 0x85 (e.g. if the last
2754        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2755        characters but of course it shouldn't be taken as a newline when it is part
2756        of another character. The bug was that, for an unlimited repeat of . in
2757        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2758        characters when looking for a newline.
2759    
2760     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2761    
2762     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2763        in debug output.
2764    
2765     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2766        long printing in the pcrecpp unittest when running under MinGW.
2767    
2768     6. ESC_K was left out of the EBCDIC table.
2769    
2770     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2771        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2772        limit also applies to "virtual nesting" when a pattern is recursive, and in
2773        this case 1000 isn't so big. I have been able to remove this limit at the
2774        expense of backing off one optimization in certain circumstances. Normally,
2775        when pcre_exec() would call its internal match() function recursively and
2776        immediately return the result unconditionally, it uses a "tail recursion"
2777        feature to save stack. However, when a subpattern that can match an empty
2778        string has an unlimited repetition quantifier, it no longer makes this
2779        optimization. That gives it a stack frame in which to save the data for
2780        checking that an empty string has been matched. Previously this was taken
2781        from the 1000-entry workspace that had been reserved. So now there is no
2782        explicit limit, but more stack is used.
2783    
2784     8. Applied Daniel's patches to solve problems with the import/export magic
2785        syntax that is required for Windows, and which was going wrong for the
2786        pcreposix and pcrecpp parts of the library. These were overlooked when this
2787        problem was solved for the main library.
2788    
2789     9. There were some crude static tests to avoid integer overflow when computing
2790        the size of patterns that contain repeated groups with explicit upper
2791        limits. As the maximum quantifier is 65535, the maximum group length was
2792        set at 30,000 so that the product of these two numbers did not overflow a
2793        32-bit integer. However, it turns out that people want to use groups that
2794        are longer than 30,000 bytes (though not repeat them that many times).
2795        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2796        made it possible to implement the integer overflow checks in a much more
2797        dynamic way, which I have now done. The artificial limitation on group
2798        length has been removed - we now have only the limit on the total length of
2799        the compiled pattern, which depends on the LINK_SIZE setting.
2800    
2801    10. Fixed a bug in the documentation for get/copy named substring when
2802        duplicate names are permitted. If none of the named substrings are set, the
2803        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2804        empty string.
2805    
2806    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2807        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2808        because the ] is interpreted as the first data character and the
2809        terminating ] is not found. PCRE has been made compatible with Perl in this
2810        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2811        cause memory overwriting.
2812    
2813    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2814        string has been matched (to stop an infinite loop). It was not recognizing
2815        a conditional subpattern that could match an empty string if that
2816        subpattern was within another subpattern. For example, it looped when
2817        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2818        condition was not nested. This bug has been fixed.
2819    
2820    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2821        past the start of the subject in the presence of bytes with the top bit
2822        set, for example "\x8aBCD".
2823    
2824    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2825        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2826    
2827    14. Optimized (?!) to (*FAIL).
2828    
2829    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2830        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2831        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2832        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2833        does: it's just the validity check that is more restrictive.
2834    
2835    16. Inserted checks for integer overflows during escape sequence (backslash)
2836        processing, and also fixed erroneous offset values for syntax errors during
2837        backslash processing.
2838    
2839    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2840        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2841    
2842    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2843        caused an overrun.
2844    
2845    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2846        something other than just ASCII characters) inside a group that had an
2847        unlimited repeat caused a loop at compile time (while checking to see
2848        whether the group could match an empty string).
2849    
2850    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2851        [\P{Any}] did so. (Error in the code for printing property names.)
2852    
2853    21. An orphan \E inside a character class could cause a crash.
2854    
2855    22. A repeated capturing bracket such as (A)? could cause a wild memory
2856        reference during compilation.
2857    
2858    23. There are several functions in pcre_compile() that scan along a compiled
2859        expression for various reasons (e.g. to see if it's fixed length for look
2860        behind). There were bugs in these functions when a repeated \p or \P was
2861        present in the pattern. These operators have additional parameters compared
2862        with \d, etc, and these were not being taken into account when moving along
2863        the compiled data. Specifically:
2864    
2865        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2866            length.
