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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1535 by ph10, Wed Mar 25 16:51:51 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Version 8.37 xx-xxx-2015
5    ------------------------
6    
7    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8        for those parentheses to be closed with whatever has been captured so far.
9        However, it was failing to mark any other groups between the hightest
10        capture so far and the currrent group as "unset". Thus, the ovector for
11        those groups contained whatever was previously there. An example is the
12        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13    
14    2.  If an assertion condition was quantified with a minimum of zero (an odd
15        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16    
17    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
18        unrecognized modifier, a crash could occur.
19    
20    4.  An attempt to do global matching in pcretest with a zero-length ovector
21        caused a crash.
22    
23    5.  Fixed a memory leak during matching that could occur for a subpattern
24        subroutine call (recursive or otherwise) if the number of captured groups
25        that had to be saved was greater than ten.
26    
27    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
28        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30    
31    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
32        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33    
34    8. When a pattern is compiled, it remembers the highest back reference so that
35       when matching, if the ovector is too small, extra memory can be obtained to
36       use instead. A conditional subpattern whose condition is a check on a
37       capture having happened, such as, for example in the pattern
38       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39       setting the highest backreference number. This mattered only if pcre_exec()
40       was called with an ovector that was too small to hold the capture, and there
41       was no other kind of back reference (a situation which is probably quite
42       rare). The effect of the bug was that the condition was always treated as
43       FALSE when the capture could not be consulted, leading to a incorrect
44       behaviour by pcre2_match(). This bug has been fixed.
45    
46    9. A reference to a duplicated named group (either a back reference or a test
47       for being set in a conditional) that occurred in a part of the pattern where
48       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49       to be incorrectly calculated, leading to overwriting.
50    
51    10. A mutually recursive set of back references such as (\2)(\1) caused a
52        segfault at study time (while trying to find the minimum matching length).
53        The infinite loop is now broken (with the minimum length unset, that is,
54        zero).
55    
56    11. If an assertion that was used as a condition was quantified with a minimum
57        of zero, matching went wrong. In particular, if the whole group had
58        unlimited repetition and could match an empty string, a segfault was
59        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60        assertions to be quantified, but not if they are being used as conditions,
61        so the above pattern is faulted by Perl. PCRE has now been changed so that
62        it also rejects such patterns.
63    
64    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65        failed to allow the zero-repeat case if pcre2_exec() was called with an
66        ovector too small to capture the group.
67    
68    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69        Red Hat Product Security:
70    
71        (a) A crash if /K and /F were both set with the option to save the compiled
72        pattern.
73    
74        (b) Another crash if the option to print captured substrings in a callout
75        was combined with setting a null ovector, for example \O\C+ as a subject
76        string.
77    
78    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79        forward reference repeated a large (but limited) number of times within a
80        repeated outer group that has a zero minimum quantifier, caused incorrect
81        code to be compiled, leading to the error "internal error:
82        previously-checked referenced subpattern not found" when an incorrect
83        memory address was read. This bug was reported as "heap overflow",
84        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
85        CVE-2015-2325.
86    
87    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
88        call within a group that also contained a recursive back reference caused
89        incorrect code to be compiled. This bug was reported as "heap overflow",
90        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
91        number CVE-2015-2326.
92    
93    24. Computing the size of the JIT read-only data in advance has been a source
94        of various issues, and new ones are still appear unfortunately. To fix
95        existing and future issues, size computation is eliminated from the code,
96        and replaced by on-demand memory allocation.
97    
98    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
99        adjacent to the end of the range, and the range contained characters with
100        more than one other case, caused incorrect behaviour when compiled in UTF
101        mode. In that example, the range a-j was left out of the class.
102    
103    26. Fix JIT compilation of conditional blocks, which assertion
104        is converted to (*FAIL). E.g: /(?(?!))/.
105    
106    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
107        discovered by the LLVM fuzzer.
108    
109    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
110        when this assertion was used as a condition, for example (?(?!)a|b). In
111        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
112        error about an unsupported item.
113    
114    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
115        possessification code could take exponential time to complete. A recursion
116        depth limit of 1000 has been imposed to limit the resources used by this
117        optimization.
118    
119    
120    Version 8.36 26-September-2014
121    ------------------------------
122    
123    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
124        -Wmissing-field-initializers and -Wunused-parameter.
125    
126    2.  The tests for quantifiers being too big (greater than 65535) were being
127        applied after reading the number, and stupidly assuming that integer
128        overflow would give a negative number. The tests are now applied as the
129        numbers are read.
130    
131    3.  Tidy code in pcre_exec.c where two branches that used to be different are
132        now the same.
133    
134    4.  The JIT compiler did not generate match limit checks for certain
135        bracketed expressions with quantifiers. This may lead to exponential
136        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
137        issue should be resolved now.
138    
139    5.  Fixed an issue, which occures when nested alternatives are optimized
140        with table jumps.
141    
142    6.  Inserted two casts and changed some ints to size_t in the light of some
143        reported 64-bit compiler warnings (Bugzilla 1477).
144    
145    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
146        an empty string, which sometimes were behaving incorrectly in the
147        interpreter (though correctly in the JIT matcher). This pcretest input is
148        an example:
149    
150          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
151          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
152    
153        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
154        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
155        for an empty string was breaking the inner loop and carrying on at a lower
156        level, when possessive repeated groups should always return to a higher
157        level as they have no backtrack points in them. The empty string test now
158        occurs at the outer level.
159    
160    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
161        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
162    
163    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
164        doesn't).
165    
166    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
167        to hold a pattern to be incorrectly computed (too small) when there were
168        named back references to duplicated names. This could cause "internal
169        error: code overflow" or "double free or corruption" or other memory
170        handling errors.
171    
172    11. When named subpatterns had the same prefixes, back references could be
173        confused. For example, in this pattern:
174    
175          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
176    
177        the reference to 'Name' was incorrectly treated as a reference to a
178        duplicate name.
179    
180    12. A pattern such as /^s?c/mi8 where the optional character has more than
181        one "other case" was incorrectly compiled such that it would only try to
182        match starting at "c".
183    
184    13. When a pattern starting with \s was studied, VT was not included in the
185        list of possible starting characters; this should have been part of the
186        8.34/18 patch.
187    
188    14. If a character class started [\Qx]... where x is any character, the class
189        was incorrectly terminated at the ].
190    
191    15. If a pattern that started with a caseless match for a character with more
192        than one "other case" was studied, PCRE did not set up the starting code
193        unit bit map for the list of possible characters. Now it does. This is an
194        optimization improvement, not a bug fix.
195    
196    16. The Unicode data tables have been updated to Unicode 7.0.0.
197    
198    17. Fixed a number of memory leaks in pcregrep.
199    
200    18. Avoid a compiler warning (from some compilers) for a function call with
201        a cast that removes "const" from an lvalue by using an intermediate
202        variable (to which the compiler does not object).
203    
204    19. Incorrect code was compiled if a group that contained an internal recursive
205        back reference was optional (had quantifier with a minimum of zero). This
206        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
207        caused segmentation faults because of stack overflows at compile time.
208    
209    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
210        group that is quantified with an indefinite repeat, caused a compile-time
211        loop which used up all the system stack and provoked a segmentation fault.
212        This was not the same bug as 19 above.
213    
214    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
215        Patch by Mike Frysinger.
216    
217    
218    Version 8.35 04-April-2014
219    --------------------------
220    
221    1.  A new flag is set, when property checks are present in an XCLASS.
222        When this flag is not set, PCRE can perform certain optimizations
223        such as studying these XCLASS-es.
224    
225    2.  The auto-possessification of character sets were improved: a normal
226        and an extended character set can be compared now. Furthermore
227        the JIT compiler optimizes more character set checks.
228    
229    3.  Got rid of some compiler warnings for potentially uninitialized variables
230        that show up only when compiled with -O2.
231    
232    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
233        of a match later then the end of the match. The pcretest program was not
234        handling the case sensibly - it was outputting from the start to the next
235        binary zero. It now reports this situation in a message, and outputs the
236        text from the end to the start.
237    
238    5.  Fast forward search is improved in JIT. Instead of the first three
239        characters, any three characters with fixed position can be searched.
240        Search order: first, last, middle.
241    
242    6.  Improve character range checks in JIT. Characters are read by an inprecise
243        function now, which returns with an unknown value if the character code is
244        above a certain threshold (e.g: 256). The only limitation is that the value
245        must be bigger than the threshold as well. This function is useful when
246        the characters above the threshold are handled in the same way.
247    
248    7.  The macros whose names start with RAWUCHAR are placeholders for a future
249        mode in which only the bottom 21 bits of 32-bit data items are used. To
250        make this more memorable for those maintaining the code, the names have
251        been changed to start with UCHAR21, and an extensive comment has been added
252        to their definition.
253    
254    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
255        to the export list in Makefile.am (they were accidentally omitted from the
256        8.34 tarball).
257    
258    9.  The informational output from pcretest used the phrase "starting byte set"
259        which is inappropriate for the 16-bit and 32-bit libraries. As the output
260        for "first char" and "need char" really means "non-UTF-char", I've changed
261        "byte" to "char", and slightly reworded the output. The documentation about
262        these values has also been (I hope) clarified.
263    
264    10. Another JIT related optimization: use table jumps for selecting the correct
265        backtracking path, when more than four alternatives are present inside a
266        bracket.
267    
268    11. Empty match is not possible, when the minimum length is greater than zero,
269        and there is no \K in the pattern. JIT should avoid empty match checks in
270        such cases.
271    
272    12. In a caseless character class with UCP support, when a character with more
273        than one alternative case was not the first character of a range, not all
274        the alternative cases were added to the class. For example, s and \x{17f}
275        are both alternative cases for S: the class [RST] was handled correctly,
276        but [R-T] was not.
277    
278    13. The configure.ac file always checked for pthread support when JIT was
279        enabled. This is not used in Windows, so I have put this test inside a
280        check for the presence of windows.h (which was already tested for).
281    
282    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
283        The algorithm provides a way to skip certain starting offsets, and usually
284        faster than linear prefix searches.
285    
286    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
287        as for 'fr_FR' and 'french'. For some reason, however, it then used the
288        Windows-specific input and output files, which have 'french' screwed in.
289        So this could never have worked. One of the problems with locales is that
290        they aren't always the same. I have now updated RunTest so that it checks
291        the output of the locale test (test 3) against three different output
292        files, and it allows the test to pass if any one of them matches. With luck
293        this should make the test pass on some versions of Solaris where it was
294        failing. Because of the uncertainty, the script did not used to stop if
295        test 3 failed; it now does. If further versions of a French locale ever
296        come to light, they can now easily be added.
297    
298    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
299        there was a message during ./configure, but it did not stop. This now
300        provokes an error. The invalid example in README has been corrected.
301        If a value less than the minimum is given, the minimum value has always
302        been used, but now a warning is given.
303    
304    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
305        was a bug in the test system, which is now fixed. Also, the list of various
306        configurations that are tested for each release did not have one with both
307        16/32 bits and --enable-bar-anycrlf. It now does.
308    
309    18. pcretest was missing "-C bsr" for displaying the \R default setting.
310    
311    19. Little endian PowerPC systems are supported now by the JIT compiler.
312    
313    20. The fast forward newline mechanism could enter to an infinite loop on
314        certain invalid UTF-8 input. Although we don't support these cases
315        this issue can be fixed by a performance optimization.
316    
317    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
318        not take account if existing stack usage. There is now a new global
319        variable called pcre_stack_guard that can be set to point to an external
320        function to check stack availability. It is called at the start of
321        processing every parenthesized group.
322    
323    22. A typo in the code meant that in ungreedy mode the max/min qualifier
324        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
325        not match "ab".
326    
327    23. When UTF was disabled, the JIT program reported some incorrect compile
328        errors. These messages are silenced now.
329    
330    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
331        compiler.
332    
333    25. Change all the temporary files used in RunGrepTest to be different to those
334        used by RunTest so that the tests can be run simultaneously, for example by
335        "make -j check".
336    
337    
338    Version 8.34 15-December-2013
339    -----------------------------
340    
341    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
342        executable memory. Patch inspired by Carsten Klein.
343    
344    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
345        this macro is never tested and has no effect, because the work to support
346        coverage involves only compiling and linking options and special targets in
347        the Makefile. The comment in config.h implied that defining the macro would
348        enable coverage support, which is totally false. There was also support for
349        setting this macro in the CMake files (my fault, I just copied it from
350        configure). SUPPORT_GCOV has now been removed.
351    
352    3.  Make a small performance improvement in strlen16() and strlen32() in
353        pcretest.
354    
355    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
356        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
357        these cases). There was also one in pcretest.c.
