/[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 1563 by ph10, Mon Jun 8 17:55:54 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.38 xx-xxx-xxxx
8    ------------------------
9    
10    1.  If a group that contained a recursive back reference also contained a
11        forward reference subroutine call followed by a non-forward-reference
12        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13        compile correct code, leading to undefined behaviour or an internally
14        detected error. This bug was discovered by the LLVM fuzzer.
15    
16    2.  Quantification of certain items (e.g. atomic back references) could cause
17        incorrect code to be compiled when recursive forward references were
18        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19        This bug was discovered by the LLVM fuzzer.
20    
21    3.  A repeated conditional group whose condition was a reference by name caused
22        a buffer overflow if there was more than one group with the given name.
23        This bug was discovered by the LLVM fuzzer.
24    
25    4.  A recursive back reference by name within a group that had the same name as
26        another group caused a buffer overflow. For example:
27        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28    
29    5.  A forward reference by name to a group whose number is the same as the
30        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31        a buffer overflow at compile time. This bug was discovered by the LLVM
32        fuzzer.
33    
34    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
35        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36    
37    7.  Another buffer overflow bug involved duplicate named groups with a
38        reference between their definition, with a group that reset capture
39        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40        fixed by always allowing for more memory, even if not needed. (A proper fix
41        is implemented in PCRE2, but it involves more refactoring.)
42    
43    8.  There was no check for integer overflow in subroutine calls such as (?123).
44    
45    
46    Version 8.37 28-April-2015
47    --------------------------
48    
49    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
50        for those parentheses to be closed with whatever has been captured so far.
51        However, it was failing to mark any other groups between the hightest
52        capture so far and the currrent group as "unset". Thus, the ovector for
53        those groups contained whatever was previously there. An example is the
54        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
55    
56    2.  If an assertion condition was quantified with a minimum of zero (an odd
57        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
58    
59    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
60        unrecognized modifier, a crash could occur.
61    
62    4.  An attempt to do global matching in pcretest with a zero-length ovector
63        caused a crash.
64    
65    5.  Fixed a memory leak during matching that could occur for a subpattern
66        subroutine call (recursive or otherwise) if the number of captured groups
67        that had to be saved was greater than ten.
68    
69    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
70        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
71        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
72    
73    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
74        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
75    
76    8. When a pattern is compiled, it remembers the highest back reference so that
77       when matching, if the ovector is too small, extra memory can be obtained to
78       use instead. A conditional subpattern whose condition is a check on a
79       capture having happened, such as, for example in the pattern
80       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
81       setting the highest backreference number. This mattered only if pcre_exec()
82       was called with an ovector that was too small to hold the capture, and there
83       was no other kind of back reference (a situation which is probably quite
84       rare). The effect of the bug was that the condition was always treated as
85       FALSE when the capture could not be consulted, leading to a incorrect
86       behaviour by pcre_exec(). This bug has been fixed.
87    
88    9. A reference to a duplicated named group (either a back reference or a test
89       for being set in a conditional) that occurred in a part of the pattern where
90       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
91       to be incorrectly calculated, leading to overwriting.
92    
93    10. A mutually recursive set of back references such as (\2)(\1) caused a
94        segfault at study time (while trying to find the minimum matching length).
95        The infinite loop is now broken (with the minimum length unset, that is,
96        zero).
97    
98    11. If an assertion that was used as a condition was quantified with a minimum
99        of zero, matching went wrong. In particular, if the whole group had
100        unlimited repetition and could match an empty string, a segfault was
101        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
102        assertions to be quantified, but not if they are being used as conditions,
103        so the above pattern is faulted by Perl. PCRE has now been changed so that
104        it also rejects such patterns.
105    
106    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
107        failed to allow the zero-repeat case if pcre2_exec() was called with an
108        ovector too small to capture the group.
109    
110    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
111        Red Hat Product Security:
112    
113        (a) A crash if /K and /F were both set with the option to save the compiled
114        pattern.
115    
116        (b) Another crash if the option to print captured substrings in a callout
117        was combined with setting a null ovector, for example \O\C+ as a subject
118        string.
119    
120    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
121        forward reference repeated a large (but limited) number of times within a
122        repeated outer group that has a zero minimum quantifier, caused incorrect
123        code to be compiled, leading to the error "internal error:
124        previously-checked referenced subpattern not found" when an incorrect
125        memory address was read. This bug was reported as "heap overflow",
126        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
127        CVE-2015-2325.
128    
129    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
130        call within a group that also contained a recursive back reference caused
131        incorrect code to be compiled. This bug was reported as "heap overflow",
132        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
133        number CVE-2015-2326.
134    
135    24. Computing the size of the JIT read-only data in advance has been a source
136        of various issues, and new ones are still appear unfortunately. To fix
137        existing and future issues, size computation is eliminated from the code,
138        and replaced by on-demand memory allocation.
139    
140    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
141        adjacent to the end of the range, and the range contained characters with
142        more than one other case, caused incorrect behaviour when compiled in UTF
143        mode. In that example, the range a-j was left out of the class.
144    
145    26. Fix JIT compilation of conditional blocks, which assertion
146        is converted to (*FAIL). E.g: /(?(?!))/.
147    
148    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
149        discovered by the LLVM fuzzer.
150    
151    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
152        when this assertion was used as a condition, for example (?(?!)a|b). In
153        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
154        error about an unsupported item.
155    
156    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
157        possessification code could take exponential time to complete. A recursion
158        depth limit of 1000 has been imposed to limit the resources used by this
159        optimization.
160    
161    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
162        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
163        because \S ensures they are all in the class. The code for doing this was
164        interacting badly with the code for computing the amount of space needed to
165        compile the pattern, leading to a buffer overflow. This bug was discovered
166        by the LLVM fuzzer.
167    
168    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
169        other kinds of group caused stack overflow at compile time. This bug was
170        discovered by the LLVM fuzzer.
171    
172    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
173        between a subroutine call and its quantifier was incorrectly compiled,
174        leading to buffer overflow or other errors. This bug was discovered by the
175        LLVM fuzzer.
176    
177    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
178        assertion after (?(. The code was failing to check the character after
179        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
180        was discovered by the LLVM fuzzer.
181    
182    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
183        a fixed maximum following a group that contains a subroutine reference was
184        incorrectly compiled and could trigger buffer overflow. This bug was
185        discovered by the LLVM fuzzer.
186    
187    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
188        caused a stack overflow instead of the diagnosis of a non-fixed length
189        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
190    
191    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
192        (e.g. /(?<=\Ka)/) could make pcregrep loop.
193    
194    37. There was a similar problem to 36 in pcretest for global matches.
195    
196    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
197        and a subsequent item in the pattern caused a non-match, backtracking over
198        the repeated \X did not stop, but carried on past the start of the subject,
199        causing reference to random memory and/or a segfault. There were also some
200        other cases where backtracking after \C could crash. This set of bugs was
201        discovered by the LLVM fuzzer.
202    
203    39. The function for finding the minimum length of a matching string could take
204        a very long time if mutual recursion was present many times in a pattern,
205        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
206        method has been implemented. This infelicity was discovered by the LLVM
207        fuzzer.
208    
209    40. Static linking against the PCRE library using the pkg-config module was
210        failing on missing pthread symbols.
211    
212    
213    Version 8.36 26-September-2014
214    ------------------------------
215    
216    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
217        -Wmissing-field-initializers and -Wunused-parameter.
218    
219    2.  The tests for quantifiers being too big (greater than 65535) were being
220        applied after reading the number, and stupidly assuming that integer
221        overflow would give a negative number. The tests are now applied as the
222        numbers are read.
223    
224    3.  Tidy code in pcre_exec.c where two branches that used to be different are
225        now the same.
226    
227    4.  The JIT compiler did not generate match limit checks for certain
228        bracketed expressions with quantifiers. This may lead to exponential
229        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
230        issue should be resolved now.
231    
232    5.  Fixed an issue, which occures when nested alternatives are optimized
233        with table jumps.
234    
235    6.  Inserted two casts and changed some ints to size_t in the light of some
236        reported 64-bit compiler warnings (Bugzilla 1477).
237    
238    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
239        an empty string, which sometimes were behaving incorrectly in the
240        interpreter (though correctly in the JIT matcher). This pcretest input is
241        an example:
242    
243          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
244          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
245    
246        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
247        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
248        for an empty string was breaking the inner loop and carrying on at a lower
249        level, when possessive repeated groups should always return to a higher
250        level as they have no backtrack points in them. The empty string test now
251        occurs at the outer level.
252    
253    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
254        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
255    
256    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
257        doesn't).
258    
259    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
260        to hold a pattern to be incorrectly computed (too small) when there were
261        named back references to duplicated names. This could cause "internal
262        error: code overflow" or "double free or corruption" or other memory
263        handling errors.
264    
265    11. When named subpatterns had the same prefixes, back references could be
266        confused. For example, in this pattern:
267    
268          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
269    
270        the reference to 'Name' was incorrectly treated as a reference to a
271        duplicate name.
272    
273    12. A pattern such as /^s?c/mi8 where the optional character has more than
274        one "other case" was incorrectly compiled such that it would only try to
275        match starting at "c".
276    
277    13. When a pattern starting with \s was studied, VT was not included in the
278        list of possible starting characters; this should have been part of the
279        8.34/18 patch.
280    
281    14. If a character class started [\Qx]... where x is any character, the class
282        was incorrectly terminated at the ].
283    
284    15. If a pattern that started with a caseless match for a character with more
285        than one "other case" was studied, PCRE did not set up the starting code
286        unit bit map for the list of possible characters. Now it does. This is an
287        optimization improvement, not a bug fix.
288    
289    16. The Unicode data tables have been updated to Unicode 7.0.0.
290    
291    17. Fixed a number of memory leaks in pcregrep.
292    
293    18. Avoid a compiler warning (from some compilers) for a function call with
294        a cast that removes "const" from an lvalue by using an intermediate
295        variable (to which the compiler does not object).
296    
297    19. Incorrect code was compiled if a group that contained an internal recursive
298        back reference was optional (had quantifier with a minimum of zero). This
299        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
300        caused segmentation faults because of stack overflows at compile time.
301    
302    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
303        group that is quantified with an indefinite repeat, caused a compile-time
304        loop which used up all the system stack and provoked a segmentation fault.
305        This was not the same bug as 19 above.
306    
307    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
308        Patch by Mike Frysinger.
309    
310    
311    Version 8.35 04-April-2014
312    --------------------------
313    
314    1.  A new flag is set, when property checks are present in an XCLASS.
315        When this flag is not set, PCRE can perform certain optimizations
316        such as studying these XCLASS-es.
317    
318    2.  The auto-possessification of character sets were improved: a normal
319        and an extended character set can be compared now. Furthermore
320        the JIT compiler optimizes more character set checks.
321    
322    3.  Got rid of some compiler warnings for potentially uninitialized variables
323        that show up only when compiled with -O2.
324    
325    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
326        of a match later then the end of the match. The pcretest program was not
327        handling the case sensibly - it was outputting from the start to the next
328        binary zero. It now reports this situation in a message, and outputs the
329        text from the end to the start.
330    
331    5.  Fast forward search is improved in JIT. Instead of the first three
332        characters, any three characters with fixed position can be searched.
333        Search order: first, last, middle.
334    
335    6.  Improve character range checks in JIT. Characters are read by an inprecise
336        function now, which returns with an unknown value if the character code is
337        above a certain threshold (e.g: 256). The only limitation is that the value
338        must be bigger than the threshold as well. This function is useful when
339        the characters above the threshold are handled in the same way.
340    
341    7.  The macros whose names start with RAWUCHAR are placeholders for a future
342        mode in which only the bottom 21 bits of 32-bit data items are used. To
343        make this more memorable for those maintaining the code, the names have
344        been changed to start with UCHAR21, and an extensive comment has been added
345        to their definition.
346    
347    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
348        to the export list in Makefile.am (they were accidentally omitted from the
349        8.34 tarball).