2867    
2868        (b) An item such as \pL+ within a repeated group could cause crashes or
2869            loops.
2870    
2871        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2872            "reference to non-existent subpattern" error.
2873    
2874        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2875    
2876    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2877        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2878    
2879    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2880    
2881    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2882        character were causing crashes (broken optimization).
2883    
2884    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2885        \p or \P) caused a compile-time loop.
2886    
2887    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2888        break. For example, the unstudied pattern [\r\n]A does not match the string
2889        "\r\nA" because change 7.0/46 below moves the current point on by two
2890        characters after failing to match at the start. However, the pattern \nA
2891        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2892        the same is true. There doesn't seem any very clean way out of this, but
2893        what I have chosen to do makes the common cases work: PCRE now takes note
2894        of whether there can be an explicit match for \r or \n anywhere in the
2895        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2896        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2897        pattern has explicit CR or LF references.
2898    
2899    29. Added (*CR) etc for changing newline setting at start of pattern.
2900    
2901    
2902    Version 7.2 19-Jun-07
2903    ---------------------
2904    
2905     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2906        which is apparently normally available under Windows.
2907    
2908     2. Re-jig the pcregrep tests with different newline settings in an attempt
2909        to make them independent of the local environment's newline setting.
2910    
2911     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2912    
2913     4. Some of the "internals" tests were previously cut out when the link size
2914        was not 2, because the output contained actual offsets. The recent new
2915        "Z" feature of pcretest means that these can be cut out, making the tests
2916        usable with all link sizes.
2917    
2918     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2919        stack recursion. This gives a massive performance boost under BSD, but just
2920        a small improvement under Linux. However, it saves one field in the frame
2921        in all cases.
2922    
2923     6. Added more features from the forthcoming Perl 5.10:
2924    
2925        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2926            recursion call. It refers to the nth most recently opened parentheses.
2927    
2928        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2929            to be opened parentheses.
2930    
2931        (c) Conditions that refer to capturing parentheses can be specified
2932            relatively, for example, (?(-2)... or (?(+3)...
2933    
2934        (d) \K resets the start of the current match so that everything before
2935            is not part of it.
2936    
2937        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2938    
2939        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2940            reference syntax.
2941    
2942        (g) (?| introduces a group in which the numbering of parentheses in each
2943            alternative starts with the same number.
2944    
2945        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2946    
2947     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2948        PCRE_INFO_JCHANGED.
2949    
2950     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2951        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2952        for detecting groups that can match an empty string.
2953    
2954     9. A pattern with a very large number of alternatives (more than several
2955        hundred) was running out of internal workspace during the pre-compile
2956        phase, where pcre_compile() figures out how much memory will be needed. A
2957        bit of new cunning has reduced the workspace needed for groups with
2958        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2959        workspace instead of running out of the 4096 that are available.
2960    
2961    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2962    
2963    11. Applied patch from Google to remove an optimization that didn't quite work.
2964        The report of the bug said:
2965    
2966          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2967          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2968          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2969    
2970    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2971        it matched the wrong number of bytes.
2972    
2973    
2974    Version 7.1 24-Apr-07
2975    ---------------------
2976    
2977     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2978        that is more "standard", making use of automake and other Autotools. There
2979        is some re-arrangement of the files and adjustment of comments consequent
2980        on this.
2981    
2982     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2983        for recursive directory scanning broke on some systems because the files
2984        are not scanned in any specific order and on different systems the order
2985        was different. A call to "sort" has been inserted into RunGrepTest for the
2986        approprate test as a short-term fix. In the longer term there may be an
2987        alternative.
2988    
2989     3. I had an email from Eric Raymond about problems translating some of PCRE's
2990        man pages to HTML (despite the fact that I distribute HTML pages, some
2991        people do their own conversions for various reasons). The problems
2992        concerned the use of low-level troff macros .br and .in. I have therefore
2993        removed all such uses from the man pages (some were redundant, some could
2994        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2995        HTML has been updated to handle .nf/.fi and to complain if it encounters
2996        .br or .in.