358    
359    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
360    
361    6.  In UTF mode, the code for checking whether a group could match an empty
362        string (which is used for indefinitely repeated groups to allow for
363        breaking an infinite loop) was broken when the group contained a repeated
364        negated single-character class with a character that occupied more than one
365        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
366        UTF-8 mode). The effect was undefined: the group might or might not be
367        deemed as matching an empty string, or the program might have crashed.
368    
369    7.  The code for checking whether a group could match an empty string was not
370        recognizing that \h, \H, \v, \V, and \R must match a character.
371    
372    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
373        an empty string. If it can, pcretest shows this in its information output.
374    
375    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
376        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
377        matched by pcre_exec() without using JIT:
378    
379        (a) If the rest of the pattern did not match after a maximal run of
380            grapheme clusters, the code for backing up to try with fewer of them
381            did not always back up over a full grapheme when characters that do not
382            have the modifier quality were involved, e.g. Hangul syllables.
383    
384        (b) If the match point in a subject started with modifier character, and
385            there was no match, the code could incorrectly back up beyond the match
386            point, and potentially beyond the first character in the subject,
387            leading to a segfault or an incorrect match result.
388    
389    10. A conditional group with an assertion condition could lead to PCRE
390        recording an incorrect first data item for a match if no other first data
391        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
392        first data item, and therefore matched "ca" after "c" instead of at the
393        start.
394    
395    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
396        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
397    
398    12. The source of pcregrep now includes z/OS-specific code so that it can be
399        compiled for z/OS as part of the special z/OS distribution.
400    
401    13. Added the -T and -TM options to pcretest.
402    
403    14. The code in pcre_compile.c for creating the table of named capturing groups
404        has been refactored. Instead of creating the table dynamically during the
405        actual compiling pass, the information is remembered during the pre-compile
406        pass (on the stack unless there are more than 20 named groups, in which
407        case malloc() is used) and the whole table is created before the actual
408        compile happens. This has simplified the code (it is now nearly 150 lines
409        shorter) and prepared the way for better handling of references to groups
410        with duplicate names.
411    
412    15. A back reference to a named subpattern when there is more than one of the
413        same name now checks them in the order in which they appear in the pattern.
414        The first one that is set is used for the reference. Previously only the
415        first one was inspected. This change makes PCRE more compatible with Perl.
416    
417    16. Unicode character properties were updated from Unicode 6.3.0.
418    
419    17. The compile-time code for auto-possessification has been refactored, based
420        on a patch by Zoltan Herczeg. It now happens after instead of during
421        compilation. The code is cleaner, and more cases are handled. The option
422        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
423        options in pcretest are provided to set it. It can also be set by
424        (*NO_AUTO_POSSESS) at the start of a pattern.
425    
426    18. The character VT has been added to the default ("C" locale) set of
427        characters that match \s and are generally treated as white space,
428        following this same change in Perl 5.18. There is now no difference between
429        "Perl space" and "POSIX space". Whether VT is treated as white space in
430        other locales depends on the locale.
431    
432    19. The code for checking named groups as conditions, either for being set or
433        for being recursed, has been refactored (this is related to 14 and 15
434        above). Processing unduplicated named groups should now be as fast at
435        numerical groups, and processing duplicated groups should be faster than
436        before.
437    
438    20. Two patches to the CMake build system, by Alexander Barkov:
439    
440          (1) Replace the "source" command by "." in CMakeLists.txt because
441              "source" is a bash-ism.
442    
443          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
444              without these the CMake build does not work on Solaris.
445    
446    21. Perl has changed its handling of \8 and \9. If there is no previously
447        encountered capturing group of those numbers, they are treated as the
448        literal characters 8 and 9 instead of a binary zero followed by the
449        literals. PCRE now does the same.
450    
451    22. Following Perl, added \o{} to specify codepoints in octal, making it
452        possible to specify values greater than 0777 and also making them
453        unambiguous.
454    
455    23. Perl now gives an error for missing closing braces after \x{... instead of
456        treating the string as literal. PCRE now does the same.
457    
458    24. RunTest used to grumble if an inappropriate test was selected explicitly,
459        but just skip it when running all tests. This make it awkward to run ranges
460        of tests when one of them was inappropriate. Now it just skips any
461        inappropriate tests, as it always did when running all tests.
462    
463    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
464        character types such as \d or \w, too many callouts were inserted, and the
465        data that they returned was rubbish.
466    
467    26. In UCP mode, \s was not matching two of the characters that Perl matches,
468        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
469        were matched by \h. The code has now been refactored so that the lists of
470        the horizontal and vertical whitespace characters used for \h and \v (which
471        are defined only in one place) are now also used for \s.
472    
473    27. Add JIT support for the 64 bit TileGX architecture.
474        Patch by Jiong Wang (Tilera Corporation).
475    
476    28. Possessive quantifiers for classes (both explicit and automatically
477        generated) now use special opcodes instead of wrapping in ONCE brackets.
478    
479    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
480        (because it's meaningless), this was not happening when PCRE_CASELESS was
481        set. Not wrong, but inefficient.
482    
483    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
484        properties for \w, \d, etc) is present in a test regex. Otherwise if the
485        test contains no characters greater than 255, Perl doesn't realise it
486        should be using Unicode semantics.
487    
488    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
489        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
490        does in Unicode mode.
491    
492    32. Added the "forbid" facility to pcretest so that putting tests into the
493        wrong test files can sometimes be quickly detected.
494    
495    33. There is now a limit (default 250) on the depth of nesting of parentheses.
496        This limit is imposed to control the amount of system stack used at compile
497        time. It can be changed at build time by --with-parens-nest-limit=xxx or
498        the equivalent in CMake.
499    
500    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
501        errors. Perl warns for these when in warning mode, but PCRE has no facility
502        for giving warnings.
503    
504    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
505        However, this was not working for (?!) because it is optimized to (*FAIL),
506        for which PCRE does not allow quantifiers. The optimization is now disabled
507        when a quantifier follows (?!). I can't see any use for this, but it makes
508        things uniform.
509    
510    36. Perl no longer allows group names to start with digits, so I have made this
511        change also in PCRE. It simplifies the code a bit.
512    
513    37. In extended mode, Perl ignores spaces before a + that indicates a
514        possessive quantifier. PCRE allowed a space before the quantifier, but not
515        before the possessive +. It now does.
516    
517    38. The use of \K (reset reported match start) within a repeated possessive
518        group such as (a\Kb)*+ was not working.
519    
520    40. Document that the same character tables must be used at compile time and
521        run time, and that the facility to pass tables to pcre_exec() and
522        pcre_dfa_exec() is for use only with saved/restored patterns.
523    
524    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
525        features for Builds with MSVC:
526    
527        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
528           builds). The .rc files can be used to set FileDescription and many other
529           attributes.
530    
531        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
532           This allows higher-level build scripts which want .pdb files to avoid
533           hard-coding the exact files needed."
534    
535    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
536        mean "start of word" and "end of word", respectively, as a transition aid.
537    
538    43. A minimizing repeat of a class containing codepoints greater than 255 in
539        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
540        compiled to use the heap for recursion.
541    
542    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
543        is configured.
544    
545    
546    Version 8.33 28-May-2013
547    ------------------------
548    
549    1.  Added 'U' to some constants that are compared to unsigned integers, to
550        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
551        variables that are added to signed variables, to ensure the result is
552        signed and can be negated.
553    
554    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
555        CMake config files.
556    
557    3.  Revise the creation of config.h.generic so that all boolean macros are
558        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
559        overriding via -D on the command line possible.
560    
561    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
562        to unsigned int is reported to make a quite noticeable speed difference in
563        a specific Windows environment. Testing on Linux did also appear to show
564        some benefit (and it is clearly not harmful). Also fixed the definition of
565        Xop which should be unsigned.
566    
567    5.  Related to (4), changing the definition of the intermediate variable cc
568        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
569        improvements.
570    
571    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
572        unnecessary spaces.
573    
574    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
575        and later.
576    
577    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
578    
579    9.  Optimizing fast_forward_start_bits in JIT.
580    
581    10. Adding support for callouts in JIT, and fixing some issues revealed
582        during this work. Namely:
583    
584        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
585    
586        (b) Minimum length was not checked before the matching is started.
587    
588    11. The value of capture_last that is passed to callouts was incorrect in some
589        cases when there was a capture on one path that was subsequently abandoned
590        after a backtrack. Also, the capture_last value is now reset after a
591        recursion, since all captures are also reset in this case.
592    
593    12. The interpreter no longer returns the "too many substrings" error in the
594        case when an overflowing capture is in a branch that is subsequently
595        abandoned after a backtrack.
596    
597    13. In the pathological case when an offset vector of size 2 is used, pcretest
598        now prints out the matched string after a yield of 0 or 1.
599    
600    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
601        Only supported by the JIT compiler at the moment.
602    
603    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
604    
605    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
606        offset of the starting point of the matching process, provided the offsets
607        vector is large enough.
608    
609    17. The \A escape now records a lookbehind value of 1, though its execution
610        does not actually inspect the previous character. This is to ensure that,
611        in partial multi-segment matching, at least one character from the old
612        segment is retained when a new segment is processed. Otherwise, if there
613        are no lookbehinds in the pattern, \A might match incorrectly at the start
614        of a new segment.
615    
616    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
617    
618    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
619        gives some modest performance improvement in 8-bit mode.
620    
621    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
622        be expressed in certain programming languages using Universal Character
623        Names.
624    
625    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
626        which points out that "non characters" are not "characters that may not
627        appear in Unicode strings" but rather "characters that are reserved for
628        internal use and have only local meaning".
629    
630    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
631        there was a conditional group that depended on an assertion, if the
632        assertion was false, the callout that immediately followed the alternation
633        in the condition was skipped when pcre_exec() was used for matching.
634    
635    23. Allow an explicit callout to be inserted before an assertion that is the
636        condition for a conditional group, for compatibility with automatic
637        callouts, which always insert a callout at this point.
638    
639    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
640        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
641    
642    25. (*PRUNE) is now supported by the JIT compiler.
643    
644    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
645    
646    27. Fix the case where there are two or more SKIPs with arguments that may be
647        ignored.
648    
649    28. (*SKIP) is now supported by the JIT compiler.
650    
651    29. (*THEN) is now supported by the JIT compiler.
652    
653    30. Update RunTest with additional test selector options.
654    
655    31. The way PCRE handles backtracking verbs has been changed in two ways.
656    
657        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
658        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
659        backtracking. In some cases this makes it more Perl-compatible, but Perl's
660        rather obscure rules do not always do the same thing.
661    
662        (2) Previously, backtracking verbs were confined within assertions. This is
663        no longer the case for positive assertions, except for (*ACCEPT). Again,
664        this sometimes improves Perl compatibility, and sometimes does not.
665    
666    32. A number of tests that were in test 2 because Perl did things differently
667        have been moved to test 1, because either Perl or PCRE has changed, and
668        these tests are now compatible.
669    
670    32. Backtracking control verbs are now handled in the same way in JIT and
671        interpreter.
672    
673    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
674        contained a forward subroutine reference caused a compile error.
675    
676    34. Auto-detect and optimize limited repetitions in JIT.
677    
678    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
679        blocking (*UTF) etc.
680    
681    36. In the interpreter, maximizing pattern repetitions for characters and
682        character types now use tail recursion, which reduces stack usage.
683    
684    37. The value of the max lookbehind was not correctly preserved if a compiled
685        and saved regex was reloaded on a host of different endianness.
686    
687    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
688        of the compiled pattern block, expand the flags field from 16 to 32 bits
689        because it was almost full.
690    
691    39. Try madvise first before posix_madvise.
692    
693    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
694        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
695        original change locked out all matches of empty strings. This has been
696        changed so that one match of an empty string per line is recognized.
697        Subsequent searches on the same line (for colouring or for --only-matching,
698        for example) do not recognize empty strings.
699    
700    41. Applied a user patch to fix a number of spelling mistakes in comments.
701    
702    42. Data lines longer than 65536 caused pcretest to crash.
703    
704    43. Clarified the data type for length and startoffset arguments for pcre_exec
705        and pcre_dfa_exec in the function-specific man pages, where they were
706        explicitly stated to be in bytes, never having been updated. I also added
707        some clarification to the pcreapi man page.
708    
709    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
710        a segmentation fault.
711    
712    
713    Version 8.32 30-November-2012
714    -----------------------------
715    
716    1.  Improved JIT compiler optimizations for first character search and single
717        character iterators.
718    
719    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
720        Patch by Daniel Richard G.
721    
722    3.  Single character iterator optimizations in the JIT compiler.
723    
724    4.  Improved JIT compiler optimizations for character ranges.
725    
726    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
727        Reported by Giuseppe D'Angelo.
728    
729    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
730        of a line, was being set incorrectly in cases where .* appeared inside
731        atomic brackets at the start of a pattern, or where there was a subsequent
732        *PRUNE or *SKIP.