350    
351    9.  The informational output from pcretest used the phrase "starting byte set"
352        which is inappropriate for the 16-bit and 32-bit libraries. As the output
353        for "first char" and "need char" really means "non-UTF-char", I've changed
354        "byte" to "char", and slightly reworded the output. The documentation about
355        these values has also been (I hope) clarified.
356    
357    10. Another JIT related optimization: use table jumps for selecting the correct
358        backtracking path, when more than four alternatives are present inside a
359        bracket.
360    
361    11. Empty match is not possible, when the minimum length is greater than zero,
362        and there is no \K in the pattern. JIT should avoid empty match checks in
363        such cases.
364    
365    12. In a caseless character class with UCP support, when a character with more
366        than one alternative case was not the first character of a range, not all
367        the alternative cases were added to the class. For example, s and \x{17f}
368        are both alternative cases for S: the class [RST] was handled correctly,
369        but [R-T] was not.
370    
371    13. The configure.ac file always checked for pthread support when JIT was
372        enabled. This is not used in Windows, so I have put this test inside a
373        check for the presence of windows.h (which was already tested for).
374    
375    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
376        The algorithm provides a way to skip certain starting offsets, and usually
377        faster than linear prefix searches.
378    
379    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
380        as for 'fr_FR' and 'french'. For some reason, however, it then used the
381        Windows-specific input and output files, which have 'french' screwed in.
382        So this could never have worked. One of the problems with locales is that
383        they aren't always the same. I have now updated RunTest so that it checks
384        the output of the locale test (test 3) against three different output
385        files, and it allows the test to pass if any one of them matches. With luck
386        this should make the test pass on some versions of Solaris where it was
387        failing. Because of the uncertainty, the script did not used to stop if
388        test 3 failed; it now does. If further versions of a French locale ever
389        come to light, they can now easily be added.
390    
391    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
392        there was a message during ./configure, but it did not stop. This now
393        provokes an error. The invalid example in README has been corrected.
394        If a value less than the minimum is given, the minimum value has always
395        been used, but now a warning is given.
396    
397    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
398        was a bug in the test system, which is now fixed. Also, the list of various
399        configurations that are tested for each release did not have one with both
400        16/32 bits and --enable-bar-anycrlf. It now does.
401    
402    18. pcretest was missing "-C bsr" for displaying the \R default setting.
403    
404    19. Little endian PowerPC systems are supported now by the JIT compiler.
405    
406    20. The fast forward newline mechanism could enter to an infinite loop on
407        certain invalid UTF-8 input. Although we don't support these cases
408        this issue can be fixed by a performance optimization.
409    
410    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
411        not take account if existing stack usage. There is now a new global
412        variable called pcre_stack_guard that can be set to point to an external
413        function to check stack availability. It is called at the start of
414        processing every parenthesized group.
415    
416    22. A typo in the code meant that in ungreedy mode the max/min qualifier
417        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
418        not match "ab".
419    
420    23. When UTF was disabled, the JIT program reported some incorrect compile
421        errors. These messages are silenced now.
422    
423    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
424        compiler.
425    
426    25. Change all the temporary files used in RunGrepTest to be different to those
427        used by RunTest so that the tests can be run simultaneously, for example by
428        "make -j check".
429    
430    
431    Version 8.34 15-December-2013
432    -----------------------------
433    
434    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
435        executable memory. Patch inspired by Carsten Klein.
436    
437    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
438        this macro is never tested and has no effect, because the work to support
439        coverage involves only compiling and linking options and special targets in
440        the Makefile. The comment in config.h implied that defining the macro would
441        enable coverage support, which is totally false. There was also support for
442        setting this macro in the CMake files (my fault, I just copied it from
443        configure). SUPPORT_GCOV has now been removed.
444    
445    3.  Make a small performance improvement in strlen16() and strlen32() in
446        pcretest.
447    
448    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
449        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
450        these cases). There was also one in pcretest.c.
451    
452    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
453    
454    6.  In UTF mode, the code for checking whether a group could match an empty
455        string (which is used for indefinitely repeated groups to allow for
456        breaking an infinite loop) was broken when the group contained a repeated
457        negated single-character class with a character that occupied more than one
458        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
459        UTF-8 mode). The effect was undefined: the group might or might not be
460        deemed as matching an empty string, or the program might have crashed.
461    
462    7.  The code for checking whether a group could match an empty string was not
463        recognizing that \h, \H, \v, \V, and \R must match a character.
464    
465    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
466        an empty string. If it can, pcretest shows this in its information output.
467    
468    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
469        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
470        matched by pcre_exec() without using JIT:
471    
472        (a) If the rest of the pattern did not match after a maximal run of
473            grapheme clusters, the code for backing up to try with fewer of them
474            did not always back up over a full grapheme when characters that do not
475            have the modifier quality were involved, e.g. Hangul syllables.
476    
477        (b) If the match point in a subject started with modifier character, and
478            there was no match, the code could incorrectly back up beyond the match
479            point, and potentially beyond the first character in the subject,
480            leading to a segfault or an incorrect match result.
481    
482    10. A conditional group with an assertion condition could lead to PCRE
483        recording an incorrect first data item for a match if no other first data
484        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
485        first data item, and therefore matched "ca" after "c" instead of at the
486        start.
487    
488    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
489        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
490    
491    12. The source of pcregrep now includes z/OS-specific code so that it can be
492        compiled for z/OS as part of the special z/OS distribution.
493    
494    13. Added the -T and -TM options to pcretest.
495    
496    14. The code in pcre_compile.c for creating the table of named capturing groups
497        has been refactored. Instead of creating the table dynamically during the
498        actual compiling pass, the information is remembered during the pre-compile
499        pass (on the stack unless there are more than 20 named groups, in which
500        case malloc() is used) and the whole table is created before the actual
501        compile happens. This has simplified the code (it is now nearly 150 lines
502        shorter) and prepared the way for better handling of references to groups
503        with duplicate names.
504    
505    15. A back reference to a named subpattern when there is more than one of the
506        same name now checks them in the order in which they appear in the pattern.
507        The first one that is set is used for the reference. Previously only the
508        first one was inspected. This change makes PCRE more compatible with Perl.
509    
510    16. Unicode character properties were updated from Unicode 6.3.0.
511    
512    17. The compile-time code for auto-possessification has been refactored, based
513        on a patch by Zoltan Herczeg. It now happens after instead of during
514        compilation. The code is cleaner, and more cases are handled. The option
515        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
516        options in pcretest are provided to set it. It can also be set by
517        (*NO_AUTO_POSSESS) at the start of a pattern.
518    
519    18. The character VT has been added to the default ("C" locale) set of
520        characters that match \s and are generally treated as white space,
521        following this same change in Perl 5.18. There is now no difference between
522        "Perl space" and "POSIX space". Whether VT is treated as white space in
523        other locales depends on the locale.
524    
525    19. The code for checking named groups as conditions, either for being set or
526        for being recursed, has been refactored (this is related to 14 and 15
527        above). Processing unduplicated named groups should now be as fast at
528        numerical groups, and processing duplicated groups should be faster than
529        before.
530    
531    20. Two patches to the CMake build system, by Alexander Barkov:
532    
533          (1) Replace the "source" command by "." in CMakeLists.txt because
534              "source" is a bash-ism.
535    
536          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
537              without these the CMake build does not work on Solaris.
538    
539    21. Perl has changed its handling of \8 and \9. If there is no previously
540        encountered capturing group of those numbers, they are treated as the
541        literal characters 8 and 9 instead of a binary zero followed by the
542        literals. PCRE now does the same.
543    
544    22. Following Perl, added \o{} to specify codepoints in octal, making it
545        possible to specify values greater than 0777 and also making them
546        unambiguous.
547    
548    23. Perl now gives an error for missing closing braces after \x{... instead of
549        treating the string as literal. PCRE now does the same.
550    
551    24. RunTest used to grumble if an inappropriate test was selected explicitly,
552        but just skip it when running all tests. This make it awkward to run ranges
553        of tests when one of them was inappropriate. Now it just skips any
554        inappropriate tests, as it always did when running all tests.
555    
556    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
557        character types such as \d or \w, too many callouts were inserted, and the
558        data that they returned was rubbish.
559    
560    26. In UCP mode, \s was not matching two of the characters that Perl matches,
561        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
562        were matched by \h. The code has now been refactored so that the lists of
563        the horizontal and vertical whitespace characters used for \h and \v (which
564        are defined only in one place) are now also used for \s.
565    
566    27. Add JIT support for the 64 bit TileGX architecture.
567        Patch by Jiong Wang (Tilera Corporation).
568    
569    28. Possessive quantifiers for classes (both explicit and automatically
570        generated) now use special opcodes instead of wrapping in ONCE brackets.
571    
572    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
573        (because it's meaningless), this was not happening when PCRE_CASELESS was
574        set. Not wrong, but inefficient.
575    
576    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
577        properties for \w, \d, etc) is present in a test regex. Otherwise if the
578        test contains no characters greater than 255, Perl doesn't realise it
579        should be using Unicode semantics.
580    
581    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
582        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
583        does in Unicode mode.
584    
585    32. Added the "forbid" facility to pcretest so that putting tests into the
586        wrong test files can sometimes be quickly detected.
587    
588    33. There is now a limit (default 250) on the depth of nesting of parentheses.
589        This limit is imposed to control the amount of system stack used at compile
590        time. It can be changed at build time by --with-parens-nest-limit=xxx or
591        the equivalent in CMake.
592    
593    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
594        errors. Perl warns for these when in warning mode, but PCRE has no facility
595        for giving warnings.
596    
597    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
598        However, this was not working for (?!) because it is optimized to (*FAIL),
599        for which PCRE does not allow quantifiers. The optimization is now disabled
600        when a quantifier follows (?!). I can't see any use for this, but it makes
601        things uniform.
602    
603    36. Perl no longer allows group names to start with digits, so I have made this
604        change also in PCRE. It simplifies the code a bit.
605    
606    37. In extended mode, Perl ignores spaces before a + that indicates a
607        possessive quantifier. PCRE allowed a space before the quantifier, but not
608        before the possessive +. It now does.
609    
610    38. The use of \K (reset reported match start) within a repeated possessive
611        group such as (a\Kb)*+ was not working.
612    
613    40. Document that the same character tables must be used at compile time and
614        run time, and that the facility to pass tables to pcre_exec() and
615        pcre_dfa_exec() is for use only with saved/restored patterns.
616    
617    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
618        features for Builds with MSVC:
619    
620        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
621           builds). The .rc files can be used to set FileDescription and many other
622           attributes.
623    
624        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
625           This allows higher-level build scripts which want .pdb files to avoid
626           hard-coding the exact files needed."
627    
628    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
629        mean "start of word" and "end of word", respectively, as a transition aid.
630    
631    43. A minimizing repeat of a class containing codepoints greater than 255 in
632        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
633        compiled to use the heap for recursion.
634    
635    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
636        is configured.
637    
638    
639    Version 8.33 28-May-2013
640    ------------------------
641    
642    1.  Added 'U' to some constants that are compared to unsigned integers, to
643        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
644        variables that are added to signed variables, to ensure the result is
645        signed and can be negated.
646    
647    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
648        CMake config files.
649    
650    3.  Revise the creation of config.h.generic so that all boolean macros are
651        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
652        overriding via -D on the command line possible.
653    
654    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
655        to unsigned int is reported to make a quite noticeable speed difference in
656        a specific Windows environment. Testing on Linux did also appear to show
657        some benefit (and it is clearly not harmful). Also fixed the definition of
658        Xop which should be unsigned.
659    
660    5.  Related to (4), changing the definition of the intermediate variable cc
661        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
662        improvements.
663    
664    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
665        unnecessary spaces.
666    
667    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
668        and later.
669    
670    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
671    
672    9.  Optimizing fast_forward_start_bits in JIT.
673    
674    10. Adding support for callouts in JIT, and fixing some issues revealed
675        during this work. Namely:
676    
677        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
678    
679        (b) Minimum length was not checked before the matching is started.