2997    
2998     4. Updated comments in configure.ac that get placed in config.h.in and also
2999        arranged for config.h to be included in the distribution, with the name
3000        config.h.generic, for the benefit of those who have to compile without
3001        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3002    
3003     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3004        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3005        makevp.bat for the current PCRE, using the additional files
3006        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3007    
3008     6. A Windows user reported a minor discrepancy with test 2, which turned out
3009        to be caused by a trailing space on an input line that had got lost in his
3010        copy. The trailing space was an accident, so I've just removed it.
3011    
3012     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3013        that is needed.
3014    
3015     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3016        as "const" (a) because they are and (b) because it helps the PHP
3017        maintainers who have recently made a script to detect big data structures
3018        in the php code that should be moved to the .rodata section. I remembered
3019        to update Builducptable as well, so it won't revert if ucptable.h is ever
3020        re-created.
3021    
3022     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3023        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3024        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3025        support is not required. This saves 1.5-2K of code, which is important in
3026        some applications.
3027    
3028        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3029        so as not to refer to the tables, even though these functions will never be
3030        called when UTF-8 support is disabled. Otherwise there are problems with a
3031        shared library.
3032    
3033    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3034    
3035        (a) It was defining its arguments as char * instead of void *.
3036    
3037        (b) It was assuming that all moves were upwards in memory; this was true
3038            a long time ago when I wrote it, but is no longer the case.
3039    
3040        The emulated memove() is provided for those environments that have neither
3041        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3042        is clearly not the case, as these two bugs were recently reported.
3043    
3044    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3045        and Detrail to create the HTML documentation, the .txt form of the man
3046        pages, and it removes trailing spaces from listed files. It also creates
3047        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3048        case, it wraps all the #defines with #ifndefs. This script should be run
3049        before "make dist".
3050    
3051    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3052        with Unicode property support.
3053    
3054        (a) For a maximizing quantifier, if the two different cases of the
3055            character were of different lengths in their UTF-8 codings (there are
3056            some cases like this - I found 11), and the matching function had to
3057            back up over a mixture of the two cases, it incorrectly assumed they
3058            were both the same length.
3059    
3060        (b) When PCRE was configured to use the heap rather than the stack for
3061            recursion during matching, it was not correctly preserving the data for
3062            the other case of a UTF-8 character when checking ahead for a match
3063            while processing a minimizing repeat. If the check also involved
3064            matching a wide character, but failed, corruption could cause an
3065            erroneous result when trying to check for a repeat of the original
3066            character.
3067    
3068    13. Some tidying changes to the testing mechanism:
3069    
3070        (a) The RunTest script now detects the internal link size and whether there
3071            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3072            values substituted by "configure". (The RunGrepTest script already did
3073            this for UTF-8.) The configure.ac script no longer substitutes the
3074            relevant variables.
3075    
3076        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3077            with length and offset values. This means that the output is different
3078            for different internal link sizes. Test 2 is skipped for link sizes
3079            other than 2 because of this, bypassing the problem. Unfortunately,
3080            there was also a test in test 3 (the locale tests) that used /B and
3081            failed for link sizes other than 2. Rather than cut the whole test out,
3082            I have added a new /Z option to pcretest that replaces the length and
3083            offset values with spaces. This is now used to make test 3 independent
3084            of link size. (Test 2 will be tidied up later.)
3085    
3086    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3087        segmentation fault instead of returning the appropriate error message.
3088    
3089    15. In multiline mode when the newline sequence was set to "any", the pattern
3090        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3091        This doesn't seem right; it now treats the CRLF combination as the line
3092        ending, and so does not match in that case. It's only a pattern such as ^$
3093        that would hit this one: something like ^ABC$ would have failed after \r
3094        and then tried again after \r\n.
3095    
3096    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3097        in an attempt to make files that differ only in their line terminators
3098        compare equal. This works on Linux.
3099    
3100    17. Under certain error circumstances pcregrep might try to free random memory
3101        as it exited. This is now fixed, thanks to valgrind.
3102    
3103    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3104        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3105        was because its rules for how to advance for /g after matching an empty
3106        string at the end of a line did not allow for this case. They now check for
3107        it specially.
3108    
3109    20. pcretest is supposed to handle patterns and data of any length, by
3110        extending its buffers when necessary. It was getting this wrong when the
3111        buffer for a data line had to be extended.