733    
734    7.  Improved instruction cache flush for POWER/PowerPC.
735        Patch by Daniel Richard G.
736    
737    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
738        grep:
739    
740        (a) There is now no limit to the number of patterns to be matched.
741    
742        (b) An error is given if a pattern is too long.
743    
744        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
745            are now supported.
746    
747        (d) --exclude-from and --include-from (multiple use) have been added.
748    
749        (e) Exclusions and inclusions now apply to all files and directories, not
750            just to those obtained from scanning a directory recursively.
751    
752        (f) Multiple uses of -f and --file-list are now supported.
753    
754        (g) In a Windows environment, the default for -d has been changed from
755            "read" (the GNU grep default) to "skip", because otherwise the presence
756            of a directory in the file list provokes an error.
757    
758        (h) The documentation has been revised and clarified in places.
759    
760    9.  Improve the matching speed of capturing brackets.
761    
762    10. Changed the meaning of \X so that it now matches a Unicode extended
763        grapheme cluster.
764    
765    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
766        out POSIX threads when JIT support is configured.
767    
768    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
769    
770    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
771        structure could go wrong in environments where size_t is not the same size
772        as int.
773    
774    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
775        set.
776    
777    15. The EBCDIC support had decayed; later updates to the code had included
778        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
779        general tidy up of EBCDIC-related issues, and the documentation was also
780        not quite right. There is now a test that can be run on ASCII systems to
781        check some of the EBCDIC-related things (but is it not a full test).
782    
783    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
784        in a small tidy to the code.
785    
786    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
787    
788    18. If the --only-matching (-o) option in pcregrep is specified multiple
789        times, each one causes appropriate output. For example, -o1 -o2 outputs the
790        substrings matched by the 1st and 2nd capturing parentheses. A separating
791        string can be specified by --om-separator (default empty).
792    
793    19. Improving the first n character searches.
794    
795    20. Turn case lists for horizontal and vertical white space into macros so that
796        they are defined only once.
797    
798    21. This set of changes together give more compatible Unicode case-folding
799        behaviour for characters that have more than one other case when UCP
800        support is available.
801    
802        (a) The Unicode property table now has offsets into a new table of sets of
803            three or more characters that are case-equivalent. The MultiStage2.py
804            script that generates these tables (the pcre_ucd.c file) now scans
805            CaseFolding.txt instead of UnicodeData.txt for character case
806            information.
807    
808        (b) The code for adding characters or ranges of characters to a character
809            class has been abstracted into a generalized function that also handles
810            case-independence. In UTF-mode with UCP support, this uses the new data
811            to handle characters with more than one other case.
812    
813        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
814            whose other case is greater than 256 are now correctly matched
815            caselessly. Previously, the high codepoint matched the low one, but not
816            vice versa.
817    
818        (d) The processing of \h, \H, \v, and \ in character classes now makes use
819            of the new class addition function, using character lists defined as
820            macros alongside the case definitions of 20 above.
821    
822        (e) Caseless back references now work with characters that have more than
823            one other case.
824    
825        (f) General caseless matching of characters with more than one other case
826            is supported.
827    
828    22. Unicode character properties were updated from Unicode 6.2.0
829    
830    23. Improved CMake support under Windows. Patch by Daniel Richard G.
831    
832    24. Add support for 32-bit character strings, and UTF-32
833    
834    25. Major JIT compiler update (code refactoring and bugfixing).
835        Experimental Sparc 32 support is added.
836    
837    26. Applied a modified version of Daniel Richard G's patch to create
838        pcre.h.generic and config.h.generic by "make" instead of in the
839        PrepareRelease script.
840    
841    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
842        pcre_compile.c when checking for a zero character.
843    
844    28. Introducing a native interface for JIT. Through this interface, the compiled
845        machine code can be directly executed. The purpose of this interface is to
846        provide fast pattern matching, so several sanity checks are not performed.
847        However, feature tests are still performed. The new interface provides
848        1.4x speedup compared to the old one.
849    
850    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
851        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
852        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
853    
854    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
855        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
856        the "old" RFC 2279). Instead, it ended up passing a negative length to
857        pcre_exec().
858    
859    31. Add support for GCC's visibility feature to hide internal functions.
860    
861    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
862        "unknown -C option" after outputting 0 or 1.
863    
864    33. There is now support for generating a code coverage report for the test
865        suite in environments where gcc is the compiler and lcov is installed. This
866        is mainly for the benefit of the developers.
867    
868    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
869        unaddressable using valgrind annotations, allowing valgrind to detect
870        invalid memory accesses. This is mainly for the benefit of the developers.
871    
872    25. (*UTF) can now be used to start a pattern in any of the three libraries.
873    
874    26. Give configure error if --enable-cpp but no C++ compiler found.
875    
876    
877    Version 8.31 06-July-2012
878    -------------------------
879    
880    1.  Fixing a wrong JIT test case and some compiler warnings.
881    
882    2.  Removed a bashism from the RunTest script.
883    
884    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
885        to unsigned type, result still unsigned" that was given by an MS compiler
886        on encountering the code "-sizeof(xxx)".
887    
888    4.  Partial matching support is added to the JIT compiler.
889    
890    5.  Fixed several bugs concerned with partial matching of items that consist
891        of more than one character:
892    
893        (a) /^(..)\1/ did not partially match "aba" because checking references was
894            done on an "all or nothing" basis. This also applied to repeated
895            references.
896    
897        (b) \R did not give a hard partial match if \r was found at the end of the
898            subject.
899    
900        (c) \X did not give a hard partial match after matching one or more
901            characters at the end of the subject.
902    
903        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
904            a partial match for the string "\r".
905    
906        (e) When newline was set to CRLF, the metacharacter "." did not recognize
907            a partial match for a CR character at the end of the subject string.
908    
909    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
910        running pcretest, the text "(JIT)" added to the output whenever JIT is
911        actually used to run the match.
912    
913    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
914        or /S+[+] with a digit between 1 and 7.
915    
916    8.  OP_NOT now supports any UTF character not just single-byte ones.
917    
918    9.  (*MARK) control verb is now supported by the JIT compiler.
919    
920    10. The command "./RunTest list" lists the available tests without actually
921        running any of them. (Because I keep forgetting what they all are.)
922    
923    11. Add PCRE_INFO_MAXLOOKBEHIND.
924    
925    12. Applied a (slightly modified) user-supplied patch that improves performance
926        when the heap is used for recursion (compiled with --disable-stack-for-
927        recursion). Instead of malloc and free for each heap frame each time a
928        logical recursion happens, frames are retained on a chain and re-used where
929        possible. This sometimes gives as much as 30% improvement.
930    
931    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
932        call.
933    
934    14. As documented, (*COMMIT) is now confined to within a positive assertion.
935    
936    15. It is now possible to link pcretest with libedit as an alternative to
937        libreadline.
938    
939    16. (*COMMIT) control verb is now supported by the JIT compiler.
940    
941    17. The Unicode data tables have been updated to Unicode 6.1.0.
942    
943    18. Added --file-list option to pcregrep.
944    
945    19. Added binary file support to pcregrep, including the -a, --binary-files,
946        -I, and --text options.
947    
948    20. The madvise function is renamed for posix_madvise for QNX compatibility
949        reasons. Fixed by Giuseppe D'Angelo.
950    
951    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
952    
953    22. Changed the option for creating symbolic links for 16-bit man pages from
954        -s to -sf so that re-installing does not cause issues.
955    
956    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
957    
958    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
959        matches in certain environments (the workspace was not being correctly
960        retained). Also added to pcre_dfa_exec() a simple plausibility check on
961        some of the workspace data at the beginning of a restart.
962    
963    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
964        was not doing so when it should - probably a typo introduced by SVN 528
965        (change 8.10/14).
966    
967    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
968        \w+ when the character tables indicated that \x{c4} was a word character.
969        There were several related cases, all because the tests for doing a table
970        lookup were testing for characters less than 127 instead of 255.
971    
972    27. If a pattern contains capturing parentheses that are not used in a match,
973        their slots in the ovector are set to -1. For those that are higher than
974        any matched groups, this happens at the end of processing. In the case when
975        there were back references that the ovector was too small to contain
976        (causing temporary malloc'd memory to be used during matching), and the
977        highest capturing number was not used, memory off the end of the ovector
978        was incorrectly being set to -1. (It was using the size of the temporary
979        memory instead of the true size.)
980    
981    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
982        ovector size, it uses memory at the end of the block that it has got.
983    
984    29. Check for an overlong MARK name and give an error at compile time. The
985        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
986    
987    30. JIT compiler update.
988    
989    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
990        Rill for the patch.
991    
992    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
993    
994    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
995    
996    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
997        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
998        was enabled.
999    
1000    35. Improve JIT code generation for greedy plus quantifier.
1001    
1002    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1003        1 to "aa" instead of to an empty string. The bug affected repeated groups
1004        that could potentially match an empty string.
1005    
1006    37. Optimizing single character iterators in JIT.
1007    
1008    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1009        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1010        codepoints that are too big for the mode are faulted, and in a UTF mode,
1011        disallowed codepoints are also faulted.
1012    
1013    39. If PCRE was compiled with UTF support, in three places in the DFA
1014        matcher there was code that should only have been obeyed in UTF mode, but
1015        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1016        processing when bytes with values greater than 127 were present. In 16-bit
1017        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1018        both cases the values are those that cannot be the first data item in a UTF
1019        character. The three items that might have provoked this were recursions,
1020        possessively repeated groups, and atomic groups.
1021    
1022    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1023        and pcregrep, because some OS require shared objects to be explicitly
1024        passed to ld, causing the link step to fail if they are not.
1025    
1026    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1027        mode, patterns that started with \h* or \R* might be incorrectly matched.
1028    
1029    
1030    Version 8.30 04-February-2012
1031    -----------------------------
1032    
1033    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1034        name is defined in ctype.h.
1035    
1036    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1037        only in quite long subpatterns.
1038    
1039    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1040        since it was replaced by pcre_fullinfo() in February 2000.
1041    
1042    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1043        match a (*MARK), and the match failed at the start of the subject, a
1044        reference to memory before the start of the subject could occur. This bug
1045        was introduced by fix 17 of release 8.21.
1046    
1047    5.  A reference to an unset group with zero minimum repetition was giving
1048        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1049        /(another)?(\1?)test/ matched against "hello world test". This bug was
1050        introduced in release 8.13.
1051    
1052    6.  Add support for 16-bit character strings (a large amount of work involving
1053        many changes and refactorings).
1054    
1055    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1056        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1057        from a file.
1058    
1059    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1060        rounding is not applied in this particular case).
1061    
1062    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1063        if they appear, or are escaped, in patterns.
1064    
1065    10. Get rid of a number of -Wunused-but-set-variable warnings.
1066    
1067    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1068        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1069        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1070        also returns the mark "x". This bug applied to capturing parentheses,
1071        non-capturing parentheses, and atomic parentheses. It also applied to some
1072        assertions.
1073    
1074    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1075        information out of configure.ac instead of relying on pcre.h.generic, which
1076        is not stored in the repository.
1077    
1078    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1079        -lreadline.
1080    
1081    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1082    
1083    15. Applied Graycode's patch to put the top-level frame on the stack rather
1084        than the heap when not using the stack for recursion. This gives a
1085        performance improvement in many cases when recursion is not deep.
1086    
1087    16. Experimental code added to "pcretest -C" to output the stack frame size.
1088    
1089    
1090    Version 8.21 12-Dec-2011
1091    ------------------------
1092    
1093    1.  Updating the JIT compiler.
1094    
1095    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1096        are added as well.
1097    
1098    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1099        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1100        calling _pcre_jit_exec. Some extra comments are added.
1101    
1102    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1103        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1104        was introduced by change 18 for 8.20.
1105    
1106    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1107        ECMA-262 standard.
1108    
1109    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1110        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1111        This bug was probably introduced by change 9 of 8.13.
1112    
1113    7.  While fixing 6 above, I noticed that a number of other items were being
1114        incorrectly rejected as "not fixed length". This arose partly because newer
1115        opcodes had not been added to the fixed-length checking code. I have (a)
1116        corrected the bug and added tests for these items, and (b) arranged for an
1117        error to occur if an unknown opcode is encountered while checking for fixed
1118        length instead of just assuming "not fixed length". The items that were
1119        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1120        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1121        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1122    
1123    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1124        being incorrectly compiled and would have given unpredicatble results.
1125    
1126    9.  A possessively repeated subpattern with minimum repeat count greater than
1127        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1128        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1129        the first (A) could occur when it should not.
1130    
1131    10. Add a cast and remove a redundant test from the code.