680    
681    11. The value of capture_last that is passed to callouts was incorrect in some
682        cases when there was a capture on one path that was subsequently abandoned
683        after a backtrack. Also, the capture_last value is now reset after a
684        recursion, since all captures are also reset in this case.
685    
686    12. The interpreter no longer returns the "too many substrings" error in the
687        case when an overflowing capture is in a branch that is subsequently
688        abandoned after a backtrack.
689    
690    13. In the pathological case when an offset vector of size 2 is used, pcretest
691        now prints out the matched string after a yield of 0 or 1.
692    
693    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
694        Only supported by the JIT compiler at the moment.
695    
696    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
697    
698    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
699        offset of the starting point of the matching process, provided the offsets
700        vector is large enough.
701    
702    17. The \A escape now records a lookbehind value of 1, though its execution
703        does not actually inspect the previous character. This is to ensure that,
704        in partial multi-segment matching, at least one character from the old
705        segment is retained when a new segment is processed. Otherwise, if there
706        are no lookbehinds in the pattern, \A might match incorrectly at the start
707        of a new segment.
708    
709    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
710    
711    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
712        gives some modest performance improvement in 8-bit mode.
713    
714    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
715        be expressed in certain programming languages using Universal Character
716        Names.
717    
718    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
719        which points out that "non characters" are not "characters that may not
720        appear in Unicode strings" but rather "characters that are reserved for
721        internal use and have only local meaning".
722    
723    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
724        there was a conditional group that depended on an assertion, if the
725        assertion was false, the callout that immediately followed the alternation
726        in the condition was skipped when pcre_exec() was used for matching.
727    
728    23. Allow an explicit callout to be inserted before an assertion that is the
729        condition for a conditional group, for compatibility with automatic
730        callouts, which always insert a callout at this point.
731    
732    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
733        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
734    
735    25. (*PRUNE) is now supported by the JIT compiler.
736    
737    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
738    
739    27. Fix the case where there are two or more SKIPs with arguments that may be
740        ignored.
741    
742    28. (*SKIP) is now supported by the JIT compiler.
743    
744    29. (*THEN) is now supported by the JIT compiler.
745    
746    30. Update RunTest with additional test selector options.
747    
748    31. The way PCRE handles backtracking verbs has been changed in two ways.
749    
750        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
751        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
752        backtracking. In some cases this makes it more Perl-compatible, but Perl's
753        rather obscure rules do not always do the same thing.
754    
755        (2) Previously, backtracking verbs were confined within assertions. This is
756        no longer the case for positive assertions, except for (*ACCEPT). Again,
757        this sometimes improves Perl compatibility, and sometimes does not.
758    
759    32. A number of tests that were in test 2 because Perl did things differently
760        have been moved to test 1, because either Perl or PCRE has changed, and
761        these tests are now compatible.
762    
763    32. Backtracking control verbs are now handled in the same way in JIT and
764        interpreter.
765    
766    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
767        contained a forward subroutine reference caused a compile error.
768    
769    34. Auto-detect and optimize limited repetitions in JIT.
770    
771    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
772        blocking (*UTF) etc.
773    
774    36. In the interpreter, maximizing pattern repetitions for characters and
775        character types now use tail recursion, which reduces stack usage.
776    
777    37. The value of the max lookbehind was not correctly preserved if a compiled
778        and saved regex was reloaded on a host of different endianness.
779    
780    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
781        of the compiled pattern block, expand the flags field from 16 to 32 bits
782        because it was almost full.
783    
784    39. Try madvise first before posix_madvise.
785    
786    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
787        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
788        original change locked out all matches of empty strings. This has been
789        changed so that one match of an empty string per line is recognized.
790        Subsequent searches on the same line (for colouring or for --only-matching,
791        for example) do not recognize empty strings.
792    
793    41. Applied a user patch to fix a number of spelling mistakes in comments.
794    
795    42. Data lines longer than 65536 caused pcretest to crash.
796    
797    43. Clarified the data type for length and startoffset arguments for pcre_exec
798        and pcre_dfa_exec in the function-specific man pages, where they were
799        explicitly stated to be in bytes, never having been updated. I also added
800        some clarification to the pcreapi man page.
801    
802    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
803        a segmentation fault.
804    
805    
806    Version 8.32 30-November-2012
807    -----------------------------
808    
809    1.  Improved JIT compiler optimizations for first character search and single
810        character iterators.
811    
812    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
813        Patch by Daniel Richard G.
814    
815    3.  Single character iterator optimizations in the JIT compiler.
816    
817    4.  Improved JIT compiler optimizations for character ranges.
818    
819    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
820        Reported by Giuseppe D'Angelo.
821    
822    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
823        of a line, was being set incorrectly in cases where .* appeared inside
824        atomic brackets at the start of a pattern, or where there was a subsequent
825        *PRUNE or *SKIP.
826    
827    7.  Improved instruction cache flush for POWER/PowerPC.
828        Patch by Daniel Richard G.
829    
830    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
831        grep:
832    
833        (a) There is now no limit to the number of patterns to be matched.
834    
835        (b) An error is given if a pattern is too long.
836    
837        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
838            are now supported.
839    
840        (d) --exclude-from and --include-from (multiple use) have been added.
841    
842        (e) Exclusions and inclusions now apply to all files and directories, not
843            just to those obtained from scanning a directory recursively.
844    
845        (f) Multiple uses of -f and --file-list are now supported.
846    
847        (g) In a Windows environment, the default for -d has been changed from
848            "read" (the GNU grep default) to "skip", because otherwise the presence
849            of a directory in the file list provokes an error.
850    
851        (h) The documentation has been revised and clarified in places.
852    
853    9.  Improve the matching speed of capturing brackets.
854    
855    10. Changed the meaning of \X so that it now matches a Unicode extended
856        grapheme cluster.
857    
858    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
859        out POSIX threads when JIT support is configured.
860    
861    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
862    
863    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
864        structure could go wrong in environments where size_t is not the same size
865        as int.
866    
867    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
868        set.
869    
870    15. The EBCDIC support had decayed; later updates to the code had included
871        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
872        general tidy up of EBCDIC-related issues, and the documentation was also
873        not quite right. There is now a test that can be run on ASCII systems to
874        check some of the EBCDIC-related things (but is it not a full test).
875    
876    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
877        in a small tidy to the code.
878    
879    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
880    
881    18. If the --only-matching (-o) option in pcregrep is specified multiple
882        times, each one causes appropriate output. For example, -o1 -o2 outputs the
883        substrings matched by the 1st and 2nd capturing parentheses. A separating
884        string can be specified by --om-separator (default empty).
885    
886    19. Improving the first n character searches.
887    
888    20. Turn case lists for horizontal and vertical white space into macros so that
889        they are defined only once.
890    
891    21. This set of changes together give more compatible Unicode case-folding
892        behaviour for characters that have more than one other case when UCP
893        support is available.
894    
895        (a) The Unicode property table now has offsets into a new table of sets of
896            three or more characters that are case-equivalent. The MultiStage2.py
897            script that generates these tables (the pcre_ucd.c file) now scans
898            CaseFolding.txt instead of UnicodeData.txt for character case
899            information.
900    
901        (b) The code for adding characters or ranges of characters to a character
902            class has been abstracted into a generalized function that also handles
903            case-independence. In UTF-mode with UCP support, this uses the new data
904            to handle characters with more than one other case.
905    
906        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
907            whose other case is greater than 256 are now correctly matched
908            caselessly. Previously, the high codepoint matched the low one, but not
909            vice versa.
910    
911        (d) The processing of \h, \H, \v, and \ in character classes now makes use
912            of the new class addition function, using character lists defined as
913            macros alongside the case definitions of 20 above.
914    
915        (e) Caseless back references now work with characters that have more than
916            one other case.
917    
918        (f) General caseless matching of characters with more than one other case
919            is supported.
920    
921    22. Unicode character properties were updated from Unicode 6.2.0
922    
923    23. Improved CMake support under Windows. Patch by Daniel Richard G.
924    
925    24. Add support for 32-bit character strings, and UTF-32
926    
927    25. Major JIT compiler update (code refactoring and bugfixing).
928        Experimental Sparc 32 support is added.
929    
930    26. Applied a modified version of Daniel Richard G's patch to create
931        pcre.h.generic and config.h.generic by "make" instead of in the
932        PrepareRelease script.
933    
934    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
935        pcre_compile.c when checking for a zero character.
936    
937    28. Introducing a native interface for JIT. Through this interface, the compiled
938        machine code can be directly executed. The purpose of this interface is to
939        provide fast pattern matching, so several sanity checks are not performed.
940        However, feature tests are still performed. The new interface provides
941        1.4x speedup compared to the old one.
942    
943    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
944        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
945        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
946    
947    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
948        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
949        the "old" RFC 2279). Instead, it ended up passing a negative length to
950        pcre_exec().
951    
952    31. Add support for GCC's visibility feature to hide internal functions.
953    
954    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
955        "unknown -C option" after outputting 0 or 1.
956    
957    33. There is now support for generating a code coverage report for the test
958        suite in environments where gcc is the compiler and lcov is installed. This
959        is mainly for the benefit of the developers.
960    
961    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
962        unaddressable using valgrind annotations, allowing valgrind to detect
963        invalid memory accesses. This is mainly for the benefit of the developers.
964    
965    25. (*UTF) can now be used to start a pattern in any of the three libraries.
966    
967    26. Give configure error if --enable-cpp but no C++ compiler found.
968    
969    
970    Version 8.31 06-July-2012
971    -------------------------
972    
973    1.  Fixing a wrong JIT test case and some compiler warnings.
974    
975    2.  Removed a bashism from the RunTest script.
976    
977    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
978        to unsigned type, result still unsigned" that was given by an MS compiler
979        on encountering the code "-sizeof(xxx)".
980    
981    4.  Partial matching support is added to the JIT compiler.
982    
983    5.  Fixed several bugs concerned with partial matching of items that consist
984        of more than one character:
985    
986        (a) /^(..)\1/ did not partially match "aba" because checking references was
987            done on an "all or nothing" basis. This also applied to repeated
988            references.
989    
990        (b) \R did not give a hard partial match if \r was found at the end of the
991            subject.
992    
993        (c) \X did not give a hard partial match after matching one or more
994            characters at the end of the subject.
995    
996        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
997            a partial match for the string "\r".
998    
999        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1000            a partial match for a CR character at the end of the subject string.
1001    
1002    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1003        running pcretest, the text "(JIT)" added to the output whenever JIT is
1004        actually used to run the match.
1005    
1006    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1007        or /S+[+] with a digit between 1 and 7.
1008    
1009    8.  OP_NOT now supports any UTF character not just single-byte ones.
1010    
1011    9.  (*MARK) control verb is now supported by the JIT compiler.
1012    
1013    10. The command "./RunTest list" lists the available tests without actually
1014        running any of them. (Because I keep forgetting what they all are.)
1015    
1016    11. Add PCRE_INFO_MAXLOOKBEHIND.
1017    
1018    12. Applied a (slightly modified) user-supplied patch that improves performance
1019        when the heap is used for recursion (compiled with --disable-stack-for-
1020        recursion). Instead of malloc and free for each heap frame each time a
1021        logical recursion happens, frames are retained on a chain and re-used where
1022        possible. This sometimes gives as much as 30% improvement.
1023    
1024    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1025        call.
1026    
1027    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1028    
1029    15. It is now possible to link pcretest with libedit as an alternative to
1030        libreadline.
1031    
1032    16. (*COMMIT) control verb is now supported by the JIT compiler.
1033    
1034    17. The Unicode data tables have been updated to Unicode 6.1.0.
1035    
1036    18. Added --file-list option to pcregrep.
1037    
1038    19. Added binary file support to pcregrep, including the -a, --binary-files,
1039        -I, and --text options.
1040    
1041    20. The madvise function is renamed for posix_madvise for QNX compatibility
1042        reasons. Fixed by Giuseppe D'Angelo.
1043    
1044    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1045    
1046    22. Changed the option for creating symbolic links for 16-bit man pages from
1047        -s to -sf so that re-installing does not cause issues.