3112    
3113    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3114        CRLF as a newline sequence.
3115    
3116    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3117        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3118        I have nevertheless tidied it up.
3119    
3120    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3121    
3122    24. Added a man page for pcre-config.
3123    
3124    
3125    Version 7.0 19-Dec-06
3126    ---------------------
3127    
3128     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3129        moving to gcc 4.1.1.
3130    
3131     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3132        sys/time.h, which is documented as needed for this function. It doesn't
3133        seem to matter on Linux, but it showed up on some releases of OS X.
3134    
3135     3. It seems that there are systems where bytes whose values are greater than
3136        127 match isprint() in the "C" locale. The "C" locale should be the
3137        default when a C program starts up. In most systems, only ASCII printing
3138        characters match isprint(). This difference caused the output from pcretest
3139        to vary, making some of the tests fail. I have changed pcretest so that:
3140    
3141        (a) When it is outputting text in the compiled version of a pattern, bytes
3142            other than 32-126 are always shown as hex escapes.
3143    
3144        (b) When it is outputting text that is a matched part of a subject string,
3145            it does the same, unless a different locale has been set for the match
3146            (using the /L modifier). In this case, it uses isprint() to decide.
3147    
3148     4. Fixed a major bug that caused incorrect computation of the amount of memory
3149        required for a compiled pattern when options that changed within the
3150        pattern affected the logic of the preliminary scan that determines the
3151        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3152        that the computed length was too small. The symptoms of this bug were
3153        either the PCRE error "internal error: code overflow" from pcre_compile(),
3154        or a glibc crash with a message such as "pcretest: free(): invalid next
3155        size (fast)". Examples of patterns that provoked this bug (shown in
3156        pcretest format) are:
3157    
3158          /(?-x: )/x
3159          /(?x)(?-x: \s*#\s*)/
3160          /((?i)[\x{c0}])/8
3161          /(?i:[\x{c0}])/8
3162    
3163        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3164        is now done differently.
3165    
3166     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3167        wrapper classes; (b) implement a new function in the C++ scanner that is
3168        more efficient than the old way of doing things because it avoids levels of
3169        recursion in the regex matching; (c) add a paragraph to the documentation
3170        for the FullMatch() function.
3171    
3172     6. The escape sequence \n was being treated as whatever was defined as
3173        "newline". Not only was this contrary to the documentation, which states
3174        that \n is character 10 (hex 0A), but it also went horribly wrong when
3175        "newline" was defined as CRLF. This has been fixed.
3176    
3177     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3178        was being set to -1 for the "end of line" case (supposedly a value that no
3179        character can have). Though this value is never used (the check for end of
3180        line is "zero bytes in current character"), it caused compiler complaints.
3181        I've changed it to 0xffffffff.
3182    
3183     8. In pcre_version.c, the version string was being built by a sequence of
3184        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3185        string (as it is for production releases) called a macro with an empty
3186        argument. The C standard says the result of this is undefined. The gcc
3187        compiler treats it as an empty string (which was what was wanted) but it is
3188        reported that Visual C gives an error. The source has been hacked around to
3189        avoid this problem.
3190    
3191     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3192        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3193        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3194        of them did).
3195    
3196    10. Originally, pcretest opened its input and output without "b"; then I was
3197        told that "b" was needed in some environments, so it was added for release
3198        5.0 to both the input and output. (It makes no difference on Unix-like
3199        systems.) Later I was told that it is wrong for the input on Windows. I've
3200        now abstracted the modes into two macros, to make it easier to fiddle with
3201        them, and removed "b" from the input mode under Windows.
3202    
3203    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3204    
3205    12. Added -help and --help to pcretest as an official way of being reminded
3206        of the options.
3207    
3208    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3209        and pcrecpp.cc because they annoy compilers at high warning levels.
3210    
3211    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3212    
3213    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3214        scripts are not C programs :-) and which was not noticed because it works
3215        on Linux.
3216    
3217    16. pcretest is supposed to handle any length of pattern and data line (as one
3218        line or as a continued sequence of lines) by extending its input buffer if
3219        necessary. This feature was broken for very long pattern lines, leading to
3220        a string of junk being passed to pcre_compile() if the pattern was longer
3221        than about 50K.
