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

Diff of /code/trunk/ChangeLog

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

revision 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 1539 by ph10, Sun Mar 29 15:44:40 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-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    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
120        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
121        because \S ensures they are all in the class. The code for doing this was
122        interacting badly with the code for computing the amount of space needed to
123        compile the pattern, leading to a buffer overflow. This bug was discovered
124        by the LLVM fuzzer.
125    
126    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
127        other kinds of group caused stack overflow at compile time. This bug was
128        discovered by the LLVM fuzzer.
129    
130    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
131        between a subroutine call and its quantifier was incorrectly compiled,
132        leading to buffer overflow or other errors. This bug was discovered by the
133        LLVM fuzzer.
134    
135    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
136        assertion after (?(. The code was failing to check the character after
137        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
138        was discovered by the LLVM fuzzer.
139    
140    
141    Version 8.36 26-September-2014
142    ------------------------------
143    
144    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
145        -Wmissing-field-initializers and -Wunused-parameter.
146    
147    2.  The tests for quantifiers being too big (greater than 65535) were being
148        applied after reading the number, and stupidly assuming that integer
149        overflow would give a negative number. The tests are now applied as the
150        numbers are read.
151    
152    3.  Tidy code in pcre_exec.c where two branches that used to be different are
153        now the same.
154    
155    4.  The JIT compiler did not generate match limit checks for certain
156        bracketed expressions with quantifiers. This may lead to exponential
157        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
158        issue should be resolved now.
159    
160    5.  Fixed an issue, which occures when nested alternatives are optimized
161        with table jumps.
162    
163    6.  Inserted two casts and changed some ints to size_t in the light of some
164        reported 64-bit compiler warnings (Bugzilla 1477).
165    
166    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
167        an empty string, which sometimes were behaving incorrectly in the
168        interpreter (though correctly in the JIT matcher). This pcretest input is
169        an example:
170    
171          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
172          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
173    
174        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
175        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
176        for an empty string was breaking the inner loop and carrying on at a lower
177        level, when possessive repeated groups should always return to a higher
178        level as they have no backtrack points in them. The empty string test now
179        occurs at the outer level.
180    
181    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
182        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
183    
184    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
185        doesn't).
186    
187    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
188        to hold a pattern to be incorrectly computed (too small) when there were
189        named back references to duplicated names. This could cause "internal
190        error: code overflow" or "double free or corruption" or other memory
191        handling errors.
192    
193    11. When named subpatterns had the same prefixes, back references could be
194        confused. For example, in this pattern:
195    
196          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
197    
198        the reference to 'Name' was incorrectly treated as a reference to a
199        duplicate name.
200    
201    12. A pattern such as /^s?c/mi8 where the optional character has more than
202        one "other case" was incorrectly compiled such that it would only try to
203        match starting at "c".
204    
205    13. When a pattern starting with \s was studied, VT was not included in the
206        list of possible starting characters; this should have been part of the
207        8.34/18 patch.
208    
209    14. If a character class started [\Qx]... where x is any character, the class
210        was incorrectly terminated at the ].
211    
212    15. If a pattern that started with a caseless match for a character with more
213        than one "other case" was studied, PCRE did not set up the starting code
214        unit bit map for the list of possible characters. Now it does. This is an
215        optimization improvement, not a bug fix.
216    
217    16. The Unicode data tables have been updated to Unicode 7.0.0.
218    
219    17. Fixed a number of memory leaks in pcregrep.
220    
221    18. Avoid a compiler warning (from some compilers) for a function call with
222        a cast that removes "const" from an lvalue by using an intermediate
223        variable (to which the compiler does not object).
224    
225    19. Incorrect code was compiled if a group that contained an internal recursive
226        back reference was optional (had quantifier with a minimum of zero). This
227        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
228        caused segmentation faults because of stack overflows at compile time.
229    
230    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
231        group that is quantified with an indefinite repeat, caused a compile-time
232        loop which used up all the system stack and provoked a segmentation fault.
233        This was not the same bug as 19 above.
234    
235    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
236        Patch by Mike Frysinger.
237    
238    
239    Version 8.35 04-April-2014
240    --------------------------
241    
242    1.  A new flag is set, when property checks are present in an XCLASS.
243        When this flag is not set, PCRE can perform certain optimizations
244        such as studying these XCLASS-es.
245    
246    2.  The auto-possessification of character sets were improved: a normal
247        and an extended character set can be compared now. Furthermore
248        the JIT compiler optimizes more character set checks.
249    
250    3.  Got rid of some compiler warnings for potentially uninitialized variables
251        that show up only when compiled with -O2.
252    
253    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
254        of a match later then the end of the match. The pcretest program was not
255        handling the case sensibly - it was outputting from the start to the next
256        binary zero. It now reports this situation in a message, and outputs the
257        text from the end to the start.
258    
259    5.  Fast forward search is improved in JIT. Instead of the first three
260        characters, any three characters with fixed position can be searched.
261        Search order: first, last, middle.
262    
263    6.  Improve character range checks in JIT. Characters are read by an inprecise
264        function now, which returns with an unknown value if the character code is
265        above a certain threshold (e.g: 256). The only limitation is that the value
266        must be bigger than the threshold as well. This function is useful when
267        the characters above the threshold are handled in the same way.
268    
269    7.  The macros whose names start with RAWUCHAR are placeholders for a future
270        mode in which only the bottom 21 bits of 32-bit data items are used. To
271        make this more memorable for those maintaining the code, the names have
272        been changed to start with UCHAR21, and an extensive comment has been added
273        to their definition.
274    
275    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
276        to the export list in Makefile.am (they were accidentally omitted from the
277        8.34 tarball).
278    
279    9.  The informational output from pcretest used the phrase "starting byte set"
280        which is inappropriate for the 16-bit and 32-bit libraries. As the output
281        for "first char" and "need char" really means "non-UTF-char", I've changed
282        "byte" to "char", and slightly reworded the output. The documentation about
283        these values has also been (I hope) clarified.
284    
285    10. Another JIT related optimization: use table jumps for selecting the correct
286        backtracking path, when more than four alternatives are present inside a
287        bracket.
288    
289    11. Empty match is not possible, when the minimum length is greater than zero,
290        and there is no \K in the pattern. JIT should avoid empty match checks in
291        such cases.
292    
293    12. In a caseless character class with UCP support, when a character with more
294        than one alternative case was not the first character of a range, not all
295        the alternative cases were added to the class. For example, s and \x{17f}
296        are both alternative cases for S: the class [RST] was handled correctly,
297        but [R-T] was not.
298    
299    13. The configure.ac file always checked for pthread support when JIT was
300        enabled. This is not used in Windows, so I have put this test inside a
301        check for the presence of windows.h (which was already tested for).
302    
303    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
304        The algorithm provides a way to skip certain starting offsets, and usually
305        faster than linear prefix searches.
306    
307    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
308        as for 'fr_FR' and 'french'. For some reason, however, it then used the
309        Windows-specific input and output files, which have 'french' screwed in.
310        So this could never have worked. One of the problems with locales is that
311        they aren't always the same. I have now updated RunTest so that it checks
312        the output of the locale test (test 3) against three different output
313        files, and it allows the test to pass if any one of them matches. With luck
314        this should make the test pass on some versions of Solaris where it was
315        failing. Because of the uncertainty, the script did not used to stop if
316        test 3 failed; it now does. If further versions of a French locale ever
317        come to light, they can now easily be added.
318    
319    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
320        there was a message during ./configure, but it did not stop. This now
321        provokes an error. The invalid example in README has been corrected.
322        If a value less than the minimum is given, the minimum value has always
323        been used, but now a warning is given.
324    
325    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
326        was a bug in the test system, which is now fixed. Also, the list of various
327        configurations that are tested for each release did not have one with both
328        16/32 bits and --enable-bar-anycrlf. It now does.
329    
330    18. pcretest was missing "-C bsr" for displaying the \R default setting.
331    
332    19. Little endian PowerPC systems are supported now by the JIT compiler.
333    
334    20. The fast forward newline mechanism could enter to an infinite loop on
335        certain invalid UTF-8 input. Although we don't support these cases
336        this issue can be fixed by a performance optimization.
337    
338    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
339        not take account if existing stack usage. There is now a new global
340        variable called pcre_stack_guard that can be set to point to an external
341        function to check stack availability. It is called at the start of
342        processing every parenthesized group.
343    
344    22. A typo in the code meant that in ungreedy mode the max/min qualifier
345        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
346        not match "ab".
347    
348    23. When UTF was disabled, the JIT program reported some incorrect compile
349        errors. These messages are silenced now.
350    
351    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
352        compiler.
353    
354    25. Change all the temporary files used in RunGrepTest to be different to those
355        used by RunTest so that the tests can be run simultaneously, for example by
356        "make -j check".
357    
358    
359    Version 8.34 15-December-2013
360    -----------------------------
361    
362    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
363        executable memory. Patch inspired by Carsten Klein.
364    
365    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
366        this macro is never tested and has no effect, because the work to support
367        coverage involves only compiling and linking options and special targets in
368        the Makefile. The comment in config.h implied that defining the macro would
369        enable coverage support, which is totally false. There was also support for
370        setting this macro in the CMake files (my fault, I just copied it from
371        configure). SUPPORT_GCOV has now been removed.
372    
373    3.  Make a small performance improvement in strlen16() and strlen32() in
374        pcretest.
375    
376    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
377        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
378        these cases). There was also one in pcretest.c.
379    
380    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
381    
382    6.  In UTF mode, the code for checking whether a group could match an empty
383        string (which is used for indefinitely repeated groups to allow for
384        breaking an infinite loop) was broken when the group contained a repeated
385        negated single-character class with a character that occupied more than one
386        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
387        UTF-8 mode). The effect was undefined: the group might or might not be
388        deemed as matching an empty string, or the program might have crashed.
389    
390    7.  The code for checking whether a group could match an empty string was not
391        recognizing that \h, \H, \v, \V, and \R must match a character.
392    
393    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
394        an empty string. If it can, pcretest shows this in its information output.
395    
396    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
397        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
398        matched by pcre_exec() without using JIT:
399    
400        (a) If the rest of the pattern did not match after a maximal run of
401            grapheme clusters, the code for backing up to try with fewer of them
402            did not always back up over a full grapheme when characters that do not
403            have the modifier quality were involved, e.g. Hangul syllables.
404    
405        (b) If the match point in a subject started with modifier character, and
406            there was no match, the code could incorrectly back up beyond the match
407            point, and potentially beyond the first character in the subject,
408            leading to a segfault or an incorrect match result.
409    
410    10. A conditional group with an assertion condition could lead to PCRE
411        recording an incorrect first data item for a match if no other first data
412        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
413        first data item, and therefore matched "ca" after "c" instead of at the
414        start.
415    
416    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
417        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
418    
419    12. The source of pcregrep now includes z/OS-specific code so that it can be
420        compiled for z/OS as part of the special z/OS distribution.
421    
422    13. Added the -T and -TM options to pcretest.
423    
424    14. The code in pcre_compile.c for creating the table of named capturing groups
425        has been refactored. Instead of creating the table dynamically during the
426        actual compiling pass, the information is remembered during the pre-compile
427        pass (on the stack unless there are more than 20 named groups, in which
428        case malloc() is used) and the whole table is created before the actual
429        compile happens. This has simplified the code (it is now nearly 150 lines
430        shorter) and prepared the way for better handling of references to groups
431        with duplicate names.
432    
433    15. A back reference to a named subpattern when there is more than one of the
434        same name now checks them in the order in which they appear in the pattern.
435        The first one that is set is used for the reference. Previously only the
436        first one was inspected. This change makes PCRE more compatible with Perl.
437    
438    16. Unicode character properties were updated from Unicode 6.3.0.
439    
440    17. The compile-time code for auto-possessification has been refactored, based
441        on a patch by Zoltan Herczeg. It now happens after instead of during
442        compilation. The code is cleaner, and more cases are handled. The option
443        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
444        options in pcretest are provided to set it. It can also be set by
445        (*NO_AUTO_POSSESS) at the start of a pattern.