1048    
1049    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1050    
1051    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1052        matches in certain environments (the workspace was not being correctly
1053        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1054        some of the workspace data at the beginning of a restart.
1055    
1056    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1057        was not doing so when it should - probably a typo introduced by SVN 528
1058        (change 8.10/14).
1059    
1060    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1061        \w+ when the character tables indicated that \x{c4} was a word character.
1062        There were several related cases, all because the tests for doing a table
1063        lookup were testing for characters less than 127 instead of 255.
1064    
1065    27. If a pattern contains capturing parentheses that are not used in a match,
1066        their slots in the ovector are set to -1. For those that are higher than
1067        any matched groups, this happens at the end of processing. In the case when
1068        there were back references that the ovector was too small to contain
1069        (causing temporary malloc'd memory to be used during matching), and the
1070        highest capturing number was not used, memory off the end of the ovector
1071        was incorrectly being set to -1. (It was using the size of the temporary
1072        memory instead of the true size.)
1073    
1074    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1075        ovector size, it uses memory at the end of the block that it has got.
1076    
1077    29. Check for an overlong MARK name and give an error at compile time. The
1078        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1079    
1080    30. JIT compiler update.
1081    
1082    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1083        Rill for the patch.
1084    
1085    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1086    
1087    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1088    
1089    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1090        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1091        was enabled.
1092    
1093    35. Improve JIT code generation for greedy plus quantifier.
1094    
1095    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1096        1 to "aa" instead of to an empty string. The bug affected repeated groups
1097        that could potentially match an empty string.
1098    
1099    37. Optimizing single character iterators in JIT.
1100    
1101    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1102        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1103        codepoints that are too big for the mode are faulted, and in a UTF mode,
1104        disallowed codepoints are also faulted.
1105    
1106    39. If PCRE was compiled with UTF support, in three places in the DFA
1107        matcher there was code that should only have been obeyed in UTF mode, but
1108        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1109        processing when bytes with values greater than 127 were present. In 16-bit
1110        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1111        both cases the values are those that cannot be the first data item in a UTF
1112        character. The three items that might have provoked this were recursions,
1113        possessively repeated groups, and atomic groups.
1114    
1115    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1116        and pcregrep, because some OS require shared objects to be explicitly
1117        passed to ld, causing the link step to fail if they are not.
1118    
1119    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1120        mode, patterns that started with \h* or \R* might be incorrectly matched.
1121    
1122    
1123    Version 8.30 04-February-2012
1124    -----------------------------
1125    
1126    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1127        name is defined in ctype.h.
1128    
1129    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1130        only in quite long subpatterns.
1131    
1132    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1133        since it was replaced by pcre_fullinfo() in February 2000.
1134    
1135    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1136        match a (*MARK), and the match failed at the start of the subject, a
1137        reference to memory before the start of the subject could occur. This bug
1138        was introduced by fix 17 of release 8.21.
1139    
1140    5.  A reference to an unset group with zero minimum repetition was giving
1141        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1142        /(another)?(\1?)test/ matched against "hello world test". This bug was
1143        introduced in release 8.13.
1144    
1145    6.  Add support for 16-bit character strings (a large amount of work involving
1146        many changes and refactorings).
1147    
1148    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1149        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1150        from a file.
1151    
1152    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1153        rounding is not applied in this particular case).
1154    
1155    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1156        if they appear, or are escaped, in patterns.
1157    
1158    10. Get rid of a number of -Wunused-but-set-variable warnings.
1159    
1160    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1161        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1162        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1163        also returns the mark "x". This bug applied to capturing parentheses,
1164        non-capturing parentheses, and atomic parentheses. It also applied to some
1165        assertions.
1166    
1167    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1168        information out of configure.ac instead of relying on pcre.h.generic, which
1169        is not stored in the repository.
1170    
1171    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1172        -lreadline.
1173    
1174    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1175    
1176    15. Applied Graycode's patch to put the top-level frame on the stack rather
1177        than the heap when not using the stack for recursion. This gives a
1178        performance improvement in many cases when recursion is not deep.
1179    
1180    16. Experimental code added to "pcretest -C" to output the stack frame size.
1181    
1182    
1183    Version 8.21 12-Dec-2011
1184    ------------------------
1185    
1186    1.  Updating the JIT compiler.
1187    
1188    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1189        are added as well.
1190    
1191    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1192        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1193        calling _pcre_jit_exec. Some extra comments are added.
1194    
1195    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1196        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1197        was introduced by change 18 for 8.20.
1198    
1199    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1200        ECMA-262 standard.
1201    
1202    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1203        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1204        This bug was probably introduced by change 9 of 8.13.
1205    
1206    7.  While fixing 6 above, I noticed that a number of other items were being
1207        incorrectly rejected as "not fixed length". This arose partly because newer
1208        opcodes had not been added to the fixed-length checking code. I have (a)
1209        corrected the bug and added tests for these items, and (b) arranged for an
1210        error to occur if an unknown opcode is encountered while checking for fixed
1211        length instead of just assuming "not fixed length". The items that were
1212        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1213        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1214        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1215    
1216    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1217        being incorrectly compiled and would have given unpredicatble results.
1218    
1219    9.  A possessively repeated subpattern with minimum repeat count greater than
1220        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1221        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1222        the first (A) could occur when it should not.
1223    
1224    10. Add a cast and remove a redundant test from the code.
1225    
1226    11. JIT should use pcre_malloc/pcre_free for allocation.
1227    
1228    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1229        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1230        is anything other than /usr, -L is still shown).
1231    
1232    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1233    
1234    14. Perl does not support \N without a following name in a [] class; PCRE now
1235        also gives an error.
1236    
1237    15. If a forward reference was repeated with an upper limit of around 2000,
1238        it caused the error "internal error: overran compiling workspace". The
1239        maximum number of forward references (including repeats) was limited by the
1240        internal workspace, and dependent on the LINK_SIZE. The code has been
1241        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1242        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1243        of around 200,000 forward references. While doing this, I also speeded up
1244        the filling in of repeated forward references.
1245    
1246    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1247        incorrectly expecting the subject to contain another "a" after the start.
1248    
1249    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1250        in the match, the SKIP should be ignored. This was not happening; instead
1251        the SKIP was being treated as NOMATCH. For patterns such as
1252        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1253        tested.
1254    
1255    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1256        now much more compatible with Perl, in particular in cases where the result
1257        is a non-match for a non-anchored pattern. For example, if
1258        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1259        "m", where previously it did not return a name. A side effect of this
1260        change is that for partial matches, the last encountered mark name is
1261        returned, as for non matches. A number of tests that were previously not
1262        Perl-compatible have been moved into the Perl-compatible test files. The
1263        refactoring has had the pleasing side effect of removing one argument from
1264        the match() function, thus reducing its stack requirements.
1265    
1266    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1267        subsequent uses of /S (without +) incorrectly behaved like /S+.
1268    
1269    21. Retrieve executable code size support for the JIT compiler and fixing
1270        some warnings.
1271    
1272    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1273        not work when the shorter character appeared right at the end of the
1274        subject string.
1275    
1276    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1277        systems.
1278    
1279    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1280        output it when the /M option is used in pcretest.
1281    
1282    25. The CheckMan script was not being included in the distribution. Also, added
1283        an explicit "perl" to run Perl scripts from the PrepareRelease script
1284        because this is reportedly needed in Windows.
1285    
1286    26. If study data was being save in a file and studying had not found a set of
1287        "starts with" bytes for the pattern, the data written to the file (though
1288        never used) was taken from uninitialized memory and so caused valgrind to
1289        complain.
1290    
1291    27. Updated RunTest.bat as provided by Sheri Pierce.
1292    
1293    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1294    
1295    29. Computation of memory usage for the table of capturing group names was
1296        giving an unnecessarily large value.
1297    
1298    
1299    Version 8.20 21-Oct-2011
1300    ------------------------
1301    
1302    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1303        a POSIX class. After further experiments with Perl, which convinced me that
1304        Perl has bugs and confusions, a closing square bracket is no longer allowed
1305        in a POSIX name. This bug also affected patterns with classes that started
1306        with full stops.
1307    
1308    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1309        captured substring, but while checking the failing first alternative,
1310        substring 1 is temporarily captured. If the output vector supplied to
1311        pcre_exec() was not big enough for this capture, the yield of the function
1312        was still zero ("insufficient space for captured substrings"). This cannot
1313        be totally fixed without adding another stack variable, which seems a lot
1314        of expense for a edge case. However, I have improved the situation in cases
1315        such as /(a)(b)x|abc/ matched against "abc", where the return code
1316        indicates that fewer than the maximum number of slots in the ovector have
1317        been set.
1318    
1319    3.  Related to (2) above: when there are more back references in a pattern than
1320        slots in the output vector, pcre_exec() uses temporary memory during
1321        matching, and copies in the captures as far as possible afterwards. It was
1322        using the entire output vector, but this conflicts with the specification
1323        that only 2/3 is used for passing back captured substrings. Now it uses
1324        only the first 2/3, for compatibility. This is, of course, another edge
1325        case.
1326    
1327    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1328        main code base, and can be used by building with --enable-jit. When this is
1329        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1330        runtime --no-jit option is given.
1331    
1332    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1333        ovector, the return from the function was zero, implying that there were
1334        other matches that did not fit. The correct "exactly full" value is now
1335        returned.
1336    
1337    6.  If a subpattern that was called recursively or as a subroutine contained
1338        (*PRUNE) or any other control that caused it to give a non-standard return,
1339        invalid errors such as "Error -26 (nested recursion at the same subject
1340        position)" or even infinite loops could occur.
1341    
1342    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1343        computing the minimum length on reaching *ACCEPT, and so ended up with the
1344        wrong value of 1 rather than 0. Further investigation indicates that
1345        computing a minimum subject length in the presence of *ACCEPT is difficult
1346        (think back references, subroutine calls), and so I have changed the code
1347        so that no minimum is registered for a pattern that contains *ACCEPT.
1348    
1349    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1350        it was not handled as intended. [But see 16 below.]
1351    
1352    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1353        Sheri Pierce.
1354    
1355    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1356        the first byte in a match must be "a".
1357    
1358    11. Change 17 for 8.13 increased the recursion depth for patterns like
1359        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1360        pattern contains any instances of (*THEN). If it does not, the old
1361        optimizations are restored. It would be nice to do this on a per-group
1362        basis, but at the moment that is not feasible.
1363    
1364    12. In some environments, the output of pcretest -C is CRLF terminated. This
1365        broke RunTest's code that checks for the link size. A single white space
1366        character after the value is now allowed for.
1367    
1368    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1369        For "fr", it uses the Windows-specific input and output files.
1370    
1371    14. If (*THEN) appeared in a group that was called recursively or as a
1372        subroutine, it did not work as intended. [But see next item.]
1373    
1374    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1375        pattern fragments (but not containing any | characters). If A and B are
1376        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1377        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1378        D. In other words, Perl considers parentheses that do not contain any |
1379        characters to be part of a surrounding alternative, whereas PCRE was
1380        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1381        differently. PCRE now behaves in the same way as Perl, except in the case
1382        of subroutine/recursion calls such as (?1) which have in any case always
1383        been different (but PCRE had them first :-).
1384    
1385    16. Related to 15 above: Perl does not treat the | in a conditional group as
1386        creating alternatives. Such a group is treated in the same way as an
1387        ordinary group without any | characters when processing (*THEN). PCRE has
1388        been changed to match Perl's behaviour.
1389    
1390    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1391        RunGrepTest script failed.
1392    
1393    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1394        inevitable for groups that contain captures, but it can lead to a lot of
1395        stack use in large patterns. The old behaviour has been restored for atomic
1396        groups that do not contain any capturing parentheses.
1397    
1398    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1399        suppress the check for a minimum subject length at run time. (If it was
1400        given to pcre_exec() or pcre_dfa_exec() it did work.)
1401    
1402    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1403        fail to work when decoding hex characters in data strings in EBCDIC
1404        environments.