3222    
3223    17. I have done a major re-factoring of the way pcre_compile() computes the
3224        amount of memory needed for a compiled pattern. Previously, there was code
3225        that made a preliminary scan of the pattern in order to do this. That was
3226        OK when PCRE was new, but as the facilities have expanded, it has become
3227        harder and harder to keep it in step with the real compile phase, and there
3228        have been a number of bugs (see for example, 4 above). I have now found a
3229        cunning way of running the real compile function in a "fake" mode that
3230        enables it to compute how much memory it would need, while actually only
3231        ever using a few hundred bytes of working memory and without too many
3232        tests of the mode. This should make future maintenance and development
3233        easier. A side effect of this work is that the limit of 200 on the nesting
3234        depth of parentheses has been removed (though this was never a serious
3235        limitation, I suspect). However, there is a downside: pcre_compile() now
3236        runs more slowly than before (30% or more, depending on the pattern). I
3237        hope this isn't a big issue. There is no effect on runtime performance.
3238    
3239    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3240        newline (only possible for the last line of a file) and it was a
3241        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3242    
3243    19. Added additional timing features to pcretest. (1) The -tm option now times
3244        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3245        separate command line item, by a number that specifies the number of
3246        repeats to use when timing. The default is 50000; this gives better
3247        precision, but takes uncomfortably long for very large patterns.
3248    
3249    20. Extended pcre_study() to be more clever in cases where a branch of a
3250        subpattern has no definite first character. For example, (a*|b*)[cd] would
3251        previously give no result from pcre_study(). Now it recognizes that the
3252        first character must be a, b, c, or d.
3253    
3254    21. There was an incorrect error "recursive call could loop indefinitely" if
3255        a subpattern (or the entire pattern) that was being tested for matching an
3256        empty string contained only one non-empty item after a nested subpattern.
3257        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3258        incorrectly, because the \d was being skipped in the check.
3259    
3260    22. The pcretest program now has a new pattern option /B and a command line
3261        option -b, which is equivalent to adding /B to every pattern. This causes
3262        it to show the compiled bytecode, without the additional information that
3263        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3264        is the same as /B/I).
3265    
3266    23. A new optimization is now able automatically to treat some sequences such
3267        as a*b as a*+b. More specifically, if something simple (such as a character
3268        or a simple class like \d) has an unlimited quantifier, and is followed by
3269        something that cannot possibly match the quantified thing, the quantifier
3270        is automatically "possessified".
3271    
3272    24. A recursive reference to a subpattern whose number was greater than 39
3273        went wrong under certain circumstances in UTF-8 mode. This bug could also
3274        have affected the operation of pcre_study().
3275    
3276    25. Realized that a little bit of performance could be had by replacing
3277        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3278    
3279    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3280    
3281    27. Possessive quantifiers such as a++ were previously implemented by turning
3282        them into atomic groups such as ($>a+). Now they have their own opcodes,
3283        which improves performance. This includes the automatically created ones
3284        from 23 above.
3285    
3286    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3287        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3288        the first matched character to be a colon. This applied both to named and
3289        numbered groups.
3290    
3291    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3292    
3293    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3294        building PCRE on Windows easier, so I have included it in the distribution.
3295    
3296    31. There is now a check in pcretest against a ridiculously large number being
3297        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3298        loop, the loop is abandoned.
3299    
3300    32. Forward references to subpatterns in conditions such as (?(2)...) where
3301        subpattern 2 is defined later cause pcre_compile() to search forwards in
3302        the pattern for the relevant set of parentheses. This search went wrong
3303        when there were unescaped parentheses in a character class, parentheses
3304        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3305    
3306    33. "Subroutine" calls and backreferences were previously restricted to
3307        referencing subpatterns earlier in the regex. This restriction has now
3308        been removed.
3309    
3310    34. Added a number of extra features that are going to be in Perl 5.10. On the
3311        whole, these are just syntactic alternatives for features that PCRE had
3312        previously implemented using the Python syntax or my own invention. The
3313        other formats are all retained for compatibility.
3314    
3315        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3316            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3317            also .NET compatible.
3318    
3319        (b) A recursion or subroutine call to a named group can now be defined as
3320            (?&name) as well as (?P>name).