446    
447    18. The character VT has been added to the default ("C" locale) set of
448        characters that match \s and are generally treated as white space,
449        following this same change in Perl 5.18. There is now no difference between
450        "Perl space" and "POSIX space". Whether VT is treated as white space in
451        other locales depends on the locale.
452    
453    19. The code for checking named groups as conditions, either for being set or
454        for being recursed, has been refactored (this is related to 14 and 15
455        above). Processing unduplicated named groups should now be as fast at
456        numerical groups, and processing duplicated groups should be faster than
457        before.
458    
459    20. Two patches to the CMake build system, by Alexander Barkov:
460    
461          (1) Replace the "source" command by "." in CMakeLists.txt because
462              "source" is a bash-ism.
463    
464          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
465              without these the CMake build does not work on Solaris.
466    
467    21. Perl has changed its handling of \8 and \9. If there is no previously
468        encountered capturing group of those numbers, they are treated as the
469        literal characters 8 and 9 instead of a binary zero followed by the
470        literals. PCRE now does the same.
471    
472    22. Following Perl, added \o{} to specify codepoints in octal, making it
473        possible to specify values greater than 0777 and also making them
474        unambiguous.
475    
476    23. Perl now gives an error for missing closing braces after \x{... instead of
477        treating the string as literal. PCRE now does the same.
478    
479    24. RunTest used to grumble if an inappropriate test was selected explicitly,
480        but just skip it when running all tests. This make it awkward to run ranges
481        of tests when one of them was inappropriate. Now it just skips any
482        inappropriate tests, as it always did when running all tests.
483    
484    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
485        character types such as \d or \w, too many callouts were inserted, and the
486        data that they returned was rubbish.
487    
488    26. In UCP mode, \s was not matching two of the characters that Perl matches,
489        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
490        were matched by \h. The code has now been refactored so that the lists of
491        the horizontal and vertical whitespace characters used for \h and \v (which
492        are defined only in one place) are now also used for \s.
493    
494    27. Add JIT support for the 64 bit TileGX architecture.
495        Patch by Jiong Wang (Tilera Corporation).
496    
497    28. Possessive quantifiers for classes (both explicit and automatically
498        generated) now use special opcodes instead of wrapping in ONCE brackets.
499    
500    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
501        (because it's meaningless), this was not happening when PCRE_CASELESS was
502        set. Not wrong, but inefficient.
503    
504    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
505        properties for \w, \d, etc) is present in a test regex. Otherwise if the
506        test contains no characters greater than 255, Perl doesn't realise it
507        should be using Unicode semantics.
508    
509    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
510        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
511        does in Unicode mode.
512    
513    32. Added the "forbid" facility to pcretest so that putting tests into the
514        wrong test files can sometimes be quickly detected.
515    
516    33. There is now a limit (default 250) on the depth of nesting of parentheses.
517        This limit is imposed to control the amount of system stack used at compile
518        time. It can be changed at build time by --with-parens-nest-limit=xxx or
519        the equivalent in CMake.
520    
521    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
522        errors. Perl warns for these when in warning mode, but PCRE has no facility
523        for giving warnings.
524    
525    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
526        However, this was not working for (?!) because it is optimized to (*FAIL),
527        for which PCRE does not allow quantifiers. The optimization is now disabled
528        when a quantifier follows (?!). I can't see any use for this, but it makes
529        things uniform.
530    
531    36. Perl no longer allows group names to start with digits, so I have made this
532        change also in PCRE. It simplifies the code a bit.
533    
534    37. In extended mode, Perl ignores spaces before a + that indicates a
535        possessive quantifier. PCRE allowed a space before the quantifier, but not
536        before the possessive +. It now does.
537    
538    38. The use of \K (reset reported match start) within a repeated possessive
539        group such as (a\Kb)*+ was not working.
540    
541    40. Document that the same character tables must be used at compile time and
542        run time, and that the facility to pass tables to pcre_exec() and
543        pcre_dfa_exec() is for use only with saved/restored patterns.
544    
545    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
546        features for Builds with MSVC:
547    
548        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
549           builds). The .rc files can be used to set FileDescription and many other
550           attributes.
551    
552        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
553           This allows higher-level build scripts which want .pdb files to avoid
554           hard-coding the exact files needed."
555    
556    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
557        mean "start of word" and "end of word", respectively, as a transition aid.
558    
559    43. A minimizing repeat of a class containing codepoints greater than 255 in
560        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
561        compiled to use the heap for recursion.
562    
563    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
564        is configured.
565    
566    
567    Version 8.33 28-May-2013
568    ------------------------
569    
570    1.  Added 'U' to some constants that are compared to unsigned integers, to
571        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
572        variables that are added to signed variables, to ensure the result is
573        signed and can be negated.
574    
575    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
576        CMake config files.
577    
578    3.  Revise the creation of config.h.generic so that all boolean macros are
579        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
580        overriding via -D on the command line possible.
581    
582    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
583        to unsigned int is reported to make a quite noticeable speed difference in
584        a specific Windows environment. Testing on Linux did also appear to show
585        some benefit (and it is clearly not harmful). Also fixed the definition of
586        Xop which should be unsigned.
587    
588    5.  Related to (4), changing the definition of the intermediate variable cc
589        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
590        improvements.
591    
592    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
593        unnecessary spaces.
594    
595    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
596        and later.
597    
598    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
599    
600    9.  Optimizing fast_forward_start_bits in JIT.
601    
602    10. Adding support for callouts in JIT, and fixing some issues revealed
603        during this work. Namely:
604    
605        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
606    
607        (b) Minimum length was not checked before the matching is started.
608    
609    11. The value of capture_last that is passed to callouts was incorrect in some
610        cases when there was a capture on one path that was subsequently abandoned
611        after a backtrack. Also, the capture_last value is now reset after a
612        recursion, since all captures are also reset in this case.
613    
614    12. The interpreter no longer returns the "too many substrings" error in the
615        case when an overflowing capture is in a branch that is subsequently
616        abandoned after a backtrack.
617    
618    13. In the pathological case when an offset vector of size 2 is used, pcretest
619        now prints out the matched string after a yield of 0 or 1.
620    
621    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
622        Only supported by the JIT compiler at the moment.
623    
624    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
625    
626    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
627        offset of the starting point of the matching process, provided the offsets
628        vector is large enough.
629    
630    17. The \A escape now records a lookbehind value of 1, though its execution
631        does not actually inspect the previous character. This is to ensure that,
632        in partial multi-segment matching, at least one character from the old
633        segment is retained when a new segment is processed. Otherwise, if there
634        are no lookbehinds in the pattern, \A might match incorrectly at the start
635        of a new segment.
636    
637    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
638    
639    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
640        gives some modest performance improvement in 8-bit mode.
641    
642    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
643        be expressed in certain programming languages using Universal Character
644        Names.
645    
646    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
647        which points out that "non characters" are not "characters that may not
648        appear in Unicode strings" but rather "characters that are reserved for
649        internal use and have only local meaning".
650    
651    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
652        there was a conditional group that depended on an assertion, if the
653        assertion was false, the callout that immediately followed the alternation
654        in the condition was skipped when pcre_exec() was used for matching.
655    
656    23. Allow an explicit callout to be inserted before an assertion that is the
657        condition for a conditional group, for compatibility with automatic
658        callouts, which always insert a callout at this point.
659    
660    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
661        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
662    
663    25. (*PRUNE) is now supported by the JIT compiler.
664    
665    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
666    
667    27. Fix the case where there are two or more SKIPs with arguments that may be
668        ignored.
669    
670    28. (*SKIP) is now supported by the JIT compiler.
671    
672    29. (*THEN) is now supported by the JIT compiler.
673    
674    30. Update RunTest with additional test selector options.
675    
676    31. The way PCRE handles backtracking verbs has been changed in two ways.
677    
678        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
679        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
680        backtracking. In some cases this makes it more Perl-compatible, but Perl's
681        rather obscure rules do not always do the same thing.
682    
683        (2) Previously, backtracking verbs were confined within assertions. This is
684        no longer the case for positive assertions, except for (*ACCEPT). Again,
685        this sometimes improves Perl compatibility, and sometimes does not.
686    
687    32. A number of tests that were in test 2 because Perl did things differently
688        have been moved to test 1, because either Perl or PCRE has changed, and
689        these tests are now compatible.
690    
691    32. Backtracking control verbs are now handled in the same way in JIT and
692        interpreter.
693    
694    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
695        contained a forward subroutine reference caused a compile error.
696    
697    34. Auto-detect and optimize limited repetitions in JIT.
698    
699    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
700        blocking (*UTF) etc.
701    
702    36. In the interpreter, maximizing pattern repetitions for characters and
703        character types now use tail recursion, which reduces stack usage.
704    
705    37. The value of the max lookbehind was not correctly preserved if a compiled
706        and saved regex was reloaded on a host of different endianness.
707    
708    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
709        of the compiled pattern block, expand the flags field from 16 to 32 bits
710        because it was almost full.
711    
712    39. Try madvise first before posix_madvise.
713    
714    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
715        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
716        original change locked out all matches of empty strings. This has been
717        changed so that one match of an empty string per line is recognized.
718        Subsequent searches on the same line (for colouring or for --only-matching,
719        for example) do not recognize empty strings.
720    
721    41. Applied a user patch to fix a number of spelling mistakes in comments.
722    
723    42. Data lines longer than 65536 caused pcretest to crash.
724    
725    43. Clarified the data type for length and startoffset arguments for pcre_exec
726        and pcre_dfa_exec in the function-specific man pages, where they were
727        explicitly stated to be in bytes, never having been updated. I also added
728        some clarification to the pcreapi man page.
729    
730    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
731        a segmentation fault.
732    
733    
734    Version 8.32 30-November-2012
735    -----------------------------
736    
737    1.  Improved JIT compiler optimizations for first character search and single
738        character iterators.
739    
740    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
741        Patch by Daniel Richard G.
742    
743    3.  Single character iterator optimizations in the JIT compiler.
744    
745    4.  Improved JIT compiler optimizations for character ranges.
746    
747    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
748        Reported by Giuseppe D'Angelo.
749    
750    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
751        of a line, was being set incorrectly in cases where .* appeared inside
752        atomic brackets at the start of a pattern, or where there was a subsequent
753        *PRUNE or *SKIP.
754    
755    7.  Improved instruction cache flush for POWER/PowerPC.
756        Patch by Daniel Richard G.
757    
758    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
759        grep:
760    
761        (a) There is now no limit to the number of patterns to be matched.
762    
763        (b) An error is given if a pattern is too long.
764    
765        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
766            are now supported.
767    
768        (d) --exclude-from and --include-from (multiple use) have been added.
769    
770        (e) Exclusions and inclusions now apply to all files and directories, not
771            just to those obtained from scanning a directory recursively.
772    
773        (f) Multiple uses of -f and --file-list are now supported.
774    
775        (g) In a Windows environment, the default for -d has been changed from
776            "read" (the GNU grep default) to "skip", because otherwise the presence
777            of a directory in the file list provokes an error.
778    
779        (h) The documentation has been revised and clarified in places.
780    
781    9.  Improve the matching speed of capturing brackets.
782    
783    10. Changed the meaning of \X so that it now matches a Unicode extended
784        grapheme cluster.
785    
786    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
787        out POSIX threads when JIT support is configured.
788    
789    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
790    
791    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
792        structure could go wrong in environments where size_t is not the same size
793        as int.
794    
795    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
796        set.
797    
798    15. The EBCDIC support had decayed; later updates to the code had included
799        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
800        general tidy up of EBCDIC-related issues, and the documentation was also
801        not quite right. There is now a test that can be run on ASCII systems to
802        check some of the EBCDIC-related things (but is it not a full test).
803    
804    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
805        in a small tidy to the code.
806    
807    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
808    
809    18. If the --only-matching (-o) option in pcregrep is specified multiple
810        times, each one causes appropriate output. For example, -o1 -o2 outputs the
811        substrings matched by the 1st and 2nd capturing parentheses. A separating
812        string can be specified by --om-separator (default empty).