1405    
1406    21. It appears that in at least one Mac OS environment, the isxdigit() function
1407        is implemented as a macro that evaluates to its argument more than once,
1408        contravening the C 90 Standard (I haven't checked a later standard). There
1409        was an instance in pcretest which caused it to go wrong when processing
1410        \x{...} escapes in subject strings. The has been rewritten to avoid using
1411        things like p++ in the argument of isxdigit().
1412    
1413    
1414    Version 8.13 16-Aug-2011
1415    ------------------------
1416    
1417    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1418    
1419    2.  Two minor typos in pcre_internal.h have been fixed.
1420    
1421    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1422        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1423        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1424    
1425    4.  There were a number of related bugs in the code for matching backrefences
1426        caselessly in UTF-8 mode when codes for the characters concerned were
1427        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1428        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1429        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1430        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1431        2-byte code at the end of the subject (it thought there wasn't enough data
1432        left).
1433    
1434    5.  Comprehensive information about what went wrong is now returned by
1435        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1436        as the output vector has at least 2 elements. The offset of the start of
1437        the failing character and a reason code are placed in the vector.
1438    
1439    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1440        now returned is for the first byte of the failing character, instead of the
1441        last byte inspected. This is an incompatible change, but I hope it is small
1442        enough not to be a problem. It makes the returned offset consistent with
1443        pcre_exec() and pcre_dfa_exec().
1444    
1445    7.  pcretest now gives a text phrase as well as the error number when
1446        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1447        failure, the offset and reason code are output.
1448    
1449    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1450        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1451        back over a single character (\n). This seems wrong (because it treated the
1452        two characters as a single entity when going forwards), conflicts with the
1453        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1454        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1455        has been changed.
1456    
1457    9.  Some internal refactoring has changed the processing so that the handling
1458        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1459        time (the PCRE_DOTALL option was changed this way some time ago: version
1460        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1461        which was always a bit of a fudge. It also means that there is one less
1462        argument for the match() function, which reduces its stack requirements
1463        slightly. This change also fixes an incompatibility with Perl: the pattern
1464        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1465    
1466    10. More internal refactoring has drastically reduced the number of recursive
1467        calls to match() for possessively repeated groups such as (abc)++ when
1468        using pcre_exec().
1469    
1470    11. While implementing 10, a number of bugs in the handling of groups were
1471        discovered and fixed:
1472    
1473        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1474        (a|)*(?1) gave a compile-time internal error.
1475        ((a|)+)+  did not notice that the outer group could match an empty string.
1476        (^a|^)+   was not marked as anchored.
1477        (.*a|.*)+ was not marked as matching at start or after a newline.
1478    
1479    12. Yet more internal refactoring has removed another argument from the match()
1480        function. Special calls to this function are now indicated by setting a
1481        value in a variable in the "match data" data block.
1482    
1483    13. Be more explicit in pcre_study() instead of relying on "default" for
1484        opcodes that mean there is no starting character; this means that when new
1485        ones are added and accidentally left out of pcre_study(), testing should
1486        pick them up.
1487    
1488    14. The -s option of pcretest has been documented for ages as being an old
1489        synonym of -m (show memory usage). I have changed it to mean "force study
1490        for every regex", that is, assume /S for every regex. This is similar to -i
1491        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1492        using it. It makes it easier to run collections of tests with and without
1493        study enabled, and thereby test pcre_study() more easily. All the standard
1494        tests are now run with and without -s (but some patterns can be marked as
1495        "never study" - see 20 below).
1496    
1497    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1498        restoration of the capturing data to the outer values was not happening
1499        correctly.
1500    
1501    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1502        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1503        pattern had matched an empty string, and so incorrectly returned a no
1504        match.
1505    
1506    17. There was optimizing code for the last branch of non-capturing parentheses,
1507        and also for the obeyed branch of a conditional subexpression, which used
1508        tail recursion to cut down on stack usage. Unfortunately, now that there is
1509        the possibility of (*THEN) occurring in these branches, tail recursion is
1510        no longer possible because the return has to be checked for (*THEN). These
1511        two optimizations have therefore been removed. [But see 8.20/11 above.]
1512    
1513    18. If a pattern containing \R was studied, it was assumed that \R always
1514        matched two bytes, thus causing the minimum subject length to be
1515        incorrectly computed because \R can also match just one byte.
1516    
1517    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1518        was incorrectly computed.
1519    
1520    20. If /S is present twice on a test pattern in pcretest input, it now
1521        *disables* studying, thereby overriding the use of -s on the command line
1522        (see 14 above). This is necessary for one or two tests to keep the output
1523        identical in both cases.
1524    
1525    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1526        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1527    
1528    22. When an atomic group that contained a capturing parenthesis was
1529        successfully matched, but the branch in which it appeared failed, the
1530        capturing was not being forgotten if a higher numbered group was later
1531        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1532        group 1 to "a", when in fact it should be unset. This applied to multi-
1533        branched capturing and non-capturing groups, repeated or not, and also to
1534        positive assertions (capturing in negative assertions does not happen
1535        in PCRE) and also to nested atomic groups.
1536    
1537    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1538        subject after a captured substring, to make it easier to tell which of a
1539        number of identical substrings has been captured.
1540    
1541    24. The way atomic groups are processed by pcre_exec() has been changed so that
1542        if they are repeated, backtracking one repetition now resets captured
1543        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1544        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1545        "aaa". Previously, it would have been "a". As part of this code
1546        refactoring, the way recursive calls are handled has also been changed.
1547    
1548    25. If an assertion condition captured any substrings, they were not passed
1549        back unless some other capturing happened later. For example, if
1550        (?(?=(a))a) was matched against "a", no capturing was returned.
1551    
1552    26. When studying a pattern that contained subroutine calls or assertions,
1553        the code for finding the minimum length of a possible match was handling
1554        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1555        group 1 called group 2 while simultaneously a separate group 2 called group
1556        1). A stack overflow occurred in this case. I have fixed this by limiting
1557        the recursion depth to 10.
1558    
1559    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1560        Fortmann. This supports explicit test numbers on the command line, and has
1561        argument validation and error reporting.
1562    
1563    28. An instance of \X with an unlimited repeat could fail if at any point the
1564        first character it looked at was a mark character.
1565    
1566    29. Some minor code refactoring concerning Unicode properties and scripts
1567        should reduce the stack requirement of match() slightly.
1568    
1569    30. Added the '=' option to pcretest to check the setting of unused capturing
1570        slots at the end of the pattern, which are documented as being -1, but are
1571        not included in the return count.
1572    
1573    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1574        compiled something random. Now it gives a compile-time error (as does
1575        Perl).
1576    
1577    32. A *MARK encountered during the processing of a positive assertion is now
1578        recorded and passed back (compatible with Perl).
1579    
1580    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1581        had alternative anchored branches, the search for a second match in a line
1582        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1583        matched the line "0102" twice. The same bug affected patterns that started
1584        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1585        twice.
1586    
1587    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1588        does, and because of capturing effects, quantifying parenthesized
1589        assertions may at times be useful. Quantifiers are now allowed for
1590        parenthesized assertions.
1591    
1592    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1593    
1594    36. \g was being checked for fancy things in a character class, when it should
1595        just be a literal "g".
1596    
1597    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1598        appearance of a nested POSIX class supersedes an apparent external class.
1599        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1600        unescaped square brackets may also appear as part of class names. For
1601        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1602        more like Perl. (But see 8.20/1 above.)
1603    
1604    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1605        was because it thought it was \N{name}, which is not supported).
1606    
1607    39. Add minix to OS list not supporting the -S option in pcretest.
1608    
1609    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1610        cannot analyze patterns in sufficient detail to catch mutual recursions
1611        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1612        subgroup is called recursively as a subpattern for a second time at the
1613        same position in the subject string. In previous releases this might have
1614        been caught by the recursion limit, or it might have run out of stack.
1615    
1616    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1617        happen only once. PCRE was, however incorrectly giving a compile time error
1618        "recursive call could loop indefinitely" because it cannot analyze the
1619        pattern in sufficient detail. The compile time test no longer happens when
1620        PCRE is compiling a conditional subpattern, but actual runaway loops are
1621        now caught at runtime (see 40 above).
1622    
1623    42. It seems that Perl allows any characters other than a closing parenthesis
1624        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1625        has been changed to be the same.
1626    
1627    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1628        as not to get warnings when autogen.sh is called. Also changed
1629        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1630    
1631    44. To help people who use pcregrep to scan files containing exceedingly long
1632        lines, the following changes have been made:
1633    
1634        (a) The default value of the buffer size parameter has been increased from
1635            8K to 20K. (The actual buffer used is three times this size.)
1636    
1637        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1638            PCRE is built.
1639    
1640        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1641            to be set at run time.
1642    
1643        (d) Numerical values in pcregrep options can be followed by K or M, for
1644            example --buffer-size=50K.
1645    
1646        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1647            given and the return code is set to 2.
1648    
1649    45. Add a pointer to the latest mark to the callout data block.
1650    
1651    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1652        partial match of an empty string instead of no match. This was specific to
1653        the use of ".".
1654    
1655    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1656        complete match instead of a partial match. This bug was dependent on both
1657        the PCRE_UTF8 and PCRE_DOTALL options being set.
1658    
1659    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1660        starting byte set, because \b was not being ignored.
1661    
1662    
1663    Version 8.12 15-Jan-2011
1664    ------------------------
1665    
1666    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1667        checks for such things as part of the documentation building process.
1668    
1669    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1670        --match-limit and --recursion-limit options (added for 8.11). In
1671        particular, this made one of the standard tests fail. (The integer value
1672        went into the wrong half of a long int.)
1673    
1674    3.  If the --colour option was given to pcregrep with -v (invert match), it
1675        did strange things, either producing crazy output, or crashing. It should,
1676        of course, ignore a request for colour when reporting lines that do not
1677        match.
1678    
1679    4.  Another pcregrep bug caused similar problems if --colour was specified with
1680        -M (multiline) and the pattern match finished with a line ending.
1681    
1682    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1683        matched in multiline mode, the following line was shown as part of the
1684        match. This seems wrong, so I have changed it.
1685    
1686    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1687        the check for further matches in the same line (so they could be coloured)
1688        to overrun the end of the current line. If another match was found, it was
1689        incorrectly shown (and then shown again when found in the next line).
1690    
1691    7.  If pcregrep was compiled under Windows, there was a reference to the
1692        function pcregrep_exit() before it was defined. I am assuming this was
1693        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1694        reported by a user. I've moved the definition above the reference.
1695    
1696    
1697    Version 8.11 10-Dec-2010
1698    ------------------------
1699    
1700    1.  (*THEN) was not working properly if there were untried alternatives prior
1701        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1702        backtracked to try for "b" instead of moving to the next alternative branch
1703        at the same level (in this case, to look for "c"). The Perl documentation
1704        is clear that when (*THEN) is backtracked onto, it goes to the "next
1705        alternative in the innermost enclosing group".
1706    
1707    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1708        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1709        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1710        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1711        (*THEN).
1712    
1713    3.  If \s appeared in a character class, it removed the VT character from
1714        the class, even if it had been included by some previous item, for example
1715        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1716        of \s, but is part of the POSIX "space" class.)
1717    
1718    4.  A partial match never returns an empty string (because you can always
1719        match an empty string at the end of the subject); however the checking for
1720        an empty string was starting at the "start of match" point. This has been
1721        changed to the "earliest inspected character" point, because the returned
1722        data for a partial match starts at this character. This means that, for
1723        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1724        (previously it gave "no match").
1725    
1726    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1727        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1728        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1729        has an implication that the given string is incomplete (because a partial
1730        match is preferred over a full match). For this reason, these items now
1731        give a partial match in this situation. [Aside: previously, the one case
1732        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1733        match rather than a full match, which was wrong by the old rules, but is
1734        now correct.]
1735    
1736    6.  There was a bug in the handling of #-introduced comments, recognized when
1737        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1738        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1739        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1740        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1741        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1742        places in pcre_compile().