1132    
1133    11. JIT should use pcre_malloc/pcre_free for allocation.
1134    
1135    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1136        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1137        is anything other than /usr, -L is still shown).
1138    
1139    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1140    
1141    14. Perl does not support \N without a following name in a [] class; PCRE now
1142        also gives an error.
1143    
1144    15. If a forward reference was repeated with an upper limit of around 2000,
1145        it caused the error "internal error: overran compiling workspace". The
1146        maximum number of forward references (including repeats) was limited by the
1147        internal workspace, and dependent on the LINK_SIZE. The code has been
1148        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1149        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1150        of around 200,000 forward references. While doing this, I also speeded up
1151        the filling in of repeated forward references.
1152    
1153    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1154        incorrectly expecting the subject to contain another "a" after the start.
1155    
1156    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1157        in the match, the SKIP should be ignored. This was not happening; instead
1158        the SKIP was being treated as NOMATCH. For patterns such as
1159        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1160        tested.
1161    
1162    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1163        now much more compatible with Perl, in particular in cases where the result
1164        is a non-match for a non-anchored pattern. For example, if
1165        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1166        "m", where previously it did not return a name. A side effect of this
1167        change is that for partial matches, the last encountered mark name is
1168        returned, as for non matches. A number of tests that were previously not
1169        Perl-compatible have been moved into the Perl-compatible test files. The
1170        refactoring has had the pleasing side effect of removing one argument from
1171        the match() function, thus reducing its stack requirements.
1172    
1173    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1174        subsequent uses of /S (without +) incorrectly behaved like /S+.
1175    
1176    21. Retrieve executable code size support for the JIT compiler and fixing
1177        some warnings.
1178    
1179    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1180        not work when the shorter character appeared right at the end of the
1181        subject string.
1182    
1183    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1184        systems.
1185    
1186    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1187        output it when the /M option is used in pcretest.
1188    
1189    25. The CheckMan script was not being included in the distribution. Also, added
1190        an explicit "perl" to run Perl scripts from the PrepareRelease script
1191        because this is reportedly needed in Windows.
1192    
1193    26. If study data was being save in a file and studying had not found a set of
1194        "starts with" bytes for the pattern, the data written to the file (though
1195        never used) was taken from uninitialized memory and so caused valgrind to
1196        complain.
1197    
1198    27. Updated RunTest.bat as provided by Sheri Pierce.
1199    
1200    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1201    
1202    29. Computation of memory usage for the table of capturing group names was
1203        giving an unnecessarily large value.
1204    
1205    
1206    Version 8.20 21-Oct-2011
1207    ------------------------
1208    
1209    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1210        a POSIX class. After further experiments with Perl, which convinced me that
1211        Perl has bugs and confusions, a closing square bracket is no longer allowed
1212        in a POSIX name. This bug also affected patterns with classes that started
1213        with full stops.
1214    
1215    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1216        captured substring, but while checking the failing first alternative,
1217        substring 1 is temporarily captured. If the output vector supplied to
1218        pcre_exec() was not big enough for this capture, the yield of the function
1219        was still zero ("insufficient space for captured substrings"). This cannot
1220        be totally fixed without adding another stack variable, which seems a lot
1221        of expense for a edge case. However, I have improved the situation in cases
1222        such as /(a)(b)x|abc/ matched against "abc", where the return code
1223        indicates that fewer than the maximum number of slots in the ovector have
1224        been set.
1225    
1226    3.  Related to (2) above: when there are more back references in a pattern than
1227        slots in the output vector, pcre_exec() uses temporary memory during
1228        matching, and copies in the captures as far as possible afterwards. It was
1229        using the entire output vector, but this conflicts with the specification
1230        that only 2/3 is used for passing back captured substrings. Now it uses
1231        only the first 2/3, for compatibility. This is, of course, another edge
1232        case.
1233    
1234    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1235        main code base, and can be used by building with --enable-jit. When this is
1236        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1237        runtime --no-jit option is given.
1238    
1239    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1240        ovector, the return from the function was zero, implying that there were
1241        other matches that did not fit. The correct "exactly full" value is now
1242        returned.
1243    
1244    6.  If a subpattern that was called recursively or as a subroutine contained
1245        (*PRUNE) or any other control that caused it to give a non-standard return,
1246        invalid errors such as "Error -26 (nested recursion at the same subject
1247        position)" or even infinite loops could occur.
1248    
1249    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1250        computing the minimum length on reaching *ACCEPT, and so ended up with the
1251        wrong value of 1 rather than 0. Further investigation indicates that
1252        computing a minimum subject length in the presence of *ACCEPT is difficult
1253        (think back references, subroutine calls), and so I have changed the code
1254        so that no minimum is registered for a pattern that contains *ACCEPT.
1255    
1256    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1257        it was not handled as intended. [But see 16 below.]
1258    
1259    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1260        Sheri Pierce.
1261    
1262    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1263        the first byte in a match must be "a".
1264    
1265    11. Change 17 for 8.13 increased the recursion depth for patterns like
1266        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1267        pattern contains any instances of (*THEN). If it does not, the old
1268        optimizations are restored. It would be nice to do this on a per-group
1269        basis, but at the moment that is not feasible.
1270    
1271    12. In some environments, the output of pcretest -C is CRLF terminated. This
1272        broke RunTest's code that checks for the link size. A single white space
1273        character after the value is now allowed for.
1274    
1275    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1276        For "fr", it uses the Windows-specific input and output files.
1277    
1278    14. If (*THEN) appeared in a group that was called recursively or as a
1279        subroutine, it did not work as intended. [But see next item.]
1280    
1281    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1282        pattern fragments (but not containing any | characters). If A and B are
1283        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1284        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1285        D. In other words, Perl considers parentheses that do not contain any |
1286        characters to be part of a surrounding alternative, whereas PCRE was
1287        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1288        differently. PCRE now behaves in the same way as Perl, except in the case
1289        of subroutine/recursion calls such as (?1) which have in any case always
1290        been different (but PCRE had them first :-).
1291    
1292    16. Related to 15 above: Perl does not treat the | in a conditional group as
1293        creating alternatives. Such a group is treated in the same way as an
1294        ordinary group without any | characters when processing (*THEN). PCRE has
1295        been changed to match Perl's behaviour.
1296    
1297    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1298        RunGrepTest script failed.
1299    
1300    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1301        inevitable for groups that contain captures, but it can lead to a lot of
1302        stack use in large patterns. The old behaviour has been restored for atomic
1303        groups that do not contain any capturing parentheses.
1304    
1305    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1306        suppress the check for a minimum subject length at run time. (If it was
1307        given to pcre_exec() or pcre_dfa_exec() it did work.)
1308    
1309    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1310        fail to work when decoding hex characters in data strings in EBCDIC
1311        environments.
1312    
1313    21. It appears that in at least one Mac OS environment, the isxdigit() function
1314        is implemented as a macro that evaluates to its argument more than once,
1315        contravening the C 90 Standard (I haven't checked a later standard). There
1316        was an instance in pcretest which caused it to go wrong when processing
1317        \x{...} escapes in subject strings. The has been rewritten to avoid using
1318        things like p++ in the argument of isxdigit().
1319    
1320    
1321    Version 8.13 16-Aug-2011
1322    ------------------------
1323    
1324    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1325    
1326    2.  Two minor typos in pcre_internal.h have been fixed.
1327    
1328    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1329        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1330        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1331    
1332    4.  There were a number of related bugs in the code for matching backrefences
1333        caselessly in UTF-8 mode when codes for the characters concerned were
1334        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1335        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1336        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1337        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1338        2-byte code at the end of the subject (it thought there wasn't enough data
1339        left).
1340    
1341    5.  Comprehensive information about what went wrong is now returned by
1342        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1343        as the output vector has at least 2 elements. The offset of the start of
1344        the failing character and a reason code are placed in the vector.
1345    
1346    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1347        now returned is for the first byte of the failing character, instead of the
1348        last byte inspected. This is an incompatible change, but I hope it is small
1349        enough not to be a problem. It makes the returned offset consistent with
1350        pcre_exec() and pcre_dfa_exec().
1351    
1352    7.  pcretest now gives a text phrase as well as the error number when
1353        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1354        failure, the offset and reason code are output.
1355    
1356    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1357        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1358        back over a single character (\n). This seems wrong (because it treated the
1359        two characters as a single entity when going forwards), conflicts with the
1360        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1361        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1362        has been changed.
1363    
1364    9.  Some internal refactoring has changed the processing so that the handling
1365        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1366        time (the PCRE_DOTALL option was changed this way some time ago: version
1367        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1368        which was always a bit of a fudge. It also means that there is one less
1369        argument for the match() function, which reduces its stack requirements
1370        slightly. This change also fixes an incompatibility with Perl: the pattern
1371        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1372    
1373    10. More internal refactoring has drastically reduced the number of recursive
1374        calls to match() for possessively repeated groups such as (abc)++ when
1375        using pcre_exec().
1376    
1377    11. While implementing 10, a number of bugs in the handling of groups were
1378        discovered and fixed:
1379    
1380        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1381        (a|)*(?1) gave a compile-time internal error.
1382        ((a|)+)+  did not notice that the outer group could match an empty string.
1383        (^a|^)+   was not marked as anchored.
1384        (.*a|.*)+ was not marked as matching at start or after a newline.
1385    
1386    12. Yet more internal refactoring has removed another argument from the match()
1387        function. Special calls to this function are now indicated by setting a
1388        value in a variable in the "match data" data block.
1389    
1390    13. Be more explicit in pcre_study() instead of relying on "default" for
1391        opcodes that mean there is no starting character; this means that when new
1392        ones are added and accidentally left out of pcre_study(), testing should
1393        pick them up.
1394    
1395    14. The -s option of pcretest has been documented for ages as being an old
1396        synonym of -m (show memory usage). I have changed it to mean "force study
1397        for every regex", that is, assume /S for every regex. This is similar to -i
1398        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1399        using it. It makes it easier to run collections of tests with and without
1400        study enabled, and thereby test pcre_study() more easily. All the standard
1401        tests are now run with and without -s (but some patterns can be marked as
1402        "never study" - see 20 below).
1403    
1404    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1405        restoration of the capturing data to the outer values was not happening
1406        correctly.
1407    
1408    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1409        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1410        pattern had matched an empty string, and so incorrectly returned a no
1411        match.
1412    
1413    17. There was optimizing code for the last branch of non-capturing parentheses,
1414        and also for the obeyed branch of a conditional subexpression, which used
1415        tail recursion to cut down on stack usage. Unfortunately, now that there is
1416        the possibility of (*THEN) occurring in these branches, tail recursion is
1417        no longer possible because the return has to be checked for (*THEN). These
1418        two optimizations have therefore been removed. [But see 8.20/11 above.]
1419    
1420    18. If a pattern containing \R was studied, it was assumed that \R always
1421        matched two bytes, thus causing the minimum subject length to be
1422        incorrectly computed because \R can also match just one byte.
1423    
1424    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1425        was incorrectly computed.
1426    
1427    20. If /S is present twice on a test pattern in pcretest input, it now
1428        *disables* studying, thereby overriding the use of -s on the command line
1429        (see 14 above). This is necessary for one or two tests to keep the output
1430        identical in both cases.
1431    
1432    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1433        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1434    
1435    22. When an atomic group that contained a capturing parenthesis was
1436        successfully matched, but the branch in which it appeared failed, the
1437        capturing was not being forgotten if a higher numbered group was later
1438        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1439        group 1 to "a", when in fact it should be unset. This applied to multi-
1440        branched capturing and non-capturing groups, repeated or not, and also to
1441        positive assertions (capturing in negative assertions does not happen
1442        in PCRE) and also to nested atomic groups.
1443    
1444    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1445        subject after a captured substring, to make it easier to tell which of a
1446        number of identical substrings has been captured.
1447    
1448    24. The way atomic groups are processed by pcre_exec() has been changed so that
1449        if they are repeated, backtracking one repetition now resets captured
1450        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1451        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1452        "aaa". Previously, it would have been "a". As part of this code
1453        refactoring, the way recursive calls are handled has also been changed.
1454    
1455    25. If an assertion condition captured any substrings, they were not passed
1456        back unless some other capturing happened later. For example, if
1457        (?(?=(a))a) was matched against "a", no capturing was returned.
1458    
1459    26. When studying a pattern that contained subroutine calls or assertions,
1460        the code for finding the minimum length of a possible match was handling
1461        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1462        group 1 called group 2 while simultaneously a separate group 2 called group
1463        1). A stack overflow occurred in this case. I have fixed this by limiting
1464        the recursion depth to 10.
1465    
1466    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1467        Fortmann. This supports explicit test numbers on the command line, and has
1468        argument validation and error reporting.