813    
814    19. Improving the first n character searches.
815    
816    20. Turn case lists for horizontal and vertical white space into macros so that
817        they are defined only once.
818    
819    21. This set of changes together give more compatible Unicode case-folding
820        behaviour for characters that have more than one other case when UCP
821        support is available.
822    
823        (a) The Unicode property table now has offsets into a new table of sets of
824            three or more characters that are case-equivalent. The MultiStage2.py
825            script that generates these tables (the pcre_ucd.c file) now scans
826            CaseFolding.txt instead of UnicodeData.txt for character case
827            information.
828    
829        (b) The code for adding characters or ranges of characters to a character
830            class has been abstracted into a generalized function that also handles
831            case-independence. In UTF-mode with UCP support, this uses the new data
832            to handle characters with more than one other case.
833    
834        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
835            whose other case is greater than 256 are now correctly matched
836            caselessly. Previously, the high codepoint matched the low one, but not
837            vice versa.
838    
839        (d) The processing of \h, \H, \v, and \ in character classes now makes use
840            of the new class addition function, using character lists defined as
841            macros alongside the case definitions of 20 above.
842    
843        (e) Caseless back references now work with characters that have more than
844            one other case.
845    
846        (f) General caseless matching of characters with more than one other case
847            is supported.
848    
849    22. Unicode character properties were updated from Unicode 6.2.0
850    
851    23. Improved CMake support under Windows. Patch by Daniel Richard G.
852    
853    24. Add support for 32-bit character strings, and UTF-32
854    
855    25. Major JIT compiler update (code refactoring and bugfixing).
856        Experimental Sparc 32 support is added.
857    
858    26. Applied a modified version of Daniel Richard G's patch to create
859        pcre.h.generic and config.h.generic by "make" instead of in the
860        PrepareRelease script.
861    
862    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
863        pcre_compile.c when checking for a zero character.
864    
865    28. Introducing a native interface for JIT. Through this interface, the compiled
866        machine code can be directly executed. The purpose of this interface is to
867        provide fast pattern matching, so several sanity checks are not performed.
868        However, feature tests are still performed. The new interface provides
869        1.4x speedup compared to the old one.
870    
871    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
872        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
873        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
874    
875    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
876        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
877        the "old" RFC 2279). Instead, it ended up passing a negative length to
878        pcre_exec().
879    
880    31. Add support for GCC's visibility feature to hide internal functions.
881    
882    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
883        "unknown -C option" after outputting 0 or 1.
884    
885    33. There is now support for generating a code coverage report for the test
886        suite in environments where gcc is the compiler and lcov is installed. This
887        is mainly for the benefit of the developers.
888    
889    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
890        unaddressable using valgrind annotations, allowing valgrind to detect
891        invalid memory accesses. This is mainly for the benefit of the developers.
892    
893    25. (*UTF) can now be used to start a pattern in any of the three libraries.
894    
895    26. Give configure error if --enable-cpp but no C++ compiler found.
896    
897    
898    Version 8.31 06-July-2012
899    -------------------------
900    
901    1.  Fixing a wrong JIT test case and some compiler warnings.
902    
903    2.  Removed a bashism from the RunTest script.
904    
905    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
906        to unsigned type, result still unsigned" that was given by an MS compiler
907        on encountering the code "-sizeof(xxx)".
908    
909    4.  Partial matching support is added to the JIT compiler.
910    
911    5.  Fixed several bugs concerned with partial matching of items that consist
912        of more than one character:
913    
914        (a) /^(..)\1/ did not partially match "aba" because checking references was
915            done on an "all or nothing" basis. This also applied to repeated
916            references.
917    
918        (b) \R did not give a hard partial match if \r was found at the end of the
919            subject.
920    
921        (c) \X did not give a hard partial match after matching one or more
922            characters at the end of the subject.
923    
924        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
925            a partial match for the string "\r".
926    
927        (e) When newline was set to CRLF, the metacharacter "." did not recognize
928            a partial match for a CR character at the end of the subject string.
929    
930    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
931        running pcretest, the text "(JIT)" added to the output whenever JIT is
932        actually used to run the match.
933    
934    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
935        or /S+[+] with a digit between 1 and 7.
936    
937    8.  OP_NOT now supports any UTF character not just single-byte ones.
938    
939    9.  (*MARK) control verb is now supported by the JIT compiler.
940    
941    10. The command "./RunTest list" lists the available tests without actually
942        running any of them. (Because I keep forgetting what they all are.)
943    
944    11. Add PCRE_INFO_MAXLOOKBEHIND.
945    
946    12. Applied a (slightly modified) user-supplied patch that improves performance
947        when the heap is used for recursion (compiled with --disable-stack-for-
948        recursion). Instead of malloc and free for each heap frame each time a
949        logical recursion happens, frames are retained on a chain and re-used where
950        possible. This sometimes gives as much as 30% improvement.
951    
952    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
953        call.
954    
955    14. As documented, (*COMMIT) is now confined to within a positive assertion.
956    
957    15. It is now possible to link pcretest with libedit as an alternative to
958        libreadline.
959    
960    16. (*COMMIT) control verb is now supported by the JIT compiler.
961    
962    17. The Unicode data tables have been updated to Unicode 6.1.0.
963    
964    18. Added --file-list option to pcregrep.
965    
966    19. Added binary file support to pcregrep, including the -a, --binary-files,
967        -I, and --text options.
968    
969    20. The madvise function is renamed for posix_madvise for QNX compatibility
970        reasons. Fixed by Giuseppe D'Angelo.
971    
972    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
973    
974    22. Changed the option for creating symbolic links for 16-bit man pages from
975        -s to -sf so that re-installing does not cause issues.
976    
977    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
978    
979    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
980        matches in certain environments (the workspace was not being correctly
981        retained). Also added to pcre_dfa_exec() a simple plausibility check on
982        some of the workspace data at the beginning of a restart.
983    
984    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
985        was not doing so when it should - probably a typo introduced by SVN 528
986        (change 8.10/14).
987    
988    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
989        \w+ when the character tables indicated that \x{c4} was a word character.
990        There were several related cases, all because the tests for doing a table
991        lookup were testing for characters less than 127 instead of 255.
992    
993    27. If a pattern contains capturing parentheses that are not used in a match,
994        their slots in the ovector are set to -1. For those that are higher than
995        any matched groups, this happens at the end of processing. In the case when
996        there were back references that the ovector was too small to contain
997        (causing temporary malloc'd memory to be used during matching), and the
998        highest capturing number was not used, memory off the end of the ovector
999        was incorrectly being set to -1. (It was using the size of the temporary
1000        memory instead of the true size.)
1001    
1002    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1003        ovector size, it uses memory at the end of the block that it has got.
1004    
1005    29. Check for an overlong MARK name and give an error at compile time. The
1006        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1007    
1008    30. JIT compiler update.
1009    
1010    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1011        Rill for the patch.
1012    
1013    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1014    
1015    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1016    
1017    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1018        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1019        was enabled.
1020    
1021    35. Improve JIT code generation for greedy plus quantifier.
1022    
1023    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1024        1 to "aa" instead of to an empty string. The bug affected repeated groups
1025        that could potentially match an empty string.
1026    
1027    37. Optimizing single character iterators in JIT.
1028    
1029    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1030        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1031        codepoints that are too big for the mode are faulted, and in a UTF mode,
1032        disallowed codepoints are also faulted.
1033    
1034    39. If PCRE was compiled with UTF support, in three places in the DFA
1035        matcher there was code that should only have been obeyed in UTF mode, but
1036        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1037        processing when bytes with values greater than 127 were present. In 16-bit
1038        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1039        both cases the values are those that cannot be the first data item in a UTF
1040        character. The three items that might have provoked this were recursions,
1041        possessively repeated groups, and atomic groups.
1042    
1043    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1044        and pcregrep, because some OS require shared objects to be explicitly
1045        passed to ld, causing the link step to fail if they are not.
1046    
1047    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1048        mode, patterns that started with \h* or \R* might be incorrectly matched.
1049    
1050    
1051    Version 8.30 04-February-2012
1052    -----------------------------
1053    
1054    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1055        name is defined in ctype.h.
1056    
1057    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1058        only in quite long subpatterns.
1059    
1060    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1061        since it was replaced by pcre_fullinfo() in February 2000.
1062    
1063    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1064        match a (*MARK), and the match failed at the start of the subject, a
1065        reference to memory before the start of the subject could occur. This bug
1066        was introduced by fix 17 of release 8.21.
1067    
1068    5.  A reference to an unset group with zero minimum repetition was giving
1069        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1070        /(another)?(\1?)test/ matched against "hello world test". This bug was
1071        introduced in release 8.13.
1072    
1073    6.  Add support for 16-bit character strings (a large amount of work involving
1074        many changes and refactorings).
1075    
1076    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1077        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1078        from a file.
1079    
1080    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1081        rounding is not applied in this particular case).
1082    
1083    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1084        if they appear, or are escaped, in patterns.
1085    
1086    10. Get rid of a number of -Wunused-but-set-variable warnings.
1087    
1088    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1089        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1090        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1091        also returns the mark "x". This bug applied to capturing parentheses,
1092        non-capturing parentheses, and atomic parentheses. It also applied to some
1093        assertions.
1094    
1095    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1096        information out of configure.ac instead of relying on pcre.h.generic, which
1097        is not stored in the repository.
1098    
1099    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1100        -lreadline.
1101    
1102    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1103    
1104    15. Applied Graycode's patch to put the top-level frame on the stack rather
1105        than the heap when not using the stack for recursion. This gives a
1106        performance improvement in many cases when recursion is not deep.
1107    
1108    16. Experimental code added to "pcretest -C" to output the stack frame size.
1109    
1110    
1111    Version 8.21 12-Dec-2011
1112    ------------------------
1113    
1114    1.  Updating the JIT compiler.
1115    
1116    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1117        are added as well.
1118    
1119    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1120        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1121        calling _pcre_jit_exec. Some extra comments are added.
1122    
1123    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1124        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1125        was introduced by change 18 for 8.20.
1126    
1127    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1128        ECMA-262 standard.
1129    
1130    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1131        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1132        This bug was probably introduced by change 9 of 8.13.
1133    
1134    7.  While fixing 6 above, I noticed that a number of other items were being
1135        incorrectly rejected as "not fixed length". This arose partly because newer
1136        opcodes had not been added to the fixed-length checking code. I have (a)
1137        corrected the bug and added tests for these items, and (b) arranged for an
1138        error to occur if an unknown opcode is encountered while checking for fixed
1139        length instead of just assuming "not fixed length". The items that were
1140        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1141        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1142        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1143    
1144    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1145        being incorrectly compiled and would have given unpredicatble results.
1146    
1147    9.  A possessively repeated subpattern with minimum repeat count greater than
1148        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1149        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1150        the first (A) could occur when it should not.
1151    
1152    10. Add a cast and remove a redundant test from the code.
1153    
1154    11. JIT should use pcre_malloc/pcre_free for allocation.
1155    
1156    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1157        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1158        is anything other than /usr, -L is still shown).
1159    
1160    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1161    
1162    14. Perl does not support \N without a following name in a [] class; PCRE now
1163        also gives an error.
1164    
1165    15. If a forward reference was repeated with an upper limit of around 2000,
1166        it caused the error "internal error: overran compiling workspace". The
1167        maximum number of forward references (including repeats) was limited by the
1168        internal workspace, and dependent on the LINK_SIZE. The code has been
1169        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1170        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1171        of around 200,000 forward references. While doing this, I also speeded up
1172        the filling in of repeated forward references.
1173    
1174    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1175        incorrectly expecting the subject to contain another "a" after the start.
1176    
1177    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1178        in the match, the SKIP should be ignored. This was not happening; instead
1179        the SKIP was being treated as NOMATCH. For patterns such as
1180        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1181        tested.