1743    
1744    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1745        comments when looking ahead for named forward references to subpatterns,
1746        the only newline sequence it recognized was NL. It now handles newlines
1747        according to the set newline convention.
1748    
1749    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1750        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1751        cater for a lack of strerror(). These oversights have been fixed.
1752    
1753    9.  Added --match-limit and --recursion-limit to pcregrep.
1754    
1755    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1756    
1757    11. When the -o option was used, pcregrep was setting a return code of 1, even
1758        when matches were found, and --line-buffered was not being honoured.
1759    
1760    12. Added an optional parentheses number to the -o and --only-matching options
1761        of pcregrep.
1762    
1763    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1764        can match an empty string. The code to do it in pcretest and pcredemo
1765        needed fixing:
1766    
1767        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1768            only one byte after an empty string match just before CRLF (this case
1769            just got forgotten; "any" and "anycrlf" were OK).
1770    
1771        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1772            mode when an empty string match preceded an ASCII character followed by
1773            a non-ASCII character. (The code for advancing by one character rather
1774            than one byte was nonsense.)
1775    
1776        (c) The pcredemo.c sample program did not have any code at all to handle
1777            the cases when CRLF is a valid newline sequence.
1778    
1779    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1780        as a starting offset was within the subject string. There is now a new
1781        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1782        negative or greater than the length of the string. In order to test this,
1783        pcretest is extended to allow the setting of negative starting offsets.
1784    
1785    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1786        starting offset points to the beginning of a UTF-8 character was
1787        unnecessarily clumsy. I tidied it up.
1788    
1789    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1790        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1791    
1792    17. Nobody had reported that the --include_dir option, which was added in
1793        release 7.7 should have been called --include-dir (hyphen, not underscore)
1794        for compatibility with GNU grep. I have changed it to --include-dir, but
1795        left --include_dir as an undocumented synonym, and the same for
1796        --exclude-dir, though that is not available in GNU grep, at least as of
1797        release 2.5.4.
1798    
1799    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1800        characters from a string of bytes) have been redefined so as not to use
1801        loops, in order to improve performance in some environments. At the same
1802        time, I abstracted some of the common code into auxiliary macros to save
1803        repetition (this should not affect the compiled code).
1804    
1805    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1806        compile-time error is now given if \c is not followed by an ASCII
1807        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1808        different, and any byte value is allowed.)
1809    
1810    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1811        START_OPTIMIZE option, which is now allowed at compile time - but just
1812        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1813        to pcregrep and other applications that have no direct access to PCRE
1814        options. The new /Y option in pcretest sets this option when calling
1815        pcre_compile().
1816    
1817    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1818        back references. Groups containing recursive back references were forced to
1819        be atomic by that change, but in the case of named groups, the amount of
1820        memory required was incorrectly computed, leading to "Failed: internal
1821        error: code overflow". This has been fixed.
1822    
1823    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1824        pcretest.c, to avoid build problems in some Borland environments.
1825    
1826    
1827    Version 8.10 25-Jun-2010
1828    ------------------------
1829    
1830    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1831        THEN.
1832    
1833    2.  (*ACCEPT) was not working when inside an atomic group.
1834    
1835    3.  Inside a character class, \B is treated as a literal by default, but
1836        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1837        causes the error). The code is unchanged, but I tidied the documentation.
1838    
1839    4.  Inside a character class, PCRE always treated \R and \X as literals,
1840        whereas Perl faults them if its -w option is set. I have changed PCRE so
1841        that it faults them when PCRE_EXTRA is set.
1842    
1843    5.  Added support for \N, which always matches any character other than
1844        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1845    
1846    6.  When compiling pcregrep with newer versions of gcc which may have
1847        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1848        declared with attribute warn_unused_result" were given. Just casting the
1849        result to (void) does not stop the warnings; a more elaborate fudge is
1850        needed. I've used a macro to implement this.
1851    
1852    7.  Minor change to pcretest.c to avoid a compiler warning.
1853    
1854    8.  Added four artifical Unicode properties to help with an option to make
1855        \s etc use properties (see next item). The new properties are: Xan
1856        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1857    
1858    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1859        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1860        this option. Modified pcretest to add /W to test this facility. Added
1861        REG_UCP to make it available via the POSIX interface.
1862    
1863    10. Added --line-buffered to pcregrep.
1864    
1865    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1866        studied, and the match started with a letter with a code point greater than
1867        127 whose first byte was different to the first byte of the other case of
1868        the letter, the other case of this starting letter was not recognized
1869        (#976).
1870    
1871    12. If a pattern that was studied started with a repeated Unicode property
1872        test, for example, \p{Nd}+, there was the theoretical possibility of
1873        setting up an incorrect bitmap of starting bytes, but fortunately it could
1874        not have actually happened in practice until change 8 above was made (it
1875        added property types that matched character-matching opcodes).
1876    
1877    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1878        possible starting bytes for non-anchored patterns.
1879    
1880    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1881        \R, and also a number of cases that involve Unicode properties, both
1882        explicit and implicit when PCRE_UCP is set.
1883    
1884    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1885        input, it could crash or give wrong results if characters with values
1886        greater than 0xc0 were present in the subject string. (Detail: it assumed
1887        UTF-8 input when processing these items.)
1888    
1889    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1890        size_t is 64-bit (#991).
1891    
1892    17. Added a check for running out of memory when PCRE is compiled with
1893        --disable-stack-for-recursion (#990).
1894    
1895    18. If the last data line in a file for pcretest does not have a newline on
1896        the end, a newline was missing in the output.
1897    
1898    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1899        less than 128) in its various bitmaps. However, there is a facility for
1900        generating tables according to the current locale when PCRE is compiled. It
1901        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1902        characters, are recognized by isspace() and therefore were getting set in
1903        these tables, and indeed these tables seem to approximate to ISO 8859. This
1904        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1905        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1906        which of course cannot start UTF-8 characters. I have changed the code so
1907        that only real ASCII characters (less than 128) and the correct starting
1908        bytes for UTF-8 encodings are set for characters greater than 127 when in
1909        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1910        altogether.)
1911    
1912    20. Added the /T option to pcretest so as to be able to run tests with non-
1913        standard character tables, thus making it possible to include the tests
1914        used for 19 above in the standard set of tests.
1915    
1916    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1917        reference to a subpattern the other side of a comment that contains an
1918        opening parenthesis caused either an internal compiling error, or a
1919        reference to the wrong subpattern.
1920    
1921    
1922    Version 8.02 19-Mar-2010
1923    ------------------------
1924    
1925    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1926    
1927    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1928        configured.
1929    
1930    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1931        original author of that file, following a query about its status.
1932    
1933    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1934        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1935    
1936    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1937        quantifier applied to a forward-referencing subroutine call, could compile
1938        incorrect code or give the error "internal error: previously-checked
1939        referenced subpattern not found".
1940    
1941    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1942        variables to point to external functions. For these systems, therefore,
1943        pcre_malloc etc. are now initialized to local functions that call the
1944        relevant global functions.
1945    
1946    7.  There were two entries missing in the vectors called coptable and poptable
1947        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1948        I've fixed the data, and added a kludgy way of testing at compile time that
1949        the lengths are correct (equal to the number of opcodes).
1950    
1951    8.  Following on from 7, I added a similar kludge to check the length of the
1952        eint vector in pcreposix.c.
1953    
1954    9.  Error texts for pcre_compile() are held as one long string to avoid too
1955        much relocation at load time. To find a text, the string is searched,
1956        counting zeros. There was no check for running off the end of the string,
1957        which could happen if a new error number was added without updating the
1958        string.
1959    
1960    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1961    
1962    11. \K was not working if it appeared in an atomic group or in a group that
1963        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1964        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1965        the assertion is positive, but not if it is negative.
1966    
1967    12. Change 11 fortuitously reduced the size of the stack frame used in the
1968        "match()" function of pcre_exec.c by one pointer. Forthcoming
1969        implementation of support for (*MARK) will need an extra pointer on the
1970        stack; I have reserved it now, so that the stack frame size does not
1971        decrease.
1972    
1973    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1974        item in branch that calls a recursion is a subroutine call - as in the
1975        second branch in the above example - was incorrectly given the compile-
1976        time error "recursive call could loop indefinitely" because pcre_compile()
1977        was not correctly checking the subroutine for matching a non-empty string.
1978    
1979    14. The checks for overrunning compiling workspace could trigger after an
1980        overrun had occurred. This is a "should never occur" error, but it can be
1981        triggered by pathological patterns such as hundreds of nested parentheses.
1982        The checks now trigger 100 bytes before the end of the workspace.
1983    
1984    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1985    
1986    
1987    Version 8.01 19-Jan-2010
1988    ------------------------
1989    
1990    1.  If a pattern contained a conditional subpattern with only one branch (in
1991        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1992        computed the wrong minimum data length (which is of course zero for such
1993        subpatterns). This could cause incorrect "no match" results.
1994    
1995    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1996        the pattern is reset in the first branch, pcre_compile() failed with
1997        "internal error: code overflow at offset...". This happened only when
1998        the reset was to the original external option setting. (An optimization
1999        abstracts leading options settings into an external setting, which was the
2000        cause of this.)
2001    
2002    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2003        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2004        assertion pattern did not match (meaning that the assertion was true), it
2005        was incorrectly treated as false if the SKIP had been reached during the
2006        matching. This also applied to assertions used as conditions.
2007    
2008    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2009        assertion subpattern, including such a pattern used as a condition,
2010        unpredictable results occurred, instead of the error return
2011        PCRE_ERROR_DFA_UITEM.
2012    
2013    5.  The C++ GlobalReplace function was not working like Perl for the special
2014        situation when an empty string is matched. It now does the fancy magic
2015        stuff that is necessary.
2016    
2017    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2018        removed. (These were left over from very, very early versions of PCRE.)
2019    
2020    7.  Some cosmetic changes to the code to make life easier when compiling it
2021        as part of something else:
2022    
2023        (a) Change DEBUG to PCRE_DEBUG.
2024    
2025        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2026            called "current" as "current_branch", to prevent a collision with the
2027            Linux macro when compiled as a kernel module.
2028    
2029        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2030            prevent a collision with the Linux macro when compiled as a kernel
2031            module.
2032    
2033    8.  In pcre_compile() there are some checks for integer overflows that used to
2034        cast potentially large values to (double). This has been changed to that
2035        when building, a check for int64_t is made, and if it is found, it is used
2036        instead, thus avoiding the use of floating point arithmetic. (There is no
2037        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2038        double.
2039    
2040    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2041        2005 (difference between two addresses compared to an unsigned value).
2042    
2043    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2044        custom one, because of the following reported problem in Windows:
2045    
2046          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2047              under Win32.
2048          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2049              therefore missing the function definition.
2050          - The compiler thus generates a "C" signature for the test function.
2051          - The linker fails to find the "C" function.
2052          - PCRE fails to configure if asked to do so against libbz2.
2053    
2054    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2055        messages were output:
2056    
2057          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2058          rerunning libtoolize, to keep the correct libtool macros in-tree.
2059          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2060    
2061        I have done both of these things.
2062    
2063    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2064        most of the time, it *can* run out if it is given a pattern that contains a
2065        runaway infinite recursion. I updated the discussion in the pcrestack man
2066        page.
2067    
2068    13. Now that we have gone to the x.xx style of version numbers, the minor
2069        version may start with zero. Using 08 or 09 is a bad idea because users
2070        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2071        interpreted as invalid octal numbers. I've updated the previous comment in
2072        configure.ac, and also added a check that gives an error if 08 or 09 are
2073        used.
2074    
2075    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2076        causing partial matching to fail when the end of the subject matched \W
2077        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2078    
2079    15. There were some discrepancies between the declarations in pcre_internal.h
2080        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2081        their definitions. The declarations used "const uschar *" and the
2082        definitions used USPTR. Even though USPTR is normally defined as "const
2083        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2084        reported that: "This difference in casting confuses some C++ compilers, for
2085        example, SunCC recognizes above declarations as different functions and
2086        generates broken code for hbpcre." I have changed the declarations to use
2087        USPTR.
2088    
2089    16. GNU libtool is named differently on some systems. The autogen.sh script now
2090        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2091        (FreeBSD).