1469    
1470    28. An instance of \X with an unlimited repeat could fail if at any point the
1471        first character it looked at was a mark character.
1472    
1473    29. Some minor code refactoring concerning Unicode properties and scripts
1474        should reduce the stack requirement of match() slightly.
1475    
1476    30. Added the '=' option to pcretest to check the setting of unused capturing
1477        slots at the end of the pattern, which are documented as being -1, but are
1478        not included in the return count.
1479    
1480    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1481        compiled something random. Now it gives a compile-time error (as does
1482        Perl).
1483    
1484    32. A *MARK encountered during the processing of a positive assertion is now
1485        recorded and passed back (compatible with Perl).
1486    
1487    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1488        had alternative anchored branches, the search for a second match in a line
1489        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1490        matched the line "0102" twice. The same bug affected patterns that started
1491        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1492        twice.
1493    
1494    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1495        does, and because of capturing effects, quantifying parenthesized
1496        assertions may at times be useful. Quantifiers are now allowed for
1497        parenthesized assertions.
1498    
1499    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1500    
1501    36. \g was being checked for fancy things in a character class, when it should
1502        just be a literal "g".
1503    
1504    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1505        appearance of a nested POSIX class supersedes an apparent external class.
1506        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1507        unescaped square brackets may also appear as part of class names. For
1508        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1509        more like Perl. (But see 8.20/1 above.)
1510    
1511    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1512        was because it thought it was \N{name}, which is not supported).
1513    
1514    39. Add minix to OS list not supporting the -S option in pcretest.
1515    
1516    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1517        cannot analyze patterns in sufficient detail to catch mutual recursions
1518        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1519        subgroup is called recursively as a subpattern for a second time at the
1520        same position in the subject string. In previous releases this might have
1521        been caught by the recursion limit, or it might have run out of stack.
1522    
1523    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1524        happen only once. PCRE was, however incorrectly giving a compile time error
1525        "recursive call could loop indefinitely" because it cannot analyze the
1526        pattern in sufficient detail. The compile time test no longer happens when
1527        PCRE is compiling a conditional subpattern, but actual runaway loops are
1528        now caught at runtime (see 40 above).
1529    
1530    42. It seems that Perl allows any characters other than a closing parenthesis
1531        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1532        has been changed to be the same.
1533    
1534    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1535        as not to get warnings when autogen.sh is called. Also changed
1536        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1537    
1538    44. To help people who use pcregrep to scan files containing exceedingly long
1539        lines, the following changes have been made:
1540    
1541        (a) The default value of the buffer size parameter has been increased from
1542            8K to 20K. (The actual buffer used is three times this size.)
1543    
1544        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1545            PCRE is built.
1546    
1547        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1548            to be set at run time.
1549    
1550        (d) Numerical values in pcregrep options can be followed by K or M, for
1551            example --buffer-size=50K.
1552    
1553        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1554            given and the return code is set to 2.
1555    
1556    45. Add a pointer to the latest mark to the callout data block.
1557    
1558    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1559        partial match of an empty string instead of no match. This was specific to
1560        the use of ".".
1561    
1562    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1563        complete match instead of a partial match. This bug was dependent on both
1564        the PCRE_UTF8 and PCRE_DOTALL options being set.
1565    
1566    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1567        starting byte set, because \b was not being ignored.
1568    
1569    
1570    Version 8.12 15-Jan-2011
1571    ------------------------
1572    
1573    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1574        checks for such things as part of the documentation building process.
1575    
1576    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1577        --match-limit and --recursion-limit options (added for 8.11). In
1578        particular, this made one of the standard tests fail. (The integer value
1579        went into the wrong half of a long int.)
1580    
1581    3.  If the --colour option was given to pcregrep with -v (invert match), it
1582        did strange things, either producing crazy output, or crashing. It should,
1583        of course, ignore a request for colour when reporting lines that do not
1584        match.
1585    
1586    4.  Another pcregrep bug caused similar problems if --colour was specified with
1587        -M (multiline) and the pattern match finished with a line ending.
1588    
1589    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1590        matched in multiline mode, the following line was shown as part of the
1591        match. This seems wrong, so I have changed it.
1592    
1593    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1594        the check for further matches in the same line (so they could be coloured)
1595        to overrun the end of the current line. If another match was found, it was
1596        incorrectly shown (and then shown again when found in the next line).
1597    
1598    7.  If pcregrep was compiled under Windows, there was a reference to the
1599        function pcregrep_exit() before it was defined. I am assuming this was
1600        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1601        reported by a user. I've moved the definition above the reference.
1602    
1603    
1604    Version 8.11 10-Dec-2010
1605    ------------------------
1606    
1607    1.  (*THEN) was not working properly if there were untried alternatives prior
1608        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1609        backtracked to try for "b" instead of moving to the next alternative branch
1610        at the same level (in this case, to look for "c"). The Perl documentation
1611        is clear that when (*THEN) is backtracked onto, it goes to the "next
1612        alternative in the innermost enclosing group".
1613    
1614    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1615        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1616        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1617        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1618        (*THEN).
1619    
1620    3.  If \s appeared in a character class, it removed the VT character from
1621        the class, even if it had been included by some previous item, for example
1622        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1623        of \s, but is part of the POSIX "space" class.)
1624    
1625    4.  A partial match never returns an empty string (because you can always
1626        match an empty string at the end of the subject); however the checking for
1627        an empty string was starting at the "start of match" point. This has been
1628        changed to the "earliest inspected character" point, because the returned
1629        data for a partial match starts at this character. This means that, for
1630        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1631        (previously it gave "no match").
1632    
1633    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1634        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1635        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1636        has an implication that the given string is incomplete (because a partial
1637        match is preferred over a full match). For this reason, these items now
1638        give a partial match in this situation. [Aside: previously, the one case
1639        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1640        match rather than a full match, which was wrong by the old rules, but is
1641        now correct.]
1642    
1643    6.  There was a bug in the handling of #-introduced comments, recognized when
1644        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1645        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1646        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1647        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1648        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1649        places in pcre_compile().
1650    
1651    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1652        comments when looking ahead for named forward references to subpatterns,
1653        the only newline sequence it recognized was NL. It now handles newlines
1654        according to the set newline convention.
1655    
1656    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1657        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1658        cater for a lack of strerror(). These oversights have been fixed.
1659    
1660    9.  Added --match-limit and --recursion-limit to pcregrep.
1661    
1662    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1663    
1664    11. When the -o option was used, pcregrep was setting a return code of 1, even
1665        when matches were found, and --line-buffered was not being honoured.
1666    
1667    12. Added an optional parentheses number to the -o and --only-matching options
1668        of pcregrep.
1669    
1670    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1671        can match an empty string. The code to do it in pcretest and pcredemo
1672        needed fixing:
1673    
1674        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1675            only one byte after an empty string match just before CRLF (this case
1676            just got forgotten; "any" and "anycrlf" were OK).
1677    
1678        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1679            mode when an empty string match preceded an ASCII character followed by
1680            a non-ASCII character. (The code for advancing by one character rather
1681            than one byte was nonsense.)
1682    
1683        (c) The pcredemo.c sample program did not have any code at all to handle
1684            the cases when CRLF is a valid newline sequence.
1685    
1686    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1687        as a starting offset was within the subject string. There is now a new
1688        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1689        negative or greater than the length of the string. In order to test this,
1690        pcretest is extended to allow the setting of negative starting offsets.
1691    
1692    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1693        starting offset points to the beginning of a UTF-8 character was
1694        unnecessarily clumsy. I tidied it up.
1695    
1696    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1697        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1698    
1699    17. Nobody had reported that the --include_dir option, which was added in
1700        release 7.7 should have been called --include-dir (hyphen, not underscore)
1701        for compatibility with GNU grep. I have changed it to --include-dir, but
1702        left --include_dir as an undocumented synonym, and the same for
1703        --exclude-dir, though that is not available in GNU grep, at least as of
1704        release 2.5.4.
1705    
1706    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1707        characters from a string of bytes) have been redefined so as not to use
1708        loops, in order to improve performance in some environments. At the same
1709        time, I abstracted some of the common code into auxiliary macros to save
1710        repetition (this should not affect the compiled code).
1711    
1712    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1713        compile-time error is now given if \c is not followed by an ASCII
1714        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1715        different, and any byte value is allowed.)
1716    
1717    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1718        START_OPTIMIZE option, which is now allowed at compile time - but just
1719        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1720        to pcregrep and other applications that have no direct access to PCRE
1721        options. The new /Y option in pcretest sets this option when calling
1722        pcre_compile().
1723    
1724    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1725        back references. Groups containing recursive back references were forced to
1726        be atomic by that change, but in the case of named groups, the amount of
1727        memory required was incorrectly computed, leading to "Failed: internal
1728        error: code overflow". This has been fixed.
1729    
1730    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1731        pcretest.c, to avoid build problems in some Borland environments.
1732    
1733    
1734    Version 8.10 25-Jun-2010
1735    ------------------------
1736    
1737    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1738        THEN.
1739    
1740    2.  (*ACCEPT) was not working when inside an atomic group.
1741    
1742    3.  Inside a character class, \B is treated as a literal by default, but
1743        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1744        causes the error). The code is unchanged, but I tidied the documentation.
1745    
1746    4.  Inside a character class, PCRE always treated \R and \X as literals,
1747        whereas Perl faults them if its -w option is set. I have changed PCRE so
1748        that it faults them when PCRE_EXTRA is set.
1749    
1750    5.  Added support for \N, which always matches any character other than
1751        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1752    
1753    6.  When compiling pcregrep with newer versions of gcc which may have
1754        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1755        declared with attribute warn_unused_result" were given. Just casting the
1756        result to (void) does not stop the warnings; a more elaborate fudge is
1757        needed. I've used a macro to implement this.
1758    
1759    7.  Minor change to pcretest.c to avoid a compiler warning.
1760    
1761    8.  Added four artifical Unicode properties to help with an option to make
1762        \s etc use properties (see next item). The new properties are: Xan
1763        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1764    
1765    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1766        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1767        this option. Modified pcretest to add /W to test this facility. Added
1768        REG_UCP to make it available via the POSIX interface.
1769    
1770    10. Added --line-buffered to pcregrep.
1771    
1772    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1773        studied, and the match started with a letter with a code point greater than
1774        127 whose first byte was different to the first byte of the other case of
1775        the letter, the other case of this starting letter was not recognized
1776        (#976).
1777    
1778    12. If a pattern that was studied started with a repeated Unicode property
1779        test, for example, \p{Nd}+, there was the theoretical possibility of
1780        setting up an incorrect bitmap of starting bytes, but fortunately it could
1781        not have actually happened in practice until change 8 above was made (it
1782        added property types that matched character-matching opcodes).
1783    
1784    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1785        possible starting bytes for non-anchored patterns.
1786    
1787    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1788        \R, and also a number of cases that involve Unicode properties, both
1789        explicit and implicit when PCRE_UCP is set.
1790    
1791    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1792        input, it could crash or give wrong results if characters with values
1793        greater than 0xc0 were present in the subject string. (Detail: it assumed
1794        UTF-8 input when processing these items.)
1795    
1796    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1797        size_t is 64-bit (#991).
1798    
1799    17. Added a check for running out of memory when PCRE is compiled with
1800        --disable-stack-for-recursion (#990).
1801    
1802    18. If the last data line in a file for pcretest does not have a newline on
1803        the end, a newline was missing in the output.
1804    
1805    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1806        less than 128) in its various bitmaps. However, there is a facility for
1807        generating tables according to the current locale when PCRE is compiled. It
1808        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1809        characters, are recognized by isspace() and therefore were getting set in
1810        these tables, and indeed these tables seem to approximate to ISO 8859. This
1811        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1812        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1813        which of course cannot start UTF-8 characters. I have changed the code so
1814        that only real ASCII characters (less than 128) and the correct starting
1815        bytes for UTF-8 encodings are set for characters greater than 127 when in
1816        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1817        altogether.)
1818    
1819    20. Added the /T option to pcretest so as to be able to run tests with non-
1820        standard character tables, thus making it possible to include the tests
1821        used for 19 above in the standard set of tests.
1822    
1823    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1824        reference to a subpattern the other side of a comment that contains an
1825        opening parenthesis caused either an internal compiling error, or a
1826        reference to the wrong subpattern.
1827    
1828    
1829    Version 8.02 19-Mar-2010
1830    ------------------------
1831    
1832    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1833    
1834    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1835        configured.
1836    
1837    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1838        original author of that file, following a query about its status.
1839    
1840    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1841        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1842    
1843    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1844        quantifier applied to a forward-referencing subroutine call, could compile
1845        incorrect code or give the error "internal error: previously-checked
1846        referenced subpattern not found".