1182    
1183    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1184        now much more compatible with Perl, in particular in cases where the result
1185        is a non-match for a non-anchored pattern. For example, if
1186        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1187        "m", where previously it did not return a name. A side effect of this
1188        change is that for partial matches, the last encountered mark name is
1189        returned, as for non matches. A number of tests that were previously not
1190        Perl-compatible have been moved into the Perl-compatible test files. The
1191        refactoring has had the pleasing side effect of removing one argument from
1192        the match() function, thus reducing its stack requirements.
1193    
1194    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1195        subsequent uses of /S (without +) incorrectly behaved like /S+.
1196    
1197    21. Retrieve executable code size support for the JIT compiler and fixing
1198        some warnings.
1199    
1200    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1201        not work when the shorter character appeared right at the end of the
1202        subject string.
1203    
1204    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1205        systems.
1206    
1207    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1208        output it when the /M option is used in pcretest.
1209    
1210    25. The CheckMan script was not being included in the distribution. Also, added
1211        an explicit "perl" to run Perl scripts from the PrepareRelease script
1212        because this is reportedly needed in Windows.
1213    
1214    26. If study data was being save in a file and studying had not found a set of
1215        "starts with" bytes for the pattern, the data written to the file (though
1216        never used) was taken from uninitialized memory and so caused valgrind to
1217        complain.
1218    
1219    27. Updated RunTest.bat as provided by Sheri Pierce.
1220    
1221    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1222    
1223    29. Computation of memory usage for the table of capturing group names was
1224        giving an unnecessarily large value.
1225    
1226    
1227    Version 8.20 21-Oct-2011
1228    ------------------------
1229    
1230    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1231        a POSIX class. After further experiments with Perl, which convinced me that
1232        Perl has bugs and confusions, a closing square bracket is no longer allowed
1233        in a POSIX name. This bug also affected patterns with classes that started
1234        with full stops.
1235    
1236    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1237        captured substring, but while checking the failing first alternative,
1238        substring 1 is temporarily captured. If the output vector supplied to
1239        pcre_exec() was not big enough for this capture, the yield of the function
1240        was still zero ("insufficient space for captured substrings"). This cannot
1241        be totally fixed without adding another stack variable, which seems a lot
1242        of expense for a edge case. However, I have improved the situation in cases
1243        such as /(a)(b)x|abc/ matched against "abc", where the return code
1244        indicates that fewer than the maximum number of slots in the ovector have
1245        been set.
1246    
1247    3.  Related to (2) above: when there are more back references in a pattern than
1248        slots in the output vector, pcre_exec() uses temporary memory during
1249        matching, and copies in the captures as far as possible afterwards. It was
1250        using the entire output vector, but this conflicts with the specification
1251        that only 2/3 is used for passing back captured substrings. Now it uses
1252        only the first 2/3, for compatibility. This is, of course, another edge
1253        case.
1254    
1255    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1256        main code base, and can be used by building with --enable-jit. When this is
1257        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1258        runtime --no-jit option is given.
1259    
1260    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1261        ovector, the return from the function was zero, implying that there were
1262        other matches that did not fit. The correct "exactly full" value is now
1263        returned.
1264    
1265    6.  If a subpattern that was called recursively or as a subroutine contained
1266        (*PRUNE) or any other control that caused it to give a non-standard return,
1267        invalid errors such as "Error -26 (nested recursion at the same subject
1268        position)" or even infinite loops could occur.
1269    
1270    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1271        computing the minimum length on reaching *ACCEPT, and so ended up with the
1272        wrong value of 1 rather than 0. Further investigation indicates that
1273        computing a minimum subject length in the presence of *ACCEPT is difficult
1274        (think back references, subroutine calls), and so I have changed the code
1275        so that no minimum is registered for a pattern that contains *ACCEPT.
1276    
1277    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1278        it was not handled as intended. [But see 16 below.]
1279    
1280    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1281        Sheri Pierce.
1282    
1283    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1284        the first byte in a match must be "a".
1285    
1286    11. Change 17 for 8.13 increased the recursion depth for patterns like
1287        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1288        pattern contains any instances of (*THEN). If it does not, the old
1289        optimizations are restored. It would be nice to do this on a per-group
1290        basis, but at the moment that is not feasible.
1291    
1292    12. In some environments, the output of pcretest -C is CRLF terminated. This
1293        broke RunTest's code that checks for the link size. A single white space
1294        character after the value is now allowed for.
1295    
1296    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1297        For "fr", it uses the Windows-specific input and output files.
1298    
1299    14. If (*THEN) appeared in a group that was called recursively or as a
1300        subroutine, it did not work as intended. [But see next item.]
1301    
1302    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1303        pattern fragments (but not containing any | characters). If A and B are
1304        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1305        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1306        D. In other words, Perl considers parentheses that do not contain any |
1307        characters to be part of a surrounding alternative, whereas PCRE was
1308        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1309        differently. PCRE now behaves in the same way as Perl, except in the case
1310        of subroutine/recursion calls such as (?1) which have in any case always
1311        been different (but PCRE had them first :-).
1312    
1313    16. Related to 15 above: Perl does not treat the | in a conditional group as
1314        creating alternatives. Such a group is treated in the same way as an
1315        ordinary group without any | characters when processing (*THEN). PCRE has
1316        been changed to match Perl's behaviour.
1317    
1318    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1319        RunGrepTest script failed.
1320    
1321    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1322        inevitable for groups that contain captures, but it can lead to a lot of
1323        stack use in large patterns. The old behaviour has been restored for atomic
1324        groups that do not contain any capturing parentheses.
1325    
1326    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1327        suppress the check for a minimum subject length at run time. (If it was
1328        given to pcre_exec() or pcre_dfa_exec() it did work.)
1329    
1330    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1331        fail to work when decoding hex characters in data strings in EBCDIC
1332        environments.
1333    
1334    21. It appears that in at least one Mac OS environment, the isxdigit() function
1335        is implemented as a macro that evaluates to its argument more than once,
1336        contravening the C 90 Standard (I haven't checked a later standard). There
1337        was an instance in pcretest which caused it to go wrong when processing
1338        \x{...} escapes in subject strings. The has been rewritten to avoid using
1339        things like p++ in the argument of isxdigit().
1340    
1341    
1342    Version 8.13 16-Aug-2011
1343    ------------------------
1344    
1345    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1346    
1347    2.  Two minor typos in pcre_internal.h have been fixed.
1348    
1349    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1350        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1351        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1352    
1353    4.  There were a number of related bugs in the code for matching backrefences
1354        caselessly in UTF-8 mode when codes for the characters concerned were
1355        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1356        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1357        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1358        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1359        2-byte code at the end of the subject (it thought there wasn't enough data
1360        left).
1361    
1362    5.  Comprehensive information about what went wrong is now returned by
1363        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1364        as the output vector has at least 2 elements. The offset of the start of
1365        the failing character and a reason code are placed in the vector.
1366    
1367    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1368        now returned is for the first byte of the failing character, instead of the
1369        last byte inspected. This is an incompatible change, but I hope it is small
1370        enough not to be a problem. It makes the returned offset consistent with
1371        pcre_exec() and pcre_dfa_exec().
1372    
1373    7.  pcretest now gives a text phrase as well as the error number when
1374        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1375        failure, the offset and reason code are output.
1376    
1377    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1378        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1379        back over a single character (\n). This seems wrong (because it treated the
1380        two characters as a single entity when going forwards), conflicts with the
1381        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1382        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1383        has been changed.
1384    
1385    9.  Some internal refactoring has changed the processing so that the handling
1386        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1387        time (the PCRE_DOTALL option was changed this way some time ago: version
1388        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1389        which was always a bit of a fudge. It also means that there is one less
1390        argument for the match() function, which reduces its stack requirements
1391        slightly. This change also fixes an incompatibility with Perl: the pattern
1392        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1393    
1394    10. More internal refactoring has drastically reduced the number of recursive
1395        calls to match() for possessively repeated groups such as (abc)++ when
1396        using pcre_exec().
1397    
1398    11. While implementing 10, a number of bugs in the handling of groups were
1399        discovered and fixed:
1400    
1401        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1402        (a|)*(?1) gave a compile-time internal error.
1403        ((a|)+)+  did not notice that the outer group could match an empty string.
1404        (^a|^)+   was not marked as anchored.
1405        (.*a|.*)+ was not marked as matching at start or after a newline.
1406    
1407    12. Yet more internal refactoring has removed another argument from the match()
1408        function. Special calls to this function are now indicated by setting a
1409        value in a variable in the "match data" data block.
1410    
1411    13. Be more explicit in pcre_study() instead of relying on "default" for
1412        opcodes that mean there is no starting character; this means that when new
1413        ones are added and accidentally left out of pcre_study(), testing should
1414        pick them up.
1415    
1416    14. The -s option of pcretest has been documented for ages as being an old
1417        synonym of -m (show memory usage). I have changed it to mean "force study
1418        for every regex", that is, assume /S for every regex. This is similar to -i
1419        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1420        using it. It makes it easier to run collections of tests with and without
1421        study enabled, and thereby test pcre_study() more easily. All the standard
1422        tests are now run with and without -s (but some patterns can be marked as
1423        "never study" - see 20 below).
1424    
1425    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1426        restoration of the capturing data to the outer values was not happening
1427        correctly.
1428    
1429    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1430        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1431        pattern had matched an empty string, and so incorrectly returned a no
1432        match.
1433    
1434    17. There was optimizing code for the last branch of non-capturing parentheses,
1435        and also for the obeyed branch of a conditional subexpression, which used
1436        tail recursion to cut down on stack usage. Unfortunately, now that there is
1437        the possibility of (*THEN) occurring in these branches, tail recursion is
1438        no longer possible because the return has to be checked for (*THEN). These
1439        two optimizations have therefore been removed. [But see 8.20/11 above.]
1440    
1441    18. If a pattern containing \R was studied, it was assumed that \R always
1442        matched two bytes, thus causing the minimum subject length to be
1443        incorrectly computed because \R can also match just one byte.
1444    
1445    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1446        was incorrectly computed.
1447    
1448    20. If /S is present twice on a test pattern in pcretest input, it now
1449        *disables* studying, thereby overriding the use of -s on the command line
1450        (see 14 above). This is necessary for one or two tests to keep the output
1451        identical in both cases.
1452    
1453    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1454        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1455    
1456    22. When an atomic group that contained a capturing parenthesis was
1457        successfully matched, but the branch in which it appeared failed, the
1458        capturing was not being forgotten if a higher numbered group was later
1459        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1460        group 1 to "a", when in fact it should be unset. This applied to multi-
1461        branched capturing and non-capturing groups, repeated or not, and also to
1462        positive assertions (capturing in negative assertions does not happen
1463        in PCRE) and also to nested atomic groups.
1464    
1465    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1466        subject after a captured substring, to make it easier to tell which of a
1467        number of identical substrings has been captured.
1468    
1469    24. The way atomic groups are processed by pcre_exec() has been changed so that
1470        if they are repeated, backtracking one repetition now resets captured
1471        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1472        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1473        "aaa". Previously, it would have been "a". As part of this code
1474        refactoring, the way recursive calls are handled has also been changed.
1475    
1476    25. If an assertion condition captured any substrings, they were not passed
1477        back unless some other capturing happened later. For example, if
1478        (?(?=(a))a) was matched against "a", no capturing was returned.
1479    
1480    26. When studying a pattern that contained subroutine calls or assertions,
1481        the code for finding the minimum length of a possible match was handling
1482        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1483        group 1 called group 2 while simultaneously a separate group 2 called group
1484        1). A stack overflow occurred in this case. I have fixed this by limiting
1485        the recursion depth to 10.
1486    
1487    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1488        Fortmann. This supports explicit test numbers on the command line, and has
1489        argument validation and error reporting.
1490    
1491    28. An instance of \X with an unlimited repeat could fail if at any point the
1492        first character it looked at was a mark character.
1493    
1494    29. Some minor code refactoring concerning Unicode properties and scripts
1495        should reduce the stack requirement of match() slightly.
1496    
1497    30. Added the '=' option to pcretest to check the setting of unused capturing
1498        slots at the end of the pattern, which are documented as being -1, but are
1499        not included in the return count.