2092    
2093    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2094        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2095        comment: "Figure out how to create a longlong from a string: strtoll and
2096        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2097        instance, but it only takes 2 args instead of 3!"
2098    
2099    18. A subtle bug concerned with back references has been fixed by a change of
2100        specification, with a corresponding code fix. A pattern such as
2101        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2102        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2103        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2104        same bug. Such groups have to be quantified to be useful, or contained
2105        inside another quantified group. (If there's no repetition, the reference
2106        can never match.) The problem arises because, having left the group and
2107        moved on to the rest of the pattern, a later failure that backtracks into
2108        the group uses the captured value from the final iteration of the group
2109        rather than the correct earlier one. I have fixed this in PCRE by forcing
2110        any group that contains a reference to itself to be an atomic group; that
2111        is, there cannot be any backtracking into it once it has completed. This is
2112        similar to recursive and subroutine calls.
2113    
2114    
2115    Version 8.00 19-Oct-09
2116    ----------------------
2117    
2118    1.  The table for translating pcre_compile() error codes into POSIX error codes
2119        was out-of-date, and there was no check on the pcre_compile() error code
2120        being within the table. This could lead to an OK return being given in
2121        error.
2122    
2123    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2124        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2125        in a Windows environment.
2126    
2127    3.  The pcregrep --count option prints the count for each file even when it is
2128        zero, as does GNU grep. However, pcregrep was also printing all files when
2129        --files-with-matches was added. Now, when both options are given, it prints
2130        counts only for those files that have at least one match. (GNU grep just
2131        prints the file name in this circumstance, but including the count seems
2132        more useful - otherwise, why use --count?) Also ensured that the
2133        combination -clh just lists non-zero counts, with no names.
2134    
2135    4.  The long form of the pcregrep -F option was incorrectly implemented as
2136        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2137        but it seems right to fix it, and I didn't think it was worth preserving
2138        the old behaviour.
2139    
2140    5.  The command line items --regex=pattern and --regexp=pattern were not
2141        recognized by pcregrep, which required --regex pattern or --regexp pattern
2142        (with a space rather than an '='). The man page documented the '=' forms,
2143        which are compatible with GNU grep; these now work.
2144    
2145    6.  No libpcreposix.pc file was created for pkg-config; there was just
2146        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2147    
2148    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2149        when UCP support is not needed, by modifying the Python script that
2150        generates it from Unicode data files. This should not matter if the module
2151        is correctly used as a library, but I received one complaint about 50K of
2152        unwanted data. My guess is that the person linked everything into his
2153        program rather than using a library. Anyway, it does no harm.
2154    
2155    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2156        was a minimum greater than 1 for a wide character in a possessive
2157        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2158        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2159        character. Chaos in the form of incorrect output or a compiling loop could
2160        result.
2161    
2162    9.  The restrictions on what a pattern can contain when partial matching is
2163        requested for pcre_exec() have been removed. All patterns can now be
2164        partially matched by this function. In addition, if there are at least two
2165        slots in the offset vector, the offset of the earliest inspected character
2166        for the match and the offset of the end of the subject are set in them when
2167        PCRE_ERROR_PARTIAL is returned.
2168    
2169    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2170        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2171        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2172        and may be more useful for multi-segment matching.
2173    
2174    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2175        used to be given if ever the end of the subject was reached; now it is
2176        given only if matching could not proceed because another character was
2177        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2178        string "Z", which now yields "no match" instead of "partial match". In the
2179        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2180        final character ended with (*FAIL).
2181    
2182    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2183        if the pattern had a "must contain" character that was already found in the
2184        earlier partial match, unless partial matching was again requested. For
2185        example, with the pattern /dog.(body)?/, the "must contain" character is
2186        "g". If the first part-match was for the string "dog", restarting with
2187        "sbody" failed. This bug has been fixed.
2188    
2189    13. The string returned by pcre_dfa_exec() after a partial match has been
2190        changed so that it starts at the first inspected character rather than the
2191        first character of the match. This makes a difference only if the pattern
2192        starts with a lookbehind assertion or \b or \B (\K is not supported by
2193        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2194        matching functions compatible, and I think it's the right thing to do.
2195    
2196    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2197        so that the demonstration program is easily available in environments where
2198        PCRE has not been installed from source.
2199    
2200    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2201        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2202        library.
2203    
2204    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2205        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2206        is not the first non-POSIX option to be added. Clearly some people find
2207        these options useful.
2208    
2209    17. If a caller to the POSIX matching function regexec() passes a non-zero
2210        value for nmatch with a NULL value for pmatch, the value of
2211        nmatch is forced to zero.
2212    
2213    18. RunGrepTest did not have a test for the availability of the -u option of
2214        the diff command, as RunTest does. It now checks in the same way as
2215        RunTest, and also checks for the -b option.
2216    
2217    19. If an odd number of negated classes containing just a single character
2218        interposed, within parentheses, between a forward reference to a named
2219        subpattern and the definition of the subpattern, compilation crashed with
2220        an internal error, complaining that it could not find the referenced
2221        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2222        [The bug was that it was starting one character too far in when skipping
2223        over the character class, thus treating the ] as data rather than
2224        terminating the class. This meant it could skip too much.]
2225    
2226    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2227        /g option in pcretest when the pattern contains \K, which makes it possible
2228        to have an empty string match not at the start, even when the pattern is
2229        anchored. Updated pcretest and pcredemo to use this option.
2230    
2231    21. If the maximum number of capturing subpatterns in a recursion was greater
2232        than the maximum at the outer level, the higher number was returned, but
2233        with unset values at the outer level. The correct (outer level) value is
2234        now given.
2235    
2236    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2237        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2238        make it do so. The string so far is captured, making this feature
2239        compatible with Perl.
2240    
2241    23. The tests have been re-organized, adding tests 11 and 12, to make it
2242        possible to check the Perl 5.10 features against Perl 5.10.
2243    
2244    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2245        pattern matches a fixed length string. PCRE did not allow this; now it
2246        does. Neither allows recursion.
2247    
2248    25. I finally figured out how to implement a request to provide the minimum
2249        length of subject string that was needed in order to match a given pattern.
2250        (It was back references and recursion that I had previously got hung up
2251        on.) This code has now been added to pcre_study(); it finds a lower bound
2252        to the length of subject needed. It is not necessarily the greatest lower
2253        bound, but using it to avoid searching strings that are too short does give
2254        some useful speed-ups. The value is available to calling programs via
2255        pcre_fullinfo().
2256    
2257    26. While implementing 25, I discovered to my embarrassment that pcretest had
2258        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2259        study optimizations had never been tested with that matching function.
2260        Oops. What is worse, even when it was passed study data, there was a bug in
2261        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2262        were also very few tests of studied patterns with pcre_dfa_exec().
2263    
2264    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2265        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2266        on the other side of the coin, they are no longer allowed to have different
2267        names, because these cannot be distinguished in PCRE, and this has caused
2268        confusion. (This is a difference from Perl.)
2269    
2270    28. When duplicate subpattern names are present (necessarily with different
2271        numbers, as required by 27 above), and a test is made by name in a
2272        conditional pattern, either for a subpattern having been matched, or for
2273        recursion in such a pattern, all the associated numbered subpatterns are
2274        tested, and the overall condition is true if the condition is true for any
2275        one of them. This is the way Perl works, and is also more like the way
2276        testing by number works.
2277    
2278    
2279    Version 7.9 11-Apr-09
2280    ---------------------
2281    
2282    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2283        (pcretest), all targets were linked against these libraries. This included
2284        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2285        libraries. This caused unwanted dependencies to be created. This problem
2286        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2287        pcretest is linked with readline.
2288    
2289    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2290        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2291        moved outside it again, because FALSE and TRUE are already defined in AIX,
2292        but BOOL is not.
2293    
2294    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2295        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2296    
2297    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2298        hyphens) following file names and line numbers when outputting matching
2299        lines. This is not true; no spaces are inserted. I have also clarified the
2300        wording for the --colour (or --color) option.
2301    
2302    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2303        was not coloured; this is different to GNU grep, so I have changed it to be
2304        the same.
2305    
2306    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2307        each matching line was coloured. Now it goes on to look for further matches
2308        of any of the test patterns, which is the same behaviour as GNU grep.
2309    
2310    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2311        doesn't make sense to accept an empty string match in pcregrep, so I have
2312        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2313        seems to be how GNU grep behaves. [But see later change 40 for release
2314        8.33.]
2315    
2316    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2317        start or after a newline", because the conditional assertion was not being
2318        correctly handled. The rule now is that both the assertion and what follows
2319        in the first alternative must satisfy the test.
2320    
2321    9.  If auto-callout was enabled in a pattern with a conditional group whose
2322        condition was an assertion, PCRE could crash during matching, both with
2323        pcre_exec() and pcre_dfa_exec().
2324    
2325    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2326        used for matching.
2327    
2328    11. Unicode property support in character classes was not working for
2329        characters (bytes) greater than 127 when not in UTF-8 mode.
2330    
2331    12. Added the -M command line option to pcretest.
2332    
2333    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2334    
2335    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2336    
2337    16. Added comments and documentation about mis-use of no_arg in the C++
2338        wrapper.
2339    
2340    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2341        from Martin Jerabek that uses macro names for all relevant character and
2342        string constants.
2343    
2344    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2345        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2346        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2347        these, but not everybody uses configure.
2348    
2349    19. A conditional group that had only one branch was not being correctly
2350        recognized as an item that could match an empty string. This meant that an
2351        enclosing group might also not be so recognized, causing infinite looping
2352        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2353        with the subject "ab", where knowledge that the repeated group can match
2354        nothing is needed in order to break the loop.
2355    
2356    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2357        exec(), but without supplying a callout function, matching went wrong.
2358    
2359    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2360        leak if the size of the offset vector was greater than 30. When the vector
2361        is smaller, the saved offsets during recursion go onto a local stack
2362        vector, but for larger vectors malloc() is used. It was failing to free
2363        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2364        error, in fact).
2365    
2366    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2367        heapframe that is used only when UTF-8 support is enabled. This caused no
2368        problem, but was untidy.
2369    
2370    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2371        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2372        included within another project.
2373    
2374    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2375        slightly modified by me:
2376    
2377          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2378              not building pcregrep.
2379    
2380          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2381              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2382    
2383    25. Forward references, both numeric and by name, in patterns that made use of
2384        duplicate group numbers, could behave incorrectly or give incorrect errors,
2385        because when scanning forward to find the reference group, PCRE was not
2386        taking into account the duplicate group numbers. A pattern such as
2387        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2388    
2389    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2390        the feature of a custom pointer more persuasive (as requested by a user).
2391    
2392    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2393        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2394        pre-defined.
2395    
2396    28. Added support for (*UTF8) at the start of a pattern.
2397    
2398    29. Arrange for flags added by the "release type" setting in CMake to be shown
2399        in the configuration summary.
2400    
2401    
2402    Version 7.8 05-Sep-08
2403    ---------------------
2404    
2405    1.  Replaced UCP searching code with optimized version as implemented for Ad
2406        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2407        stage table and inline lookup instead of a function, giving speed ups of 2
2408        to 5 times on some simple patterns that I tested. Permission was given to
2409        distribute the MultiStage2.py script that generates the tables (it's not in
2410        the tarball, but is in the Subversion repository).
2411    
2412    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2413        scripts.
2414    
2415    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2416        a group with a zero qualifier. The result of the study could be incorrect,
2417        or the function might crash, depending on the pattern.
2418    
2419    4.  Caseless matching was not working for non-ASCII characters in back
2420        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2421        It now works when Unicode Property Support is available.
2422    
2423    5.  In pcretest, an escape such as \x{de} in the data was always generating
2424        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2425        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2426        truncation.
2427    
2428    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2429    
2430    7.  Added two (int) casts to pcregrep when printing the difference of two
2431        pointers, in case they are 64-bit values.