1847    
1848    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1849        variables to point to external functions. For these systems, therefore,
1850        pcre_malloc etc. are now initialized to local functions that call the
1851        relevant global functions.
1852    
1853    7.  There were two entries missing in the vectors called coptable and poptable
1854        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1855        I've fixed the data, and added a kludgy way of testing at compile time that
1856        the lengths are correct (equal to the number of opcodes).
1857    
1858    8.  Following on from 7, I added a similar kludge to check the length of the
1859        eint vector in pcreposix.c.
1860    
1861    9.  Error texts for pcre_compile() are held as one long string to avoid too
1862        much relocation at load time. To find a text, the string is searched,
1863        counting zeros. There was no check for running off the end of the string,
1864        which could happen if a new error number was added without updating the
1865        string.
1866    
1867    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1868    
1869    11. \K was not working if it appeared in an atomic group or in a group that
1870        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1871        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1872        the assertion is positive, but not if it is negative.
1873    
1874    12. Change 11 fortuitously reduced the size of the stack frame used in the
1875        "match()" function of pcre_exec.c by one pointer. Forthcoming
1876        implementation of support for (*MARK) will need an extra pointer on the
1877        stack; I have reserved it now, so that the stack frame size does not
1878        decrease.
1879    
1880    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1881        item in branch that calls a recursion is a subroutine call - as in the
1882        second branch in the above example - was incorrectly given the compile-
1883        time error "recursive call could loop indefinitely" because pcre_compile()
1884        was not correctly checking the subroutine for matching a non-empty string.
1885    
1886    14. The checks for overrunning compiling workspace could trigger after an
1887        overrun had occurred. This is a "should never occur" error, but it can be
1888        triggered by pathological patterns such as hundreds of nested parentheses.
1889        The checks now trigger 100 bytes before the end of the workspace.
1890    
1891    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1892    
1893    
1894    Version 8.01 19-Jan-2010
1895    ------------------------
1896    
1897    1.  If a pattern contained a conditional subpattern with only one branch (in
1898        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1899        computed the wrong minimum data length (which is of course zero for such
1900        subpatterns). This could cause incorrect "no match" results.
1901    
1902    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1903        the pattern is reset in the first branch, pcre_compile() failed with
1904        "internal error: code overflow at offset...". This happened only when
1905        the reset was to the original external option setting. (An optimization
1906        abstracts leading options settings into an external setting, which was the
1907        cause of this.)
1908    
1909    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1910        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1911        assertion pattern did not match (meaning that the assertion was true), it
1912        was incorrectly treated as false if the SKIP had been reached during the
1913        matching. This also applied to assertions used as conditions.
1914    
1915    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1916        assertion subpattern, including such a pattern used as a condition,
1917        unpredictable results occurred, instead of the error return
1918        PCRE_ERROR_DFA_UITEM.
1919    
1920    5.  The C++ GlobalReplace function was not working like Perl for the special
1921        situation when an empty string is matched. It now does the fancy magic
1922        stuff that is necessary.
1923    
1924    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1925        removed. (These were left over from very, very early versions of PCRE.)
1926    
1927    7.  Some cosmetic changes to the code to make life easier when compiling it
1928        as part of something else:
1929    
1930        (a) Change DEBUG to PCRE_DEBUG.
1931    
1932        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1933            called "current" as "current_branch", to prevent a collision with the
1934            Linux macro when compiled as a kernel module.
1935    
1936        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1937            prevent a collision with the Linux macro when compiled as a kernel
1938            module.
1939    
1940    8.  In pcre_compile() there are some checks for integer overflows that used to
1941        cast potentially large values to (double). This has been changed to that
1942        when building, a check for int64_t is made, and if it is found, it is used
1943        instead, thus avoiding the use of floating point arithmetic. (There is no
1944        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1945        double.
1946    
1947    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1948        2005 (difference between two addresses compared to an unsigned value).
1949    
1950    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1951        custom one, because of the following reported problem in Windows:
1952    
1953          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1954              under Win32.
1955          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1956              therefore missing the function definition.
1957          - The compiler thus generates a "C" signature for the test function.
1958          - The linker fails to find the "C" function.
1959          - PCRE fails to configure if asked to do so against libbz2.
1960    
1961    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1962        messages were output:
1963    
1964          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1965          rerunning libtoolize, to keep the correct libtool macros in-tree.
1966          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1967    
1968        I have done both of these things.
1969    
1970    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1971        most of the time, it *can* run out if it is given a pattern that contains a
1972        runaway infinite recursion. I updated the discussion in the pcrestack man
1973        page.
1974    
1975    13. Now that we have gone to the x.xx style of version numbers, the minor
1976        version may start with zero. Using 08 or 09 is a bad idea because users
1977        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1978        interpreted as invalid octal numbers. I've updated the previous comment in
1979        configure.ac, and also added a check that gives an error if 08 or 09 are
1980        used.
1981    
1982    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1983        causing partial matching to fail when the end of the subject matched \W
1984        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1985    
1986    15. There were some discrepancies between the declarations in pcre_internal.h
1987        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1988        their definitions. The declarations used "const uschar *" and the
1989        definitions used USPTR. Even though USPTR is normally defined as "const
1990        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1991        reported that: "This difference in casting confuses some C++ compilers, for
1992        example, SunCC recognizes above declarations as different functions and
1993        generates broken code for hbpcre." I have changed the declarations to use
1994        USPTR.
1995    
1996    16. GNU libtool is named differently on some systems. The autogen.sh script now
1997        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1998        (FreeBSD).
1999    
2000    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2001        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2002        comment: "Figure out how to create a longlong from a string: strtoll and
2003        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2004        instance, but it only takes 2 args instead of 3!"
2005    
2006    18. A subtle bug concerned with back references has been fixed by a change of
2007        specification, with a corresponding code fix. A pattern such as
2008        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2009        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2010        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2011        same bug. Such groups have to be quantified to be useful, or contained
2012        inside another quantified group. (If there's no repetition, the reference
2013        can never match.) The problem arises because, having left the group and
2014        moved on to the rest of the pattern, a later failure that backtracks into
2015        the group uses the captured value from the final iteration of the group
2016        rather than the correct earlier one. I have fixed this in PCRE by forcing
2017        any group that contains a reference to itself to be an atomic group; that
2018        is, there cannot be any backtracking into it once it has completed. This is
2019        similar to recursive and subroutine calls.
2020    
2021    
2022    Version 8.00 19-Oct-09
2023    ----------------------
2024    
2025    1.  The table for translating pcre_compile() error codes into POSIX error codes
2026        was out-of-date, and there was no check on the pcre_compile() error code
2027        being within the table. This could lead to an OK return being given in
2028        error.
2029    
2030    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2031        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2032        in a Windows environment.
2033    
2034    3.  The pcregrep --count option prints the count for each file even when it is
2035        zero, as does GNU grep. However, pcregrep was also printing all files when
2036        --files-with-matches was added. Now, when both options are given, it prints
2037        counts only for those files that have at least one match. (GNU grep just
2038        prints the file name in this circumstance, but including the count seems
2039        more useful - otherwise, why use --count?) Also ensured that the
2040        combination -clh just lists non-zero counts, with no names.
2041    
2042    4.  The long form of the pcregrep -F option was incorrectly implemented as
2043        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2044        but it seems right to fix it, and I didn't think it was worth preserving
2045        the old behaviour.
2046    
2047    5.  The command line items --regex=pattern and --regexp=pattern were not
2048        recognized by pcregrep, which required --regex pattern or --regexp pattern
2049        (with a space rather than an '='). The man page documented the '=' forms,
2050        which are compatible with GNU grep; these now work.
2051    
2052    6.  No libpcreposix.pc file was created for pkg-config; there was just
2053        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2054    
2055    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2056        when UCP support is not needed, by modifying the Python script that
2057        generates it from Unicode data files. This should not matter if the module
2058        is correctly used as a library, but I received one complaint about 50K of
2059        unwanted data. My guess is that the person linked everything into his
2060        program rather than using a library. Anyway, it does no harm.
2061    
2062    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2063        was a minimum greater than 1 for a wide character in a possessive
2064        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2065        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2066        character. Chaos in the form of incorrect output or a compiling loop could
2067        result.
2068    
2069    9.  The restrictions on what a pattern can contain when partial matching is
2070        requested for pcre_exec() have been removed. All patterns can now be
2071        partially matched by this function. In addition, if there are at least two
2072        slots in the offset vector, the offset of the earliest inspected character
2073        for the match and the offset of the end of the subject are set in them when
2074        PCRE_ERROR_PARTIAL is returned.
2075    
2076    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2077        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2078        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2079        and may be more useful for multi-segment matching.
2080    
2081    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2082        used to be given if ever the end of the subject was reached; now it is
2083        given only if matching could not proceed because another character was
2084        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2085        string "Z", which now yields "no match" instead of "partial match". In the
2086        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2087        final character ended with (*FAIL).
2088    
2089    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2090        if the pattern had a "must contain" character that was already found in the
2091        earlier partial match, unless partial matching was again requested. For
2092        example, with the pattern /dog.(body)?/, the "must contain" character is
2093        "g". If the first part-match was for the string "dog", restarting with
2094        "sbody" failed. This bug has been fixed.
2095    
2096    13. The string returned by pcre_dfa_exec() after a partial match has been
2097        changed so that it starts at the first inspected character rather than the
2098        first character of the match. This makes a difference only if the pattern
2099        starts with a lookbehind assertion or \b or \B (\K is not supported by
2100        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2101        matching functions compatible, and I think it's the right thing to do.
2102    
2103    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2104        so that the demonstration program is easily available in environments where
2105        PCRE has not been installed from source.
2106    
2107    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2108        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2109        library.
2110    
2111    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2112        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2113        is not the first non-POSIX option to be added. Clearly some people find
2114        these options useful.
2115    
2116    17. If a caller to the POSIX matching function regexec() passes a non-zero
2117        value for nmatch with a NULL value for pmatch, the value of
2118        nmatch is forced to zero.
2119    
2120    18. RunGrepTest did not have a test for the availability of the -u option of
2121        the diff command, as RunTest does. It now checks in the same way as
2122        RunTest, and also checks for the -b option.
2123    
2124    19. If an odd number of negated classes containing just a single character
2125        interposed, within parentheses, between a forward reference to a named
2126        subpattern and the definition of the subpattern, compilation crashed with
2127        an internal error, complaining that it could not find the referenced
2128        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2129        [The bug was that it was starting one character too far in when skipping
2130        over the character class, thus treating the ] as data rather than
2131        terminating the class. This meant it could skip too much.]
2132    
2133    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2134        /g option in pcretest when the pattern contains \K, which makes it possible
2135        to have an empty string match not at the start, even when the pattern is
2136        anchored. Updated pcretest and pcredemo to use this option.
2137    
2138    21. If the maximum number of capturing subpatterns in a recursion was greater
2139        than the maximum at the outer level, the higher number was returned, but
2140        with unset values at the outer level. The correct (outer level) value is
2141        now given.
2142    
2143    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2144        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2145        make it do so. The string so far is captured, making this feature
2146        compatible with Perl.
2147    
2148    23. The tests have been re-organized, adding tests 11 and 12, to make it
2149        possible to check the Perl 5.10 features against Perl 5.10.
2150    
2151    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2152        pattern matches a fixed length string. PCRE did not allow this; now it
2153        does. Neither allows recursion.
2154    
2155    25. I finally figured out how to implement a request to provide the minimum
2156        length of subject string that was needed in order to match a given pattern.
2157        (It was back references and recursion that I had previously got hung up
2158        on.) This code has now been added to pcre_study(); it finds a lower bound
2159        to the length of subject needed. It is not necessarily the greatest lower
2160        bound, but using it to avoid searching strings that are too short does give
2161        some useful speed-ups. The value is available to calling programs via
2162        pcre_fullinfo().
2163    
2164    26. While implementing 25, I discovered to my embarrassment that pcretest had
2165        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2166        study optimizations had never been tested with that matching function.
2167        Oops. What is worse, even when it was passed study data, there was a bug in
2168        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2169        were also very few tests of studied patterns with pcre_dfa_exec().
2170    
2171    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2172        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2173        on the other side of the coin, they are no longer allowed to have different
2174        names, because these cannot be distinguished in PCRE, and this has caused
2175        confusion. (This is a difference from Perl.)
2176    
2177    28. When duplicate subpattern names are present (necessarily with different
2178        numbers, as required by 27 above), and a test is made by name in a
2179        conditional pattern, either for a subpattern having been matched, or for
2180        recursion in such a pattern, all the associated numbered subpatterns are
2181        tested, and the overall condition is true if the condition is true for any
2182        one of them. This is the way Perl works, and is also more like the way
2183        testing by number works.