1500    
1501    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1502        compiled something random. Now it gives a compile-time error (as does
1503        Perl).
1504    
1505    32. A *MARK encountered during the processing of a positive assertion is now
1506        recorded and passed back (compatible with Perl).
1507    
1508    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1509        had alternative anchored branches, the search for a second match in a line
1510        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1511        matched the line "0102" twice. The same bug affected patterns that started
1512        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1513        twice.
1514    
1515    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1516        does, and because of capturing effects, quantifying parenthesized
1517        assertions may at times be useful. Quantifiers are now allowed for
1518        parenthesized assertions.
1519    
1520    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1521    
1522    36. \g was being checked for fancy things in a character class, when it should
1523        just be a literal "g".
1524    
1525    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1526        appearance of a nested POSIX class supersedes an apparent external class.
1527        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1528        unescaped square brackets may also appear as part of class names. For
1529        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1530        more like Perl. (But see 8.20/1 above.)
1531    
1532    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1533        was because it thought it was \N{name}, which is not supported).
1534    
1535    39. Add minix to OS list not supporting the -S option in pcretest.
1536    
1537    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1538        cannot analyze patterns in sufficient detail to catch mutual recursions
1539        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1540        subgroup is called recursively as a subpattern for a second time at the
1541        same position in the subject string. In previous releases this might have
1542        been caught by the recursion limit, or it might have run out of stack.
1543    
1544    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1545        happen only once. PCRE was, however incorrectly giving a compile time error
1546        "recursive call could loop indefinitely" because it cannot analyze the
1547        pattern in sufficient detail. The compile time test no longer happens when
1548        PCRE is compiling a conditional subpattern, but actual runaway loops are
1549        now caught at runtime (see 40 above).
1550    
1551    42. It seems that Perl allows any characters other than a closing parenthesis
1552        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1553        has been changed to be the same.
1554    
1555    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1556        as not to get warnings when autogen.sh is called. Also changed
1557        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1558    
1559    44. To help people who use pcregrep to scan files containing exceedingly long
1560        lines, the following changes have been made:
1561    
1562        (a) The default value of the buffer size parameter has been increased from
1563            8K to 20K. (The actual buffer used is three times this size.)
1564    
1565        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1566            PCRE is built.
1567    
1568        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1569            to be set at run time.
1570    
1571        (d) Numerical values in pcregrep options can be followed by K or M, for
1572            example --buffer-size=50K.
1573    
1574        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1575            given and the return code is set to 2.
1576    
1577    45. Add a pointer to the latest mark to the callout data block.
1578    
1579    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1580        partial match of an empty string instead of no match. This was specific to
1581        the use of ".".
1582    
1583    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1584        complete match instead of a partial match. This bug was dependent on both
1585        the PCRE_UTF8 and PCRE_DOTALL options being set.
1586    
1587    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1588        starting byte set, because \b was not being ignored.
1589    
1590    
1591    Version 8.12 15-Jan-2011
1592    ------------------------
1593    
1594    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1595        checks for such things as part of the documentation building process.
1596    
1597    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1598        --match-limit and --recursion-limit options (added for 8.11). In
1599        particular, this made one of the standard tests fail. (The integer value
1600        went into the wrong half of a long int.)
1601    
1602    3.  If the --colour option was given to pcregrep with -v (invert match), it
1603        did strange things, either producing crazy output, or crashing. It should,
1604        of course, ignore a request for colour when reporting lines that do not
1605        match.
1606    
1607    4.  Another pcregrep bug caused similar problems if --colour was specified with
1608        -M (multiline) and the pattern match finished with a line ending.
1609    
1610    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1611        matched in multiline mode, the following line was shown as part of the
1612        match. This seems wrong, so I have changed it.
1613    
1614    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1615        the check for further matches in the same line (so they could be coloured)
1616        to overrun the end of the current line. If another match was found, it was
1617        incorrectly shown (and then shown again when found in the next line).
1618    
1619    7.  If pcregrep was compiled under Windows, there was a reference to the
1620        function pcregrep_exit() before it was defined. I am assuming this was
1621        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1622        reported by a user. I've moved the definition above the reference.
1623    
1624    
1625    Version 8.11 10-Dec-2010
1626    ------------------------
1627    
1628    1.  (*THEN) was not working properly if there were untried alternatives prior
1629        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1630        backtracked to try for "b" instead of moving to the next alternative branch
1631        at the same level (in this case, to look for "c"). The Perl documentation
1632        is clear that when (*THEN) is backtracked onto, it goes to the "next
1633        alternative in the innermost enclosing group".
1634    
1635    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1636        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1637        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1638        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1639        (*THEN).
1640    
1641    3.  If \s appeared in a character class, it removed the VT character from
1642        the class, even if it had been included by some previous item, for example
1643        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1644        of \s, but is part of the POSIX "space" class.)
1645    
1646    4.  A partial match never returns an empty string (because you can always
1647        match an empty string at the end of the subject); however the checking for
1648        an empty string was starting at the "start of match" point. This has been
1649        changed to the "earliest inspected character" point, because the returned
1650        data for a partial match starts at this character. This means that, for
1651        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1652        (previously it gave "no match").
1653    
1654    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1655        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1656        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1657        has an implication that the given string is incomplete (because a partial
1658        match is preferred over a full match). For this reason, these items now
1659        give a partial match in this situation. [Aside: previously, the one case
1660        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1661        match rather than a full match, which was wrong by the old rules, but is
1662        now correct.]
1663    
1664    6.  There was a bug in the handling of #-introduced comments, recognized when
1665        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1666        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1667        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1668        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1669        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1670        places in pcre_compile().
1671    
1672    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1673        comments when looking ahead for named forward references to subpatterns,
1674        the only newline sequence it recognized was NL. It now handles newlines
1675        according to the set newline convention.
1676    
1677    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1678        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1679        cater for a lack of strerror(). These oversights have been fixed.
1680    
1681    9.  Added --match-limit and --recursion-limit to pcregrep.
1682    
1683    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1684    
1685    11. When the -o option was used, pcregrep was setting a return code of 1, even
1686        when matches were found, and --line-buffered was not being honoured.
1687    
1688    12. Added an optional parentheses number to the -o and --only-matching options
1689        of pcregrep.
1690    
1691    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1692        can match an empty string. The code to do it in pcretest and pcredemo
1693        needed fixing:
1694    
1695        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1696            only one byte after an empty string match just before CRLF (this case
1697            just got forgotten; "any" and "anycrlf" were OK).
1698    
1699        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1700            mode when an empty string match preceded an ASCII character followed by
1701            a non-ASCII character. (The code for advancing by one character rather
1702            than one byte was nonsense.)
1703    
1704        (c) The pcredemo.c sample program did not have any code at all to handle
1705            the cases when CRLF is a valid newline sequence.
1706    
1707    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1708        as a starting offset was within the subject string. There is now a new
1709        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1710        negative or greater than the length of the string. In order to test this,
1711        pcretest is extended to allow the setting of negative starting offsets.
1712    
1713    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1714        starting offset points to the beginning of a UTF-8 character was
1715        unnecessarily clumsy. I tidied it up.
1716    
1717    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1718        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1719    
1720    17. Nobody had reported that the --include_dir option, which was added in
1721        release 7.7 should have been called --include-dir (hyphen, not underscore)
1722        for compatibility with GNU grep. I have changed it to --include-dir, but
1723        left --include_dir as an undocumented synonym, and the same for
1724        --exclude-dir, though that is not available in GNU grep, at least as of
1725        release 2.5.4.
1726    
1727    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1728        characters from a string of bytes) have been redefined so as not to use
1729        loops, in order to improve performance in some environments. At the same
1730        time, I abstracted some of the common code into auxiliary macros to save
1731        repetition (this should not affect the compiled code).
1732    
1733    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1734        compile-time error is now given if \c is not followed by an ASCII
1735        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1736        different, and any byte value is allowed.)
1737    
1738    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1739        START_OPTIMIZE option, which is now allowed at compile time - but just
1740        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1741        to pcregrep and other applications that have no direct access to PCRE
1742        options. The new /Y option in pcretest sets this option when calling
1743        pcre_compile().
1744    
1745    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1746        back references. Groups containing recursive back references were forced to
1747        be atomic by that change, but in the case of named groups, the amount of
1748        memory required was incorrectly computed, leading to "Failed: internal
1749        error: code overflow". This has been fixed.
1750    
1751    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1752        pcretest.c, to avoid build problems in some Borland environments.
1753    
1754    
1755    Version 8.10 25-Jun-2010
1756    ------------------------
1757    
1758    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1759        THEN.
1760    
1761    2.  (*ACCEPT) was not working when inside an atomic group.
1762    
1763    3.  Inside a character class, \B is treated as a literal by default, but
1764        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1765        causes the error). The code is unchanged, but I tidied the documentation.
1766    
1767    4.  Inside a character class, PCRE always treated \R and \X as literals,
1768        whereas Perl faults them if its -w option is set. I have changed PCRE so
1769        that it faults them when PCRE_EXTRA is set.
1770    
1771    5.  Added support for \N, which always matches any character other than
1772        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1773    
1774    6.  When compiling pcregrep with newer versions of gcc which may have
1775        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1776        declared with attribute warn_unused_result" were given. Just casting the
1777        result to (void) does not stop the warnings; a more elaborate fudge is
1778        needed. I've used a macro to implement this.
1779    
1780    7.  Minor change to pcretest.c to avoid a compiler warning.
1781    
1782    8.  Added four artifical Unicode properties to help with an option to make
1783        \s etc use properties (see next item). The new properties are: Xan
1784        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1785    
1786    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1787        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1788        this option. Modified pcretest to add /W to test this facility. Added
1789        REG_UCP to make it available via the POSIX interface.
1790    
1791    10. Added --line-buffered to pcregrep.
1792    
1793    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1794        studied, and the match started with a letter with a code point greater than
1795        127 whose first byte was different to the first byte of the other case of
1796        the letter, the other case of this starting letter was not recognized
1797        (#976).
1798    
1799    12. If a pattern that was studied started with a repeated Unicode property
1800        test, for example, \p{Nd}+, there was the theoretical possibility of
1801        setting up an incorrect bitmap of starting bytes, but fortunately it could
1802        not have actually happened in practice until change 8 above was made (it
1803        added property types that matched character-matching opcodes).
1804    
1805    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1806        possible starting bytes for non-anchored patterns.
1807    
1808    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1809        \R, and also a number of cases that involve Unicode properties, both
1810        explicit and implicit when PCRE_UCP is set.
1811    
1812    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1813        input, it could crash or give wrong results if characters with values
1814        greater than 0xc0 were present in the subject string. (Detail: it assumed
1815        UTF-8 input when processing these items.)
1816    
1817    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1818        size_t is 64-bit (#991).
1819    
1820    17. Added a check for running out of memory when PCRE is compiled with
1821        --disable-stack-for-recursion (#990).
1822    
1823    18. If the last data line in a file for pcretest does not have a newline on
1824        the end, a newline was missing in the output.
1825    
1826    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1827        less than 128) in its various bitmaps. However, there is a facility for
1828        generating tables according to the current locale when PCRE is compiled. It
1829        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1830        characters, are recognized by isspace() and therefore were getting set in
1831        these tables, and indeed these tables seem to approximate to ISO 8859. This
1832        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1833        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1834        which of course cannot start UTF-8 characters. I have changed the code so
1835        that only real ASCII characters (less than 128) and the correct starting
1836        bytes for UTF-8 encodings are set for characters greater than 127 when in
1837        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1838        altogether.)
1839    
1840    20. Added the /T option to pcretest so as to be able to run tests with non-
1841        standard character tables, thus making it possible to include the tests
1842        used for 19 above in the standard set of tests.
1843    
1844    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1845        reference to a subpattern the other side of a comment that contains an
1846        opening parenthesis caused either an internal compiling error, or a
1847        reference to the wrong subpattern.
1848    
1849    
1850    Version 8.02 19-Mar-2010
1851    ------------------------
1852    
1853    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1854    
1855    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1856        configured.