2432    
2433    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2434        test 2 if it fails.
2435    
2436    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2437        and a #define of that name to empty if it is not externally set. This is to
2438        allow users of MSVC to set it if necessary.
2439    
2440    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2441        the convenience functions in the pcre_get.c source file.
2442    
2443    11. An option change at the start of a pattern that had top-level alternatives
2444        could cause overwriting and/or a crash. This command provoked a crash in
2445        some environments:
2446    
2447          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2448    
2449        This potential security problem was recorded as CVE-2008-2371.
2450    
2451    12. For a pattern where the match had to start at the beginning or immediately
2452        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2453        pcre_dfa_exec() could read past the end of the passed subject if there was
2454        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2455        pcretest so that it places the subject at the end of its malloc-ed buffer.
2456    
2457    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2458        exec() might read past the end of the data buffer in UTF-8 mode.
2459    
2460    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2461        the data contained the byte 0x85 as part of a UTF-8 character within its
2462        first line. This applied both to normal and DFA matching.
2463    
2464    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2465        /^[^d]*?$/8 failed to match "abc".
2466    
2467    16. Added a missing copyright notice to pcrecpp_internal.h.
2468    
2469    17. Make it more clear in the documentation that values returned from
2470        pcre_exec() in ovector are byte offsets, not character counts.
2471    
2472    18. Tidied a few places to stop certain compilers from issuing warnings.
2473    
2474    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2475        supplied by Stefan Weber. I made a further small update for 7.8 because
2476        there is a change of source arrangements: the pcre_searchfuncs.c module is
2477        replaced by pcre_ucd.c.
2478    
2479    
2480    Version 7.7 07-May-08
2481    ---------------------
2482    
2483    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2484        a string to a long long, pretend we don't even have a long long." This is
2485        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2486    
2487    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2488        pre-7.6 versions, which defined a global no_arg variable instead of putting
2489        it in the RE class. (See also #8 below.)
2490    
2491    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2492        Lopes.
2493    
2494    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2495    
2496        (1) The include/exclude patterns were being applied to the whole pathnames
2497            of files, instead of just to the final components.
2498    
2499        (2) If there was more than one level of directory, the subdirectories were
2500            skipped unless they satisfied the include/exclude conditions. This is
2501            inconsistent with GNU grep (and could even be seen as contrary to the
2502            pcregrep specification - which I improved to make it absolutely clear).
2503            The action now is always to scan all levels of directory, and just
2504            apply the include/exclude patterns to regular files.
2505    
2506    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2507        --exclude_dir in the tests to avoid scanning .svn directories.
2508    
2509    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2510        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2511        doesn't support NULs in patterns.
2512    
2513    7.  Added some missing "const"s to declarations of static tables in
2514        pcre_compile.c and pcre_dfa_exec.c.
2515    
2516    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2517        caused by fix #2  above. (Subsequently also a second patch to fix the
2518        first patch. And a third patch - this was a messy problem.)
2519    
2520    9.  Applied Craig's patch to remove the use of push_back().
2521    
2522    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2523        matching function regexec().
2524    
2525    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2526        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2527        references. PCRE supports relative numbers with this syntax (I don't think
2528        Oniguruma does).
2529    
2530    12. Previously, a group with a zero repeat such as (...){0} was completely
2531        omitted from the compiled regex. However, this means that if the group
2532        was called as a subroutine from elsewhere in the pattern, things went wrong
2533        (an internal error was given). Such groups are now left in the compiled
2534        pattern, with a new opcode that causes them to be skipped at execution
2535        time.
2536    
2537    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2538        to the way PCRE behaves:
2539    
2540        (a) A lone ] character is dis-allowed (Perl treats it as data).
2541    
2542        (b) A back reference to an unmatched subpattern matches an empty string
2543            (Perl fails the current match path).
2544    
2545        (c) A data ] in a character class must be notated as \] because if the
2546            first data character in a class is ], it defines an empty class. (In
2547            Perl it is not possible to have an empty class.) The empty class []
2548            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2549            The negative empty class [^] matches any one character, independently
2550            of the DOTALL setting.
2551    
2552    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2553        non-existent subpattern following a character class starting with ']' and
2554        containing () gave an internal compiling error instead of "reference to
2555        non-existent subpattern". Fortunately, when the pattern did exist, the
2556        compiled code was correct. (When scanning forwards to check for the
2557        existence of the subpattern, it was treating the data ']' as terminating
2558        the class, so got the count wrong. When actually compiling, the reference
2559        was subsequently set up correctly.)
2560    
2561    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2562        it was being rejected as not supported by pcre_dfa_exec(), even though
2563        other assertions are supported. I have made pcre_dfa_exec() support
2564        (*FAIL).
2565    
2566    16. The implementation of 13c above involved the invention of a new opcode,
2567        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2568        cannot be changed at match time, I realized I could make a small
2569        improvement to matching performance by compiling OP_ALLANY instead of
2570        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2571        on the OP_ANY path.
2572    
2573    17. Compiling pcretest on Windows with readline support failed without the
2574        following two fixes: (1) Make the unistd.h include conditional on
2575        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2576    
2577    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2578        ncurses library to be included for pcretest when ReadLine support is
2579        requested, but also to allow for it to be overridden. This patch came from
2580        Daniel Bergström.
2581    
2582    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2583        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2584        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2585        spotting this.
2586    
2587    
2588    Version 7.6 28-Jan-08
2589    ---------------------
2590    
2591    1.  A character class containing a very large number of characters with
2592        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2593        overflow.
2594    
2595    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2596        HAVE_LONG_LONG is not defined.
2597    
2598    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2599        bring them up to date and include new features. This patch includes:
2600    
2601        - Fixed PH's badly added libz and libbz2 support.
2602        - Fixed a problem with static linking.
2603        - Added pcredemo. [But later removed - see 7 below.]
2604        - Fixed dftables problem and added an option.
2605        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2606            HAVE_LONG_LONG.
2607        - Added readline support for pcretest.
2608        - Added an listing of the option settings after cmake has run.
2609    
2610    4.  A user submitted a patch to Makefile that makes it easy to create
2611        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2612        Makefile.am that cause it to include this special target, without
2613        affecting anything else. Note that the same mingw target plus all
2614        the other distribution libraries and programs are now supported
2615        when configuring with CMake (see 6 below) instead of with
2616        Configure/Make.
2617    
2618    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2619        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2620        exported in the Windows port". It has not yet been confirmed that the patch
2621        solves the problem, but it does no harm.
2622    
2623    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2624        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2625        with CMake, and also correct the comment about stack recursion.
2626    
2627    7.  Remove the automatic building of pcredemo from the ./configure system and
2628        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2629        of a program that users should build themselves after PCRE is installed, so
2630        building it automatically is not really right. What is more, it gave
2631        trouble in some build environments.
2632    
2633    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2634    
2635    
2636    Version 7.5 10-Jan-08
2637    ---------------------
2638    
2639    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2640        values in parens when parsing an RE using the C++ wrapper."
2641    
2642    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2643        Characters greater than 255 were excluded from the class instead of being
2644        included.
2645    
2646    3.  The same bug as (2) above applied to negated POSIX classes such as
2647        [:^space:].
2648    
2649    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2650        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2651        I have changed it.
2652    
2653    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2654        first named subpattern) and a construct such as (?&a) would reference the
2655        first named subpattern whose name started with "a" (in other words, the
2656        length check was missing). Both these problems are fixed. "Subpattern name
2657        expected" is now given for (?&) (a zero-length name), and this patch also
2658        makes it give the same error for \k'' (previously it complained that that
2659        was a reference to a non-existent subpattern).
2660    
2661    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2662        this is right because (?- can be followed by option settings as well as by
2663        digits. I have, however, made the messages clearer.
2664    
2665    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2666        than the number used in the conditional) now cause a compile-time error.
2667        This is actually not compatible with Perl, which accepts such patterns, but
2668        treats the conditional as always being FALSE (as PCRE used to), but it
2669        seems to me that giving a diagnostic is better.
2670    
2671    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2672        and messages.
2673    
2674    9.  Fix two occurrences of "backslash" in comments that should have been
2675        "backspace".
2676    
2677    10. Remove two redundant lines of code that can never be obeyed (their function
2678        was moved elsewhere).
2679    
2680    11. The program that makes PCRE's Unicode character property table had a bug
2681        which caused it to generate incorrect table entries for sequences of
2682        characters that have the same character type, but are in different scripts.
2683        It amalgamated them into a single range, with the script of the first of
2684        them. In other words, some characters were in the wrong script. There were
2685        thirteen such cases, affecting characters in the following ranges:
2686    
2687          U+002b0 - U+002c1
2688          U+0060c - U+0060d
2689          U+0061e - U+00612
2690          U+0064b - U+0065e
2691          U+0074d - U+0076d
2692          U+01800 - U+01805
2693          U+01d00 - U+01d77
2694          U+01d9b - U+01dbf
2695          U+0200b - U+0200f
2696          U+030fc - U+030fe
2697          U+03260 - U+0327f
2698          U+0fb46 - U+0fbb1
2699          U+10450 - U+1049d
2700    
2701    12. The -o option (show only the matching part of a line) for pcregrep was not
2702        compatible with GNU grep in that, if there was more than one match in a
2703        line, it showed only the first of them. It now behaves in the same way as
2704        GNU grep.
2705    
2706    13. If the -o and -v options were combined for pcregrep, it printed a blank
2707        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2708        does the same. The return code can be used to tell if there were any
2709        non-matching lines.
2710    
2711    14. Added --file-offsets and --line-offsets to pcregrep.
2712    
2713    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2714        infinitely looping recursion. The bug was that positive lookaheads were not
2715        being skipped when checking for a possible empty match (negative lookaheads
2716        and both kinds of lookbehind were skipped).
2717    
2718    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2719        inclusion of <windows.h> to before rather than after the definition of
2720        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2721    
2722    17. Specifying a possessive quantifier with a specific limit for a Unicode
2723        character property caused pcre_compile() to compile bad code, which led at
2724        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2725        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2726        caused the error; without that there was no problem.
2727    
2728    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2729    
2730    19. Added --enable-pcretest-libreadline.
2731    
2732    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2733        RE::GlobalReplace(). As a result, the number of replacements returned was
2734        double what it should be. I removed one of the increments, but Craig sent a
2735        later patch that removed the other one (the right fix) and added unit tests
2736        that check the return values (which was not done before).
2737    
2738    21. Several CMake things:
2739    
2740        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2741            the names libpcre and libpcreposix, not just pcre and pcreposix.
2742    
2743        (2) The above change means that pcretest and pcregrep are now correctly
2744            linked with the newly-built libraries, not previously installed ones.
2745    
2746        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2747    
2748    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2749        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2750        UTF-8 newline character). The key issue is that the pattern starts .*;
2751        this means that the match must be either at the beginning, or after a
2752        newline. The bug was in the code for advancing after a failed match and
2753        checking that the new position followed a newline. It was not taking
2754        account of UTF-8 characters correctly.
2755    
2756    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2757        character classes. PCRE was not treating the sequence [:...:] as a
2758        character class unless the ... were all letters. Perl, however, seems to
2759        allow any characters between [: and :], though of course it rejects as
2760        unknown any "names" that contain non-letters, because all the known class
2761        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2762        for example, whereas PCRE did not - it did not recognize a POSIX character
2763        class. This seemed a bit dangerous, so the code has been changed to be
2764        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2765        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2766        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2767        Perl does, and where it didn't before.
2768    
2769    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2770        Windows environments %n is disabled by default.
2771    
2772    
2773    Version 7.4 21-Sep-07
2774    ---------------------
2775    
2776    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2777        means that a class such as [\s] counted as "explicit reference to CR or
2778        LF". That isn't really right - the whole point of the change was to try to
2779        help when there was an actual mention of one of the two characters. So now
2780        the change happens only if \r or \n (or a literal CR or LF) character is
2781        encountered.
2782    
2783    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2784        of both had grown to the point where there were only 3 bits left.
2785        Fortunately, there was spare space in the data structure, and so I