3321    
3322        (c) A backreference to a named group can now be defined as \k<name> or
3323            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3324            5.10, are also .NET compatible.
3325    
3326        (d) A conditional reference to a named group can now use the syntax
3327            (?(<name>) or (?('name') as well as (?(name).
3328    
3329        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3330            groups (named and numbered) that are never evaluated inline, but can be
3331            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3332            is always false. There may be only one alternative in such a group.
3333    
3334        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3335            as the simple (?(R). The condition is true only if the most recent
3336            recursion is that of the given number or name. It does not search out
3337            through the entire recursion stack.
3338    
3339        (g) The escape \gN or \g{N} has been added, where N is a positive or
3340            negative number, specifying an absolute or relative reference.
3341    
3342    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3343        some "unreachable code" warnings.
3344    
3345    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3346        things, this adds five new scripts.
3347    
3348    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3349        There were also incompatibilities regarding the handling of \Q..\E inside
3350        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3351        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3352    
3353    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3354        matches an empty string, and forcibly breaks the loop. There were bugs in
3355        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3356        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3357        separate and independent bugs (that affected different cases) have been
3358        fixed.
3359    
3360    39. Refactored the code to abolish the use of different opcodes for small
3361        capturing bracket numbers. This is a tidy that I avoided doing when I
3362        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3363        The new approach is not only tidier, it makes it possible to reduce the
3364        memory needed to fix the previous bug (38).
3365    
3366    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3367        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3368        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3369        mode.
3370    
3371    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3372        report.
3373    
3374    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3375        copy construction and assignment in the C++ wrapper.
3376    
3377    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3378        couple of bugs that could have given wrong results in the "--newline=crlf"
3379        case.
3380    
3381    44. Added a number of casts and did some reorganization of signed/unsigned int
3382        variables following suggestions from Dair Grant. Also renamed the variable
3383        "this" as "item" because it is a C++ keyword.
3384    
3385    45. Arranged for dftables to add
3386    
3387          #include "pcre_internal.h"
3388    
3389        to pcre_chartables.c because without it, gcc 4.x may remove the array
3390        definition from the final binary if PCRE is built into a static library and
3391        dead code stripping is activated.
3392    
3393    46. For an unanchored pattern, if a match attempt fails at the start of a
3394        newline sequence, and the newline setting is CRLF or ANY, and the next two
3395        characters are CRLF, advance by two characters instead of one.
3396    
3397    
3398    Version 6.7 04-Jul-06
3399    ---------------------
3400    
3401     1. In order to handle tests when input lines are enormously long, pcretest has
3402        been re-factored so that it automatically extends its buffers when
3403        necessary. The code is crude, but this _is_ just a test program. The
3404        default size has been increased from 32K to 50K.
3405    
3406     2. The code in pcre_study() was using the value of the re argument before
3407        testing it for NULL. (Of course, in any sensible call of the function, it
3408        won't be NULL.)
3409    
3410     3. The memmove() emulation function in pcre_internal.h, which is used on
3411        systems that lack both memmove() and bcopy() - that is, hardly ever -
3412        was missing a "static" storage class specifier.
3413    
3414     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3415        containing an extended class (one that cannot be represented by a bitmap
3416        because it contains high-valued characters or Unicode property items, e.g.
3417        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3418        pattern, but PCRE should not loop if you do not (it no longer does).
3419        [Detail: two cases were found: (a) a repeated subpattern containing an
3420        extended class; (b) a recursive reference to a subpattern that followed a
3421        previous extended class. It wasn't skipping over the extended class
3422        correctly when UTF-8 mode was not set.]
3423    
3424     5. A negated single-character class was not being recognized as fixed-length
3425        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3426        compile error "lookbehind assertion is not fixed length".
3427    
3428     6. The RunPerlTest auxiliary script was showing an unexpected difference
3429        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3430        write a Perl script that can interpret lines of an input file either as
3431        byte characters or as UTF-8, which is what "perltest" was being required to
3432        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3433        can't do is switch easily at run time between having the "use utf8;" pragma
3434        or not. In the end, I fudged it by using the RunPerlTest script to insert
3435        "use utf8;" explicitly for the UTF-8 tests.