1857    
1858    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1859        original author of that file, following a query about its status.
1860    
1861    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1862        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1863    
1864    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1865        quantifier applied to a forward-referencing subroutine call, could compile
1866        incorrect code or give the error "internal error: previously-checked
1867        referenced subpattern not found".
1868    
1869    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1870        variables to point to external functions. For these systems, therefore,
1871        pcre_malloc etc. are now initialized to local functions that call the
1872        relevant global functions.
1873    
1874    7.  There were two entries missing in the vectors called coptable and poptable
1875        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1876        I've fixed the data, and added a kludgy way of testing at compile time that
1877        the lengths are correct (equal to the number of opcodes).
1878    
1879    8.  Following on from 7, I added a similar kludge to check the length of the
1880        eint vector in pcreposix.c.
1881    
1882    9.  Error texts for pcre_compile() are held as one long string to avoid too
1883        much relocation at load time. To find a text, the string is searched,
1884        counting zeros. There was no check for running off the end of the string,
1885        which could happen if a new error number was added without updating the
1886        string.
1887    
1888    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1889    
1890    11. \K was not working if it appeared in an atomic group or in a group that
1891        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1892        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1893        the assertion is positive, but not if it is negative.
1894    
1895    12. Change 11 fortuitously reduced the size of the stack frame used in the
1896        "match()" function of pcre_exec.c by one pointer. Forthcoming
1897        implementation of support for (*MARK) will need an extra pointer on the
1898        stack; I have reserved it now, so that the stack frame size does not
1899        decrease.
1900    
1901    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1902        item in branch that calls a recursion is a subroutine call - as in the
1903        second branch in the above example - was incorrectly given the compile-
1904        time error "recursive call could loop indefinitely" because pcre_compile()
1905        was not correctly checking the subroutine for matching a non-empty string.
1906    
1907    14. The checks for overrunning compiling workspace could trigger after an
1908        overrun had occurred. This is a "should never occur" error, but it can be
1909        triggered by pathological patterns such as hundreds of nested parentheses.
1910        The checks now trigger 100 bytes before the end of the workspace.
1911    
1912    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1913    
1914    
1915    Version 8.01 19-Jan-2010
1916    ------------------------
1917    
1918    1.  If a pattern contained a conditional subpattern with only one branch (in
1919        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1920        computed the wrong minimum data length (which is of course zero for such
1921        subpatterns). This could cause incorrect "no match" results.
1922    
1923    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1924        the pattern is reset in the first branch, pcre_compile() failed with
1925        "internal error: code overflow at offset...". This happened only when
1926        the reset was to the original external option setting. (An optimization
1927        abstracts leading options settings into an external setting, which was the
1928        cause of this.)
1929    
1930    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1931        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1932        assertion pattern did not match (meaning that the assertion was true), it
1933        was incorrectly treated as false if the SKIP had been reached during the
1934        matching. This also applied to assertions used as conditions.
1935    
1936    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1937        assertion subpattern, including such a pattern used as a condition,
1938        unpredictable results occurred, instead of the error return
1939        PCRE_ERROR_DFA_UITEM.
1940    
1941    5.  The C++ GlobalReplace function was not working like Perl for the special
1942        situation when an empty string is matched. It now does the fancy magic
1943        stuff that is necessary.
1944    
1945    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1946        removed. (These were left over from very, very early versions of PCRE.)
1947    
1948    7.  Some cosmetic changes to the code to make life easier when compiling it
1949        as part of something else:
1950    
1951        (a) Change DEBUG to PCRE_DEBUG.
1952    
1953        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1954            called "current" as "current_branch", to prevent a collision with the
1955            Linux macro when compiled as a kernel module.
1956    
1957        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1958            prevent a collision with the Linux macro when compiled as a kernel
1959            module.
1960    
1961    8.  In pcre_compile() there are some checks for integer overflows that used to
1962        cast potentially large values to (double). This has been changed to that
1963        when building, a check for int64_t is made, and if it is found, it is used
1964        instead, thus avoiding the use of floating point arithmetic. (There is no
1965        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1966        double.
1967    
1968    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1969        2005 (difference between two addresses compared to an unsigned value).
1970    
1971    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1972        custom one, because of the following reported problem in Windows:
1973    
1974          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1975              under Win32.
1976          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1977              therefore missing the function definition.
1978          - The compiler thus generates a "C" signature for the test function.
1979          - The linker fails to find the "C" function.
1980          - PCRE fails to configure if asked to do so against libbz2.
1981    
1982    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1983        messages were output:
1984    
1985          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1986          rerunning libtoolize, to keep the correct libtool macros in-tree.
1987          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1988    
1989        I have done both of these things.
1990    
1991    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1992        most of the time, it *can* run out if it is given a pattern that contains a
1993        runaway infinite recursion. I updated the discussion in the pcrestack man
1994        page.
1995    
1996    13. Now that we have gone to the x.xx style of version numbers, the minor
1997        version may start with zero. Using 08 or 09 is a bad idea because users
1998        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1999        interpreted as invalid octal numbers. I've updated the previous comment in
2000        configure.ac, and also added a check that gives an error if 08 or 09 are
2001        used.
2002    
2003    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2004        causing partial matching to fail when the end of the subject matched \W
2005        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2006    
2007    15. There were some discrepancies between the declarations in pcre_internal.h
2008        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2009        their definitions. The declarations used "const uschar *" and the
2010        definitions used USPTR. Even though USPTR is normally defined as "const
2011        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2012        reported that: "This difference in casting confuses some C++ compilers, for
2013        example, SunCC recognizes above declarations as different functions and
2014        generates broken code for hbpcre." I have changed the declarations to use
2015        USPTR.
2016    
2017    16. GNU libtool is named differently on some systems. The autogen.sh script now
2018        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2019        (FreeBSD).
2020    
2021    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2022        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2023        comment: "Figure out how to create a longlong from a string: strtoll and
2024        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2025        instance, but it only takes 2 args instead of 3!"
2026    
2027    18. A subtle bug concerned with back references has been fixed by a change of
2028        specification, with a corresponding code fix. A pattern such as
2029        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2030        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2031        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2032        same bug. Such groups have to be quantified to be useful, or contained
2033        inside another quantified group. (If there's no repetition, the reference
2034        can never match.) The problem arises because, having left the group and
2035        moved on to the rest of the pattern, a later failure that backtracks into
2036        the group uses the captured value from the final iteration of the group
2037        rather than the correct earlier one. I have fixed this in PCRE by forcing
2038        any group that contains a reference to itself to be an atomic group; that
2039        is, there cannot be any backtracking into it once it has completed. This is
2040        similar to recursive and subroutine calls.
2041    
2042    
2043    Version 8.00 19-Oct-09
2044    ----------------------
2045    
2046    1.  The table for translating pcre_compile() error codes into POSIX error codes
2047        was out-of-date, and there was no check on the pcre_compile() error code
2048        being within the table. This could lead to an OK return being given in
2049        error.
2050    
2051    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2052        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2053        in a Windows environment.
2054    
2055    3.  The pcregrep --count option prints the count for each file even when it is
2056        zero, as does GNU grep. However, pcregrep was also printing all files when
2057        --files-with-matches was added. Now, when both options are given, it prints
2058        counts only for those files that have at least one match. (GNU grep just
2059        prints the file name in this circumstance, but including the count seems
2060        more useful - otherwise, why use --count?) Also ensured that the
2061        combination -clh just lists non-zero counts, with no names.
2062    
2063    4.  The long form of the pcregrep -F option was incorrectly implemented as
2064        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2065        but it seems right to fix it, and I didn't think it was worth preserving
2066        the old behaviour.
2067    
2068    5.  The command line items --regex=pattern and --regexp=pattern were not
2069        recognized by pcregrep, which required --regex pattern or --regexp pattern
2070        (with a space rather than an '='). The man page documented the '=' forms,
2071        which are compatible with GNU grep; these now work.
2072    
2073    6.  No libpcreposix.pc file was created for pkg-config; there was just
2074        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2075    
2076    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2077        when UCP support is not needed, by modifying the Python script that
2078        generates it from Unicode data files. This should not matter if the module
2079        is correctly used as a library, but I received one complaint about 50K of
2080        unwanted data. My guess is that the person linked everything into his
2081        program rather than using a library. Anyway, it does no harm.
2082    
2083    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2084        was a minimum greater than 1 for a wide character in a possessive
2085        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2086        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2087        character. Chaos in the form of incorrect output or a compiling loop could
2088        result.
2089    
2090    9.  The restrictions on what a pattern can contain when partial matching is
2091        requested for pcre_exec() have been removed. All patterns can now be
2092        partially matched by this function. In addition, if there are at least two
2093        slots in the offset vector, the offset of the earliest inspected character
2094        for the match and the offset of the end of the subject are set in them when
2095        PCRE_ERROR_PARTIAL is returned.
2096    
2097    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2098        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2099        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2100        and may be more useful for multi-segment matching.
2101    
2102    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2103        used to be given if ever the end of the subject was reached; now it is
2104        given only if matching could not proceed because another character was
2105        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2106        string "Z", which now yields "no match" instead of "partial match". In the
2107        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2108        final character ended with (*FAIL).
2109    
2110    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2111        if the pattern had a "must contain" character that was already found in the
2112        earlier partial match, unless partial matching was again requested. For
2113        example, with the pattern /dog.(body)?/, the "must contain" character is
2114        "g". If the first part-match was for the string "dog", restarting with
2115        "sbody" failed. This bug has been fixed.
2116    
2117    13. The string returned by pcre_dfa_exec() after a partial match has been
2118        changed so that it starts at the first inspected character rather than the
2119        first character of the match. This makes a difference only if the pattern
2120        starts with a lookbehind assertion or \b or \B (\K is not supported by
2121        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2122        matching functions compatible, and I think it's the right thing to do.
2123    
2124    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2125        so that the demonstration program is easily available in environments where
2126        PCRE has not been installed from source.
2127    
2128    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2129        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2130        library.
2131    
2132    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2133        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2134        is not the first non-POSIX option to be added. Clearly some people find
2135        these options useful.
2136    
2137    17. If a caller to the POSIX matching function regexec() passes a non-zero
2138        value for nmatch with a NULL value for pmatch, the value of
2139        nmatch is forced to zero.
2140    
2141    18. RunGrepTest did not have a test for the availability of the -u option of
2142        the diff command, as RunTest does. It now checks in the same way as
2143        RunTest, and also checks for the -b option.
2144    
2145    19. If an odd number of negated classes containing just a single character
2146        interposed, within parentheses, between a forward reference to a named
2147        subpattern and the definition of the subpattern, compilation crashed with
2148        an internal error, complaining that it could not find the referenced
2149        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2150        [The bug was that it was starting one character too far in when skipping
2151        over the character class, thus treating the ] as data rather than
2152        terminating the class. This meant it could skip too much.]
2153    
2154    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2155        /g option in pcretest when the pattern contains \K, which makes it possible
2156        to have an empty string match not at the start, even when the pattern is
2157        anchored. Updated pcretest and pcredemo to use this option.
2158    
2159    21. If the maximum number of capturing subpatterns in a recursion was greater
2160        than the maximum at the outer level, the higher number was returned, but
2161        with unset values at the outer level. The correct (outer level) value is
2162        now given.
2163    
2164    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2165        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2166        make it do so. The string so far is captured, making this feature
2167        compatible with Perl.
2168    
2169    23. The tests have been re-organized, adding tests 11 and 12, to make it
2170        possible to check the Perl 5.10 features against Perl 5.10.
2171    
2172    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2173        pattern matches a fixed length string. PCRE did not allow this; now it
2174        does. Neither allows recursion.
2175    
2176    25. I finally figured out how to implement a request to provide the minimum
2177        length of subject string that was needed in order to match a given pattern.
2178        (It was back references and recursion that I had previously got hung up
2179        on.) This code has now been added to pcre_study(); it finds a lower bound
2180        to the length of subject needed. It is not necessarily the greatest lower
2181        bound, but using it to avoid searching strings that are too short does give
2182        some useful speed-ups. The value is available to calling programs via
2183        pcre_fullinfo().