2184    
2185    
2186    Version 7.9 11-Apr-09
2187    ---------------------
2188    
2189    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2190        (pcretest), all targets were linked against these libraries. This included
2191        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2192        libraries. This caused unwanted dependencies to be created. This problem
2193        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2194        pcretest is linked with readline.
2195    
2196    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2197        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2198        moved outside it again, because FALSE and TRUE are already defined in AIX,
2199        but BOOL is not.
2200    
2201    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2202        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2203    
2204    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2205        hyphens) following file names and line numbers when outputting matching
2206        lines. This is not true; no spaces are inserted. I have also clarified the
2207        wording for the --colour (or --color) option.
2208    
2209    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2210        was not coloured; this is different to GNU grep, so I have changed it to be
2211        the same.
2212    
2213    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2214        each matching line was coloured. Now it goes on to look for further matches
2215        of any of the test patterns, which is the same behaviour as GNU grep.
2216    
2217    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2218        doesn't make sense to accept an empty string match in pcregrep, so I have
2219        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2220        seems to be how GNU grep behaves. [But see later change 40 for release
2221        8.33.]
2222    
2223    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2224        start or after a newline", because the conditional assertion was not being
2225        correctly handled. The rule now is that both the assertion and what follows
2226        in the first alternative must satisfy the test.
2227    
2228    9.  If auto-callout was enabled in a pattern with a conditional group whose
2229        condition was an assertion, PCRE could crash during matching, both with
2230        pcre_exec() and pcre_dfa_exec().
2231    
2232    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2233        used for matching.
2234    
2235    11. Unicode property support in character classes was not working for
2236        characters (bytes) greater than 127 when not in UTF-8 mode.
2237    
2238    12. Added the -M command line option to pcretest.
2239    
2240    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2241    
2242    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2243    
2244    16. Added comments and documentation about mis-use of no_arg in the C++
2245        wrapper.
2246    
2247    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2248        from Martin Jerabek that uses macro names for all relevant character and
2249        string constants.
2250    
2251    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2252        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2253        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2254        these, but not everybody uses configure.
2255    
2256    19. A conditional group that had only one branch was not being correctly
2257        recognized as an item that could match an empty string. This meant that an
2258        enclosing group might also not be so recognized, causing infinite looping
2259        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2260        with the subject "ab", where knowledge that the repeated group can match
2261        nothing is needed in order to break the loop.
2262    
2263    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2264        exec(), but without supplying a callout function, matching went wrong.
2265    
2266    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2267        leak if the size of the offset vector was greater than 30. When the vector
2268        is smaller, the saved offsets during recursion go onto a local stack
2269        vector, but for larger vectors malloc() is used. It was failing to free
2270        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2271        error, in fact).
2272    
2273    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2274        heapframe that is used only when UTF-8 support is enabled. This caused no
2275        problem, but was untidy.
2276    
2277    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2278        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2279        included within another project.
2280    
2281    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2282        slightly modified by me:
2283    
2284          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2285              not building pcregrep.
2286    
2287          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2288              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2289    
2290    25. Forward references, both numeric and by name, in patterns that made use of
2291        duplicate group numbers, could behave incorrectly or give incorrect errors,
2292        because when scanning forward to find the reference group, PCRE was not
2293        taking into account the duplicate group numbers. A pattern such as
2294        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2295    
2296    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2297        the feature of a custom pointer more persuasive (as requested by a user).
2298    
2299    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2300        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2301        pre-defined.
2302    
2303    28. Added support for (*UTF8) at the start of a pattern.
2304    
2305    29. Arrange for flags added by the "release type" setting in CMake to be shown
2306        in the configuration summary.
2307    
2308    
2309    Version 7.8 05-Sep-08
2310    ---------------------
2311    
2312    1.  Replaced UCP searching code with optimized version as implemented for Ad
2313        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2314        stage table and inline lookup instead of a function, giving speed ups of 2
2315        to 5 times on some simple patterns that I tested. Permission was given to
2316        distribute the MultiStage2.py script that generates the tables (it's not in
2317        the tarball, but is in the Subversion repository).
2318    
2319    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2320        scripts.
2321    
2322    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2323        a group with a zero qualifier. The result of the study could be incorrect,
2324        or the function might crash, depending on the pattern.
2325    
2326    4.  Caseless matching was not working for non-ASCII characters in back
2327        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2328        It now works when Unicode Property Support is available.
2329    
2330    5.  In pcretest, an escape such as \x{de} in the data was always generating
2331        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2332        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2333        truncation.
2334    
2335    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2336    
2337    7.  Added two (int) casts to pcregrep when printing the difference of two
2338        pointers, in case they are 64-bit values.
2339    
2340    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2341        test 2 if it fails.
2342    
2343    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2344        and a #define of that name to empty if it is not externally set. This is to
2345        allow users of MSVC to set it if necessary.
2346    
2347    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2348        the convenience functions in the pcre_get.c source file.
2349    
2350    11. An option change at the start of a pattern that had top-level alternatives
2351        could cause overwriting and/or a crash. This command provoked a crash in
2352        some environments:
2353    
2354          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2355    
2356        This potential security problem was recorded as CVE-2008-2371.
2357    
2358    12. For a pattern where the match had to start at the beginning or immediately
2359        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2360        pcre_dfa_exec() could read past the end of the passed subject if there was
2361        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2362        pcretest so that it places the subject at the end of its malloc-ed buffer.
2363    
2364    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2365        exec() might read past the end of the data buffer in UTF-8 mode.
2366    
2367    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2368        the data contained the byte 0x85 as part of a UTF-8 character within its
2369        first line. This applied both to normal and DFA matching.
2370    
2371    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2372        /^[^d]*?$/8 failed to match "abc".
2373    
2374    16. Added a missing copyright notice to pcrecpp_internal.h.
2375    
2376    17. Make it more clear in the documentation that values returned from
2377        pcre_exec() in ovector are byte offsets, not character counts.
2378    
2379    18. Tidied a few places to stop certain compilers from issuing warnings.
2380    
2381    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2382        supplied by Stefan Weber. I made a further small update for 7.8 because
2383        there is a change of source arrangements: the pcre_searchfuncs.c module is
2384        replaced by pcre_ucd.c.
2385    
2386    
2387    Version 7.7 07-May-08
2388    ---------------------
2389    
2390    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2391        a string to a long long, pretend we don't even have a long long." This is
2392        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2393    
2394    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2395        pre-7.6 versions, which defined a global no_arg variable instead of putting
2396        it in the RE class. (See also #8 below.)
2397    
2398    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2399        Lopes.
2400    
2401    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2402    
2403        (1) The include/exclude patterns were being applied to the whole pathnames
2404            of files, instead of just to the final components.
2405    
2406        (2) If there was more than one level of directory, the subdirectories were
2407            skipped unless they satisfied the include/exclude conditions. This is
2408            inconsistent with GNU grep (and could even be seen as contrary to the
2409            pcregrep specification - which I improved to make it absolutely clear).
2410            The action now is always to scan all levels of directory, and just
2411            apply the include/exclude patterns to regular files.
2412    
2413    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2414        --exclude_dir in the tests to avoid scanning .svn directories.
2415    
2416    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2417        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2418        doesn't support NULs in patterns.
2419    
2420    7.  Added some missing "const"s to declarations of static tables in
2421        pcre_compile.c and pcre_dfa_exec.c.
2422    
2423    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2424        caused by fix #2  above. (Subsequently also a second patch to fix the
2425        first patch. And a third patch - this was a messy problem.)
2426    
2427    9.  Applied Craig's patch to remove the use of push_back().
2428    
2429    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2430        matching function regexec().
2431    
2432    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2433        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2434        references. PCRE supports relative numbers with this syntax (I don't think
2435        Oniguruma does).
2436    
2437    12. Previously, a group with a zero repeat such as (...){0} was completely
2438        omitted from the compiled regex. However, this means that if the group
2439        was called as a subroutine from elsewhere in the pattern, things went wrong
2440        (an internal error was given). Such groups are now left in the compiled
2441        pattern, with a new opcode that causes them to be skipped at execution
2442        time.
2443    
2444    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2445        to the way PCRE behaves:
2446    
2447        (a) A lone ] character is dis-allowed (Perl treats it as data).
2448    
2449        (b) A back reference to an unmatched subpattern matches an empty string
2450            (Perl fails the current match path).
2451    
2452        (c) A data ] in a character class must be notated as \] because if the
2453            first data character in a class is ], it defines an empty class. (In
2454            Perl it is not possible to have an empty class.) The empty class []
2455            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2456            The negative empty class [^] matches any one character, independently
2457            of the DOTALL setting.
2458    
2459    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2460        non-existent subpattern following a character class starting with ']' and
2461        containing () gave an internal compiling error instead of "reference to
2462        non-existent subpattern". Fortunately, when the pattern did exist, the
2463        compiled code was correct. (When scanning forwards to check for the
2464        existence of the subpattern, it was treating the data ']' as terminating
2465        the class, so got the count wrong. When actually compiling, the reference
2466        was subsequently set up correctly.)
2467    
2468    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2469        it was being rejected as not supported by pcre_dfa_exec(), even though
2470        other assertions are supported. I have made pcre_dfa_exec() support
2471        (*FAIL).
2472    
2473    16. The implementation of 13c above involved the invention of a new opcode,
2474        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2475        cannot be changed at match time, I realized I could make a small
2476        improvement to matching performance by compiling OP_ALLANY instead of
2477        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2478        on the OP_ANY path.
2479    
2480    17. Compiling pcretest on Windows with readline support failed without the
2481        following two fixes: (1) Make the unistd.h include conditional on
2482        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2483    
2484    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2485        ncurses library to be included for pcretest when ReadLine support is
2486        requested, but also to allow for it to be overridden. This patch came from
2487        Daniel Bergström.
2488    
2489    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2490        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2491        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2492        spotting this.
2493    
2494    
2495    Version 7.6 28-Jan-08
2496    ---------------------
2497    
2498    1.  A character class containing a very large number of characters with
2499        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2500        overflow.
2501    
2502    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2503        HAVE_LONG_LONG is not defined.
2504    
2505    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2506        bring them up to date and include new features. This patch includes:
2507    
2508        - Fixed PH's badly added libz and libbz2 support.
2509        - Fixed a problem with static linking.
2510        - Added pcredemo. [But later removed - see 7 below.]
2511        - Fixed dftables problem and added an option.
2512        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2513            HAVE_LONG_LONG.
2514        - Added readline support for pcretest.
2515        - Added an listing of the option settings after cmake has run.
2516    
2517    4.  A user submitted a patch to Makefile that makes it easy to create
2518        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2519        Makefile.am that cause it to include this special target, without
2520        affecting anything else. Note that the same mingw target plus all
2521        the other distribution libraries and programs are now supported
2522        when configuring with CMake (see 6 below) instead of with
2523        Configure/Make.
2524    
2525    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2526        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2527        exported in the Windows port". It has not yet been confirmed that the patch
2528        solves the problem, but it does no harm.
2529    
2530    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2531        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2532        with CMake, and also correct the comment about stack recursion.
2533    
2534    7.  Remove the automatic building of pcredemo from the ./configure system and
2535        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2536        of a program that users should build themselves after PCRE is installed, so
2537        building it automatically is not really right. What is more, it gave
2538        trouble in some build environments.
2539    
2540    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2541    
2542    
2543    Version 7.5 10-Jan-08
2544    ---------------------
2545    
2546    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2547        values in parens when parsing an RE using the C++ wrapper."
2548    
2549    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2550        Characters greater than 255 were excluded from the class instead of being
2551        included.
2552    
2553    3.  The same bug as (2) above applied to negated POSIX classes such as
2554        [:^space:].
2555    
2556    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2557        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2558        I have changed it.
2559    
2560    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2561        first named subpattern) and a construct such as (?&a) would reference the
2562        first named subpattern whose name started with "a" (in other words, the
2563        length check was missing). Both these problems are fixed. "Subpattern name
2564        expected" is now given for (?&) (a zero-length name), and this patch also
2565        makes it give the same error for \k'' (previously it complained that that
2566        was a reference to a non-existent subpattern).
2567    
2568    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2569        this is right because (?- can be followed by option settings as well as by
2570        digits. I have, however, made the messages clearer.
2571    
2572    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2573        than the number used in the conditional) now cause a compile-time error.
2574        This is actually not compatible with Perl, which accepts such patterns, but
2575        treats the conditional as always being FALSE (as PCRE used to), but it
2576        seems to me that giving a diagnostic is better.
2577    
2578    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2579        and messages.
2580    
2581    9.  Fix two occurrences of "backslash" in comments that should have been
2582        "backspace".
2583    
2584    10. Remove two redundant lines of code that can never be obeyed (their function
2585        was moved elsewhere).