3436    
3437     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3438        the end of the subject string, contrary to the documentation and to what
3439        Perl does. This was true of both matching functions. Now it matches only at
3440        the start of the subject and immediately after *internal* newlines.
3441    
3442     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3443        a pointer to an int instead of a pointer to an unsigned long int. This
3444        caused problems on 64-bit systems.
3445    
3446     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3447        instance of the 'standard' template library not being so standard".
3448    
3449    10. There was no check on the number of named subpatterns nor the maximum
3450        length of a subpattern name. The product of these values is used to compute
3451        the size of the memory block for a compiled pattern. By supplying a very
3452        long subpattern name and a large number of named subpatterns, the size
3453        computation could be caused to overflow. This is now prevented by limiting
3454        the length of names to 32 characters, and the number of named subpatterns
3455        to 10,000.
3456    
3457    11. Subpatterns that are repeated with specific counts have to be replicated in
3458        the compiled pattern. The size of memory for this was computed from the
3459        length of the subpattern and the repeat count. The latter is limited to
3460        65535, but there was no limit on the former, meaning that integer overflow
3461        could in principle occur. The compiled length of a repeated subpattern is
3462        now limited to 30,000 bytes in order to prevent this.
3463    
3464    12. Added the optional facility to have named substrings with the same name.
3465    
3466    13. Added the ability to use a named substring as a condition, using the
3467        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3468        are numbers (not recommended). Forward references are permitted.
3469    
3470    14. Added forward references in named backreferences (if you see what I mean).
3471    
3472    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3473        pattern could run off the end of the subject. For example, the pattern
3474        "(?s)(.{1,5})"8 did this with the subject "ab".
3475    
3476    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3477        PCRE_CASELESS was set when matching characters that were quantified with ?
3478        or *.
3479    
3480    17. A character class other than a single negated character that had a minimum
3481        but no maximum quantifier - for example [ab]{6,} - was not handled
3482        correctly by pce_dfa_exec(). It would match only one character.
3483    
3484    18. A valid (though odd) pattern that looked like a POSIX character
3485        class but used an invalid character after [ (for example [[,abc,]]) caused
3486        pcre_compile() to give the error "Failed: internal error: code overflow" or
3487        in some cases to crash with a glibc free() error. This could even happen if
3488        the pattern terminated after [[ but there just happened to be a sequence of
3489        letters, a binary zero, and a closing ] in the memory that followed.
3490    
3491    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3492        over the years. Originally (before any Unicode support), just the bottom 8
3493        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3494        output from "man perlunicode" includes this:
3495    
3496          The regular expression compiler produces polymorphic opcodes.  That
3497          is, the pattern adapts to the data and automatically switches to
3498          the Unicode character scheme when presented with Unicode data--or
3499          instead uses a traditional byte scheme when presented with byte
3500          data.
3501    
3502        Sadly, a wide octal escape does not cause a switch, and in a string with
3503        no other multibyte characters, these octal escapes are treated as before.
3504        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3505        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3506        Unicode string.
3507    
3508        I have not perpetrated such confusion in PCRE. Up till now, it took just
3509        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3510        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3511        translate to the appropriate multibyte character.
3512    
3513    29. Applied some refactoring to reduce the number of warnings from Microsoft
3514        and Borland compilers. This has included removing the fudge introduced
3515        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3516        a warning about an unused variable.
3517    
3518    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3519        characters since release 4.0, because Perl (from release 5.004) does not.
3520        [Or at least, is documented not to: some releases seem to be in conflict
3521        with the documentation.] However, when a pattern was studied with
3522        pcre_study() and all its branches started with \s, PCRE still included VT
3523        as a possible starting character. Of course, this did no harm; it just
3524        caused an unnecessary match attempt.
3525    
3526    22. Removed a now-redundant internal flag bit that recorded the fact that case
3527        dependency changed within the pattern. This was once needed for "required
3528        byte" processing, but is no longer used. This recovers a now-scarce options
3529        bit. Also moved the least significant internal flag bit to the most-
3530        significant bit of the word, which was not previously used (hangover from
3531        the days when it was an int rather than a uint) to free up another bit for
3532        the future.