2184    
2185    26. While implementing 25, I discovered to my embarrassment that pcretest had
2186        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2187        study optimizations had never been tested with that matching function.
2188        Oops. What is worse, even when it was passed study data, there was a bug in
2189        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2190        were also very few tests of studied patterns with pcre_dfa_exec().
2191    
2192    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2193        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2194        on the other side of the coin, they are no longer allowed to have different
2195        names, because these cannot be distinguished in PCRE, and this has caused
2196        confusion. (This is a difference from Perl.)
2197    
2198    28. When duplicate subpattern names are present (necessarily with different
2199        numbers, as required by 27 above), and a test is made by name in a
2200        conditional pattern, either for a subpattern having been matched, or for
2201        recursion in such a pattern, all the associated numbered subpatterns are
2202        tested, and the overall condition is true if the condition is true for any
2203        one of them. This is the way Perl works, and is also more like the way
2204        testing by number works.
2205    
2206    
2207    Version 7.9 11-Apr-09
2208    ---------------------
2209    
2210    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2211        (pcretest), all targets were linked against these libraries. This included
2212        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2213        libraries. This caused unwanted dependencies to be created. This problem
2214        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2215        pcretest is linked with readline.
2216    
2217    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2218        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2219        moved outside it again, because FALSE and TRUE are already defined in AIX,
2220        but BOOL is not.
2221    
2222    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2223        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2224    
2225    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2226        hyphens) following file names and line numbers when outputting matching
2227        lines. This is not true; no spaces are inserted. I have also clarified the
2228        wording for the --colour (or --color) option.
2229    
2230    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2231        was not coloured; this is different to GNU grep, so I have changed it to be
2232        the same.
2233    
2234    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2235        each matching line was coloured. Now it goes on to look for further matches
2236        of any of the test patterns, which is the same behaviour as GNU grep.
2237    
2238    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2239        doesn't make sense to accept an empty string match in pcregrep, so I have
2240        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2241        seems to be how GNU grep behaves. [But see later change 40 for release
2242        8.33.]
2243    
2244    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2245        start or after a newline", because the conditional assertion was not being
2246        correctly handled. The rule now is that both the assertion and what follows
2247        in the first alternative must satisfy the test.
2248    
2249    9.  If auto-callout was enabled in a pattern with a conditional group whose
2250        condition was an assertion, PCRE could crash during matching, both with
2251        pcre_exec() and pcre_dfa_exec().
2252    
2253    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2254        used for matching.
2255    
2256    11. Unicode property support in character classes was not working for
2257        characters (bytes) greater than 127 when not in UTF-8 mode.
2258    
2259    12. Added the -M command line option to pcretest.
2260    
2261    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2262    
2263    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2264    
2265    16. Added comments and documentation about mis-use of no_arg in the C++
2266        wrapper.
2267    
2268    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2269        from Martin Jerabek that uses macro names for all relevant character and
2270        string constants.
2271    
2272    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2273        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2274        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2275        these, but not everybody uses configure.
2276    
2277    19. A conditional group that had only one branch was not being correctly
2278        recognized as an item that could match an empty string. This meant that an
2279        enclosing group might also not be so recognized, causing infinite looping
2280        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2281        with the subject "ab", where knowledge that the repeated group can match
2282        nothing is needed in order to break the loop.
2283    
2284    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2285        exec(), but without supplying a callout function, matching went wrong.
2286    
2287    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2288        leak if the size of the offset vector was greater than 30. When the vector
2289        is smaller, the saved offsets during recursion go onto a local stack
2290        vector, but for larger vectors malloc() is used. It was failing to free
2291        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2292        error, in fact).
2293    
2294    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2295        heapframe that is used only when UTF-8 support is enabled. This caused no
2296        problem, but was untidy.
2297    
2298    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2299        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2300        included within another project.
2301    
2302    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2303        slightly modified by me:
2304    
2305          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2306              not building pcregrep.
2307    
2308          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2309              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2310    
2311    25. Forward references, both numeric and by name, in patterns that made use of
2312        duplicate group numbers, could behave incorrectly or give incorrect errors,
2313        because when scanning forward to find the reference group, PCRE was not
2314        taking into account the duplicate group numbers. A pattern such as
2315        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2316    
2317    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2318        the feature of a custom pointer more persuasive (as requested by a user).
2319    
2320    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2321        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2322        pre-defined.
2323    
2324    28. Added support for (*UTF8) at the start of a pattern.
2325    
2326    29. Arrange for flags added by the "release type" setting in CMake to be shown
2327        in the configuration summary.
2328    
2329    
2330    Version 7.8 05-Sep-08
2331    ---------------------
2332    
2333    1.  Replaced UCP searching code with optimized version as implemented for Ad
2334        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2335        stage table and inline lookup instead of a function, giving speed ups of 2
2336        to 5 times on some simple patterns that I tested. Permission was given to
2337        distribute the MultiStage2.py script that generates the tables (it's not in
2338        the tarball, but is in the Subversion repository).
2339    
2340    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2341        scripts.
2342    
2343    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2344        a group with a zero qualifier. The result of the study could be incorrect,
2345        or the function might crash, depending on the pattern.
2346    
2347    4.  Caseless matching was not working for non-ASCII characters in back
2348        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2349        It now works when Unicode Property Support is available.
2350    
2351    5.  In pcretest, an escape such as \x{de} in the data was always generating
2352        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2353        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2354        truncation.
2355    
2356    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2357    
2358    7.  Added two (int) casts to pcregrep when printing the difference of two
2359        pointers, in case they are 64-bit values.
2360    
2361    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2362        test 2 if it fails.
2363    
2364    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2365        and a #define of that name to empty if it is not externally set. This is to
2366        allow users of MSVC to set it if necessary.
2367    
2368    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2369        the convenience functions in the pcre_get.c source file.
2370    
2371    11. An option change at the start of a pattern that had top-level alternatives
2372        could cause overwriting and/or a crash. This command provoked a crash in
2373        some environments:
2374    
2375          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2376    
2377        This potential security problem was recorded as CVE-2008-2371.
2378    
2379    12. For a pattern where the match had to start at the beginning or immediately
2380        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2381        pcre_dfa_exec() could read past the end of the passed subject if there was
2382        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2383        pcretest so that it places the subject at the end of its malloc-ed buffer.
2384    
2385    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2386        exec() might read past the end of the data buffer in UTF-8 mode.
2387    
2388    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2389        the data contained the byte 0x85 as part of a UTF-8 character within its
2390        first line. This applied both to normal and DFA matching.
2391    
2392    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2393        /^[^d]*?$/8 failed to match "abc".
2394    
2395    16. Added a missing copyright notice to pcrecpp_internal.h.
2396    
2397    17. Make it more clear in the documentation that values returned from
2398        pcre_exec() in ovector are byte offsets, not character counts.
2399    
2400    18. Tidied a few places to stop certain compilers from issuing warnings.
2401    
2402    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2403        supplied by Stefan Weber. I made a further small update for 7.8 because
2404        there is a change of source arrangements: the pcre_searchfuncs.c module is
2405        replaced by pcre_ucd.c.
2406    
2407    
2408    Version 7.7 07-May-08
2409    ---------------------
2410    
2411    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2412        a string to a long long, pretend we don't even have a long long." This is
2413        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2414    
2415    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2416        pre-7.6 versions, which defined a global no_arg variable instead of putting
2417        it in the RE class. (See also #8 below.)
2418    
2419    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2420        Lopes.
2421    
2422    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2423    
2424        (1) The include/exclude patterns were being applied to the whole pathnames
2425            of files, instead of just to the final components.
2426    
2427        (2) If there was more than one level of directory, the subdirectories were
2428            skipped unless they satisfied the include/exclude conditions. This is
2429            inconsistent with GNU grep (and could even be seen as contrary to the
2430            pcregrep specification - which I improved to make it absolutely clear).
2431            The action now is always to scan all levels of directory, and just
2432            apply the include/exclude patterns to regular files.
2433    
2434    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2435        --exclude_dir in the tests to avoid scanning .svn directories.
2436    
2437    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2438        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2439        doesn't support NULs in patterns.
2440    
2441    7.  Added some missing "const"s to declarations of static tables in
2442        pcre_compile.c and pcre_dfa_exec.c.
2443    
2444    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2445        caused by fix #2  above. (Subsequently also a second patch to fix the
2446        first patch. And a third patch - this was a messy problem.)
2447    
2448    9.  Applied Craig's patch to remove the use of push_back().
2449    
2450    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2451        matching function regexec().
2452    
2453    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2454        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2455        references. PCRE supports relative numbers with this syntax (I don't think
2456        Oniguruma does).
2457    
2458    12. Previously, a group with a zero repeat such as (...){0} was completely
2459        omitted from the compiled regex. However, this means that if the group
2460        was called as a subroutine from elsewhere in the pattern, things went wrong
2461        (an internal error was given). Such groups are now left in the compiled
2462        pattern, with a new opcode that causes them to be skipped at execution
2463        time.
2464    
2465    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2466        to the way PCRE behaves:
2467    
2468        (a) A lone ] character is dis-allowed (Perl treats it as data).
2469    
2470        (b) A back reference to an unmatched subpattern matches an empty string
2471            (Perl fails the current match path).
2472    
2473        (c) A data ] in a character class must be notated as \] because if the
2474            first data character in a class is ], it defines an empty class. (In
2475            Perl it is not possible to have an empty class.) The empty class []
2476            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2477            The negative empty class [^] matches any one character, independently
2478            of the DOTALL setting.
2479    
2480    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2481        non-existent subpattern following a character class starting with ']' and
2482        containing () gave an internal compiling error instead of "reference to
2483        non-existent subpattern". Fortunately, when the pattern did exist, the
2484        compiled code was correct. (When scanning forwards to check for the
2485        existence of the subpattern, it was treating the data ']' as terminating
2486        the class, so got the count wrong. When actually compiling, the reference
2487        was subsequently set up correctly.)
2488    
2489    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2490        it was being rejected as not supported by pcre_dfa_exec(), even though
2491        other assertions are supported. I have made pcre_dfa_exec() support
2492        (*FAIL).
2493    
2494    16. The implementation of 13c above involved the invention of a new opcode,
2495        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2496        cannot be changed at match time, I realized I could make a small
2497        improvement to matching performance by compiling OP_ALLANY instead of
2498        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2499        on the OP_ANY path.
2500    
2501    17. Compiling pcretest on Windows with readline support failed without the
2502        following two fixes: (1) Make the unistd.h include conditional on
2503        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2504    
2505    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2506        ncurses library to be included for pcretest when ReadLine support is
2507        requested, but also to allow for it to be overridden. This patch came from
2508        Daniel Bergström.
2509    
2510    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2511        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2512        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2513        spotting this.
2514    
2515    
2516    Version 7.6 28-Jan-08
2517    ---------------------
2518    
2519    1.  A character class containing a very large number of characters with
2520        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2521        overflow.
2522    
2523    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2524        HAVE_LONG_LONG is not defined.
2525    
2526    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2527        bring them up to date and include new features. This patch includes:
2528    
2529        - Fixed PH's badly added libz and libbz2 support.
2530        - Fixed a problem with static linking.
2531        - Added pcredemo. [But later removed - see 7 below.]
2532        - Fixed dftables problem and added an option.
2533        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2534            HAVE_LONG_LONG.
2535        - Added readline support for pcretest.
2536        - Added an listing of the option settings after cmake has run.
2537    
2538    4.  A user submitted a patch to Makefile that makes it easy to create
2539        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2540        Makefile.am that cause it to include this special target, without
2541        affecting anything else. Note that the same mingw target plus all
2542        the other distribution libraries and programs are now supported
2543        when configuring with CMake (see 6 below) instead of with
2544        Configure/Make.
2545    
2546    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2547        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2548        exported in the Windows port". It has not yet been confirmed that the patch
2549        solves the problem, but it does no harm.
2550    
2551    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2552        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2553        with CMake, and also correct the comment about stack recursion.