2586    
2587    11. The program that makes PCRE's Unicode character property table had a bug
2588        which caused it to generate incorrect table entries for sequences of
2589        characters that have the same character type, but are in different scripts.
2590        It amalgamated them into a single range, with the script of the first of
2591        them. In other words, some characters were in the wrong script. There were
2592        thirteen such cases, affecting characters in the following ranges:
2593    
2594          U+002b0 - U+002c1
2595          U+0060c - U+0060d
2596          U+0061e - U+00612
2597          U+0064b - U+0065e
2598          U+0074d - U+0076d
2599          U+01800 - U+01805
2600          U+01d00 - U+01d77
2601          U+01d9b - U+01dbf
2602          U+0200b - U+0200f
2603          U+030fc - U+030fe
2604          U+03260 - U+0327f
2605          U+0fb46 - U+0fbb1
2606          U+10450 - U+1049d
2607    
2608    12. The -o option (show only the matching part of a line) for pcregrep was not
2609        compatible with GNU grep in that, if there was more than one match in a
2610        line, it showed only the first of them. It now behaves in the same way as
2611        GNU grep.
2612    
2613    13. If the -o and -v options were combined for pcregrep, it printed a blank
2614        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2615        does the same. The return code can be used to tell if there were any
2616        non-matching lines.
2617    
2618    14. Added --file-offsets and --line-offsets to pcregrep.
2619    
2620    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2621        infinitely looping recursion. The bug was that positive lookaheads were not
2622        being skipped when checking for a possible empty match (negative lookaheads
2623        and both kinds of lookbehind were skipped).
2624    
2625    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2626        inclusion of <windows.h> to before rather than after the definition of
2627        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2628    
2629    17. Specifying a possessive quantifier with a specific limit for a Unicode
2630        character property caused pcre_compile() to compile bad code, which led at
2631        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2632        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2633        caused the error; without that there was no problem.
2634    
2635    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2636    
2637    19. Added --enable-pcretest-libreadline.
2638    
2639    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2640        RE::GlobalReplace(). As a result, the number of replacements returned was
2641        double what it should be. I removed one of the increments, but Craig sent a
2642        later patch that removed the other one (the right fix) and added unit tests
2643        that check the return values (which was not done before).
2644    
2645    21. Several CMake things:
2646    
2647        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2648            the names libpcre and libpcreposix, not just pcre and pcreposix.
2649    
2650        (2) The above change means that pcretest and pcregrep are now correctly
2651            linked with the newly-built libraries, not previously installed ones.
2652    
2653        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2654    
2655    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2656        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2657        UTF-8 newline character). The key issue is that the pattern starts .*;
2658        this means that the match must be either at the beginning, or after a
2659        newline. The bug was in the code for advancing after a failed match and
2660        checking that the new position followed a newline. It was not taking
2661        account of UTF-8 characters correctly.
2662    
2663    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2664        character classes. PCRE was not treating the sequence [:...:] as a
2665        character class unless the ... were all letters. Perl, however, seems to
2666        allow any characters between [: and :], though of course it rejects as
2667        unknown any "names" that contain non-letters, because all the known class
2668        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2669        for example, whereas PCRE did not - it did not recognize a POSIX character
2670        class. This seemed a bit dangerous, so the code has been changed to be
2671        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2672        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2673        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2674        Perl does, and where it didn't before.
2675    
2676    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2677        Windows environments %n is disabled by default.
2678    
2679    
2680    Version 7.4 21-Sep-07
2681    ---------------------
2682    
2683    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2684        means that a class such as [\s] counted as "explicit reference to CR or
2685        LF". That isn't really right - the whole point of the change was to try to
2686        help when there was an actual mention of one of the two characters. So now
2687        the change happens only if \r or \n (or a literal CR or LF) character is
2688        encountered.
2689    
2690    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2691        of both had grown to the point where there were only 3 bits left.
2692        Fortunately, there was spare space in the data structure, and so I have
2693        moved the internal flags into a new 16-bit field to free up more option
2694        bits.
2695    
2696    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2697        but did not set the internal JCHANGED flag - either of these is enough to
2698        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2699        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2700        start sets both bits.
2701    
2702    4.  Added options (at build time, compile time, exec time) to change \R from
2703        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2704    
2705    5.  doc/pcresyntax.html was missing from the distribution.
2706    
2707    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2708        compatibility, even though it is no longer used.
2709    
2710    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2711        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2712        windows.h file is present (where different names are used). [This was
2713        reversed later after testing - see 16 below.]
2714    
2715    8.  Changed all #include <config.h> to #include "config.h". There were also
2716        some further <pcre.h> cases that I changed to "pcre.h".
2717    
2718    9.  When pcregrep was used with the --colour option, it missed the line ending
2719        sequence off the lines that it output.
2720    
2721    10. It was pointed out to me that arrays of string pointers cause lots of
2722        relocations when a shared library is dynamically loaded. A technique of
2723        using a single long string with a table of offsets can drastically reduce
2724        these. I have refactored PCRE in four places to do this. The result is
2725        dramatic:
2726    
2727          Originally:                          290
2728          After changing UCP table:            187
2729          After changing error message table:   43
2730          After changing table of "verbs"       36
2731          After changing table of Posix names   22
2732    
2733        Thanks to the folks working on Gregex for glib for this insight.
2734    
2735    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2736        unicode-properties was also set.
2737    
2738    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2739    
2740    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2741        checked only for CRLF.
2742    
2743    14. Added casts to pcretest.c to avoid compiler warnings.
2744    
2745    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2746    
2747    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2748        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2749        entirely. This removes changes made in 7 above.
2750    
2751    17. The CMake files have been updated, and there is now more information about
2752        building with CMake in the NON-UNIX-USE document.
2753    
2754    
2755    Version 7.3 28-Aug-07
2756    ---------------------
2757    
2758     1. In the rejigging of the build system that eventually resulted in 7.1, the
2759        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2760        brackets there is not right, since it causes compilers to look for an
2761        installed pcre.h, not the version that is in the source that is being
2762        compiled (which of course may be different). I have changed it back to:
2763    
2764          #include "pcre.h"
2765    
2766        I have a vague recollection that the change was concerned with compiling in
2767        different directories, but in the new build system, that is taken care of
2768        by the VPATH setting the Makefile.
2769    
2770     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2771        when the subject happened to end in the byte 0x85 (e.g. if the last
2772        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2773        characters but of course it shouldn't be taken as a newline when it is part
2774        of another character. The bug was that, for an unlimited repeat of . in
2775        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2776        characters when looking for a newline.
2777    
2778     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2779    
2780     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2781        in debug output.
2782    
2783     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2784        long printing in the pcrecpp unittest when running under MinGW.
2785    
2786     6. ESC_K was left out of the EBCDIC table.
2787    
2788     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2789        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2790        limit also applies to "virtual nesting" when a pattern is recursive, and in
2791        this case 1000 isn't so big. I have been able to remove this limit at the
2792        expense of backing off one optimization in certain circumstances. Normally,
2793        when pcre_exec() would call its internal match() function recursively and
2794        immediately return the result unconditionally, it uses a "tail recursion"
2795        feature to save stack. However, when a subpattern that can match an empty
2796        string has an unlimited repetition quantifier, it no longer makes this
2797        optimization. That gives it a stack frame in which to save the data for
2798        checking that an empty string has been matched. Previously this was taken
2799        from the 1000-entry workspace that had been reserved. So now there is no
2800        explicit limit, but more stack is used.
2801    
2802     8. Applied Daniel's patches to solve problems with the import/export magic
2803        syntax that is required for Windows, and which was going wrong for the
2804        pcreposix and pcrecpp parts of the library. These were overlooked when this
2805        problem was solved for the main library.
2806    
2807     9. There were some crude static tests to avoid integer overflow when computing
2808        the size of patterns that contain repeated groups with explicit upper
2809        limits. As the maximum quantifier is 65535, the maximum group length was
2810        set at 30,000 so that the product of these two numbers did not overflow a
2811        32-bit integer. However, it turns out that people want to use groups that
2812        are longer than 30,000 bytes (though not repeat them that many times).
2813        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2814        made it possible to implement the integer overflow checks in a much more
2815        dynamic way, which I have now done. The artificial limitation on group
2816        length has been removed - we now have only the limit on the total length of
2817        the compiled pattern, which depends on the LINK_SIZE setting.
2818    
2819    10. Fixed a bug in the documentation for get/copy named substring when
2820        duplicate names are permitted. If none of the named substrings are set, the
2821        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2822        empty string.
2823    
2824    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2825        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2826        because the ] is interpreted as the first data character and the
2827        terminating ] is not found. PCRE has been made compatible with Perl in this
2828        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2829        cause memory overwriting.
2830    
2831    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2832        string has been matched (to stop an infinite loop). It was not recognizing
2833        a conditional subpattern that could match an empty string if that
2834        subpattern was within another subpattern. For example, it looped when
2835        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2836        condition was not nested. This bug has been fixed.
2837    
2838    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2839        past the start of the subject in the presence of bytes with the top bit
2840        set, for example "\x8aBCD".
2841    
2842    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2843        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2844    
2845    14. Optimized (?!) to (*FAIL).
2846    
2847    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2848        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2849        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2850        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2851        does: it's just the validity check that is more restrictive.
2852    
2853    16. Inserted checks for integer overflows during escape sequence (backslash)
2854        processing, and also fixed erroneous offset values for syntax errors during
2855        backslash processing.
2856    
2857    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2858        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2859    
2860    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2861        caused an overrun.
2862    
2863    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2864        something other than just ASCII characters) inside a group that had an
2865        unlimited repeat caused a loop at compile time (while checking to see
2866        whether the group could match an empty string).
2867    
2868    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2869        [\P{Any}] did so. (Error in the code for printing property names.)
2870    
2871    21. An orphan \E inside a character class could cause a crash.
2872    
2873    22. A repeated capturing bracket such as (A)? could cause a wild memory
2874        reference during compilation.
2875    
2876    23. There are several functions in pcre_compile() that scan along a compiled
2877        expression for various reasons (e.g. to see if it's fixed length for look
2878        behind). There were bugs in these functions when a repeated \p or \P was
2879        present in the pattern. These operators have additional parameters compared
2880        with \d, etc, and these were not being taken into account when moving along
2881        the compiled data. Specifically:
2882    
2883        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2884            length.
2885    
2886        (b) An item such as \pL+ within a repeated group could cause crashes or
2887            loops.
2888    
2889        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2890            "reference to non-existent subpattern" error.
2891    
2892        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2893    
2894    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2895        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2896    
2897    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2898    
2899    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2900        character were causing crashes (broken optimization).
2901    
2902    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2903        \p or \P) caused a compile-time loop.
2904    
2905    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2906        break. For example, the unstudied pattern [\r\n]A does not match the string
2907        "\r\nA" because change 7.0/46 below moves the current point on by two
2908        characters after failing to match at the start. However, the pattern \nA
2909        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2910        the same is true. There doesn't seem any very clean way out of this, but
2911        what I have chosen to do makes the common cases work: PCRE now takes note
2912        of whether there can be an explicit match for \r or \n anywhere in the
2913        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2914        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2915        pattern has explicit CR or LF references.
2916    
2917    29. Added (*CR) etc for changing newline setting at start of pattern.
2918    
2919    
2920    Version 7.2 19-Jun-07
2921  ---------------------  ---------------------
2922    
2923   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 2937  Version 7.2 01-May-07
2937      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
2938      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
2939      in all cases.      in all cases.
2940    
2941   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
2942    
2943      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
2944          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
2945    
2946      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
2947          to be opened parentheses.          to be opened parentheses.
2948    
2949      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
2950          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
2951    
2952        (d) \K resets the start of the current match so that everything before
2953            is not part of it.
2954    
2955        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2956    
2957        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2958            reference syntax.
2959    
2960        (g) (?| introduces a group in which the numbering of parentheses in each
2961            alternative starts with the same number.
2962    
2963        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2964    
2965     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2966        PCRE_INFO_JCHANGED.
2967    
2968     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2969        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2970        for detecting groups that can match an empty string.
2971    
2972     9. A pattern with a very large number of alternatives (more than several
2973        hundred) was running out of internal workspace during the pre-compile
2974        phase, where pcre_compile() figures out how much memory will be needed. A
2975        bit of new cunning has reduced the workspace needed for groups with
2976        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2977        workspace instead of running out of the 4096 that are available.
2978    
2979    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2980    
2981    11. Applied patch from Google to remove an optimization that didn't quite work.
2982        The report of the bug said:
2983    
2984          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2985          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2986          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2987    
2988    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2989        it matched the wrong number of bytes.
2990    
2991    
2992  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
2993  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1535

  ViewVC Help
Powered by ViewVC 1.1.5