2554    
2555    7.  Remove the automatic building of pcredemo from the ./configure system and
2556        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2557        of a program that users should build themselves after PCRE is installed, so
2558        building it automatically is not really right. What is more, it gave
2559        trouble in some build environments.
2560    
2561    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2562    
2563    
2564    Version 7.5 10-Jan-08
2565    ---------------------
2566    
2567    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2568        values in parens when parsing an RE using the C++ wrapper."
2569    
2570    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2571        Characters greater than 255 were excluded from the class instead of being
2572        included.
2573    
2574    3.  The same bug as (2) above applied to negated POSIX classes such as
2575        [:^space:].
2576    
2577    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2578        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2579        I have changed it.
2580    
2581    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2582        first named subpattern) and a construct such as (?&a) would reference the
2583        first named subpattern whose name started with "a" (in other words, the
2584        length check was missing). Both these problems are fixed. "Subpattern name
2585        expected" is now given for (?&) (a zero-length name), and this patch also
2586        makes it give the same error for \k'' (previously it complained that that
2587        was a reference to a non-existent subpattern).
2588    
2589    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2590        this is right because (?- can be followed by option settings as well as by
2591        digits. I have, however, made the messages clearer.
2592    
2593    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2594        than the number used in the conditional) now cause a compile-time error.
2595        This is actually not compatible with Perl, which accepts such patterns, but
2596        treats the conditional as always being FALSE (as PCRE used to), but it
2597        seems to me that giving a diagnostic is better.
2598    
2599    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2600        and messages.
2601    
2602    9.  Fix two occurrences of "backslash" in comments that should have been
2603        "backspace".
2604    
2605    10. Remove two redundant lines of code that can never be obeyed (their function
2606        was moved elsewhere).
2607    
2608    11. The program that makes PCRE's Unicode character property table had a bug
2609        which caused it to generate incorrect table entries for sequences of
2610        characters that have the same character type, but are in different scripts.
2611        It amalgamated them into a single range, with the script of the first of
2612        them. In other words, some characters were in the wrong script. There were
2613        thirteen such cases, affecting characters in the following ranges:
2614    
2615          U+002b0 - U+002c1
2616          U+0060c - U+0060d
2617          U+0061e - U+00612
2618          U+0064b - U+0065e
2619          U+0074d - U+0076d
2620          U+01800 - U+01805
2621          U+01d00 - U+01d77
2622          U+01d9b - U+01dbf
2623          U+0200b - U+0200f
2624          U+030fc - U+030fe
2625          U+03260 - U+0327f
2626          U+0fb46 - U+0fbb1
2627          U+10450 - U+1049d
2628    
2629    12. The -o option (show only the matching part of a line) for pcregrep was not
2630        compatible with GNU grep in that, if there was more than one match in a
2631        line, it showed only the first of them. It now behaves in the same way as
2632        GNU grep.
2633    
2634    13. If the -o and -v options were combined for pcregrep, it printed a blank
2635        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2636        does the same. The return code can be used to tell if there were any
2637        non-matching lines.
2638    
2639    14. Added --file-offsets and --line-offsets to pcregrep.
2640    
2641    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2642        infinitely looping recursion. The bug was that positive lookaheads were not
2643        being skipped when checking for a possible empty match (negative lookaheads
2644        and both kinds of lookbehind were skipped).
2645    
2646    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2647        inclusion of <windows.h> to before rather than after the definition of
2648        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2649    
2650    17. Specifying a possessive quantifier with a specific limit for a Unicode
2651        character property caused pcre_compile() to compile bad code, which led at
2652        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2653        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2654        caused the error; without that there was no problem.
2655    
2656    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2657    
2658    19. Added --enable-pcretest-libreadline.
2659    
2660    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2661        RE::GlobalReplace(). As a result, the number of replacements returned was
2662        double what it should be. I removed one of the increments, but Craig sent a
2663        later patch that removed the other one (the right fix) and added unit tests
2664        that check the return values (which was not done before).
2665    
2666    21. Several CMake things:
2667    
2668        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2669            the names libpcre and libpcreposix, not just pcre and pcreposix.
2670    
2671        (2) The above change means that pcretest and pcregrep are now correctly
2672            linked with the newly-built libraries, not previously installed ones.
2673    
2674        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2675    
2676    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2677        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2678        UTF-8 newline character). The key issue is that the pattern starts .*;
2679        this means that the match must be either at the beginning, or after a
2680        newline. The bug was in the code for advancing after a failed match and
2681        checking that the new position followed a newline. It was not taking
2682        account of UTF-8 characters correctly.
2683    
2684    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2685        character classes. PCRE was not treating the sequence [:...:] as a
2686        character class unless the ... were all letters. Perl, however, seems to
2687        allow any characters between [: and :], though of course it rejects as
2688        unknown any "names" that contain non-letters, because all the known class
2689        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2690        for example, whereas PCRE did not - it did not recognize a POSIX character
2691        class. This seemed a bit dangerous, so the code has been changed to be
2692        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2693        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2694        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2695        Perl does, and where it didn't before.
2696    
2697    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2698        Windows environments %n is disabled by default.
2699    
2700    
2701    Version 7.4 21-Sep-07
2702    ---------------------
2703    
2704    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2705        means that a class such as [\s] counted as "explicit reference to CR or
2706        LF". That isn't really right - the whole point of the change was to try to
2707        help when there was an actual mention of one of the two characters. So now
2708        the change happens only if \r or \n (or a literal CR or LF) character is
2709        encountered.
2710    
2711    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2712        of both had grown to the point where there were only 3 bits left.
2713        Fortunately, there was spare space in the data structure, and so I have
2714        moved the internal flags into a new 16-bit field to free up more option
2715        bits.
2716    
2717    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2718        but did not set the internal JCHANGED flag - either of these is enough to
2719        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2720        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2721        start sets both bits.
2722    
2723    4.  Added options (at build time, compile time, exec time) to change \R from
2724        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2725    
2726    5.  doc/pcresyntax.html was missing from the distribution.
2727    
2728    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2729        compatibility, even though it is no longer used.
2730    
2731    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2732        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2733        windows.h file is present (where different names are used). [This was
2734        reversed later after testing - see 16 below.]
2735    
2736    8.  Changed all #include <config.h> to #include "config.h". There were also
2737        some further <pcre.h> cases that I changed to "pcre.h".
2738    
2739    9.  When pcregrep was used with the --colour option, it missed the line ending
2740        sequence off the lines that it output.
2741    
2742    10. It was pointed out to me that arrays of string pointers cause lots of
2743        relocations when a shared library is dynamically loaded. A technique of
2744        using a single long string with a table of offsets can drastically reduce
2745        these. I have refactored PCRE in four places to do this. The result is
2746        dramatic:
2747    
2748          Originally:                          290
2749          After changing UCP table:            187
2750          After changing error message table:   43
2751          After changing table of "verbs"       36
2752          After changing table of Posix names   22
2753    
2754        Thanks to the folks working on Gregex for glib for this insight.
2755    
2756    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2757        unicode-properties was also set.
2758    
2759    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2760    
2761    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2762        checked only for CRLF.
2763    
2764    14. Added casts to pcretest.c to avoid compiler warnings.
2765    
2766    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2767    
2768    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2769        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2770        entirely. This removes changes made in 7 above.
2771    
2772    17. The CMake files have been updated, and there is now more information about
2773        building with CMake in the NON-UNIX-USE document.
2774    
2775    
2776    Version 7.3 28-Aug-07
2777  ---------------------  ---------------------
2778    
2779   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 29  Version 7.3 05-Jul-07 Line 2801  Version 7.3 05-Jul-07
2801   4. Debugging: adjusted the names of opcodes for different kinds of parentheses   4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2802      in debug output.      in debug output.
2803    
2804   5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp   5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2805      unittest when running under MinGW.      long printing in the pcrecpp unittest when running under MinGW.
2806    
2807   6. ESC_K was left out of the EBCDIC table.   6. ESC_K was left out of the EBCDIC table.
2808    
# Line 53  Version 7.3 05-Jul-07 Line 2825  Version 7.3 05-Jul-07
2825      pcreposix and pcrecpp parts of the library. These were overlooked when this      pcreposix and pcrecpp parts of the library. These were overlooked when this
2826      problem was solved for the main library.      problem was solved for the main library.
2827    
2828     9. There were some crude static tests to avoid integer overflow when computing
2829        the size of patterns that contain repeated groups with explicit upper
2830        limits. As the maximum quantifier is 65535, the maximum group length was
2831        set at 30,000 so that the product of these two numbers did not overflow a
2832        32-bit integer. However, it turns out that people want to use groups that
2833        are longer than 30,000 bytes (though not repeat them that many times).
2834        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2835        made it possible to implement the integer overflow checks in a much more
2836        dynamic way, which I have now done. The artificial limitation on group
2837        length has been removed - we now have only the limit on the total length of
2838        the compiled pattern, which depends on the LINK_SIZE setting.
2839    
2840    10. Fixed a bug in the documentation for get/copy named substring when
2841        duplicate names are permitted. If none of the named substrings are set, the
2842        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2843        empty string.
2844    
2845    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2846        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2847        because the ] is interpreted as the first data character and the
2848        terminating ] is not found. PCRE has been made compatible with Perl in this
2849        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2850        cause memory overwriting.
2851    
2852    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2853        string has been matched (to stop an infinite loop). It was not recognizing
2854        a conditional subpattern that could match an empty string if that
2855        subpattern was within another subpattern. For example, it looped when
2856        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2857        condition was not nested. This bug has been fixed.
2858    
2859    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2860        past the start of the subject in the presence of bytes with the top bit
2861        set, for example "\x8aBCD".
2862    
2863    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2864        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2865    
2866    14. Optimized (?!) to (*FAIL).
2867    
2868    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2869        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2870        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2871        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2872        does: it's just the validity check that is more restrictive.
2873    
2874    16. Inserted checks for integer overflows during escape sequence (backslash)
2875        processing, and also fixed erroneous offset values for syntax errors during
2876        backslash processing.
2877    
2878    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2879        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2880    
2881    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2882        caused an overrun.
2883    
2884    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2885        something other than just ASCII characters) inside a group that had an
2886        unlimited repeat caused a loop at compile time (while checking to see
2887        whether the group could match an empty string).
2888    
2889    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2890        [\P{Any}] did so. (Error in the code for printing property names.)
2891    
2892    21. An orphan \E inside a character class could cause a crash.
2893    
2894    22. A repeated capturing bracket such as (A)? could cause a wild memory
2895        reference during compilation.
2896    
2897    23. There are several functions in pcre_compile() that scan along a compiled
2898        expression for various reasons (e.g. to see if it's fixed length for look
2899        behind). There were bugs in these functions when a repeated \p or \P was
2900        present in the pattern. These operators have additional parameters compared
2901        with \d, etc, and these were not being taken into account when moving along
2902        the compiled data. Specifically:
2903    
2904        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2905            length.
2906    
2907        (b) An item such as \pL+ within a repeated group could cause crashes or
2908            loops.
2909    
2910        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2911            "reference to non-existent subpattern" error.
2912    
2913        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2914    
2915    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2916        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2917    
2918    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2919    
2920    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2921        character were causing crashes (broken optimization).
2922    
2923    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2924        \p or \P) caused a compile-time loop.
2925    
2926    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2927        break. For example, the unstudied pattern [\r\n]A does not match the string
2928        "\r\nA" because change 7.0/46 below moves the current point on by two
2929        characters after failing to match at the start. However, the pattern \nA
2930        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2931        the same is true. There doesn't seem any very clean way out of this, but
2932        what I have chosen to do makes the common cases work: PCRE now takes note
2933        of whether there can be an explicit match for \r or \n anywhere in the
2934        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2935        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2936        pattern has explicit CR or LF references.
2937    
2938    29. Added (*CR) etc for changing newline setting at start of pattern.
2939    
2940    
2941  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
2942  ---------------------  ---------------------

Legend:
Removed from v.200  
changed lines
  Added in v.1539

  ViewVC Help
Powered by ViewVC 1.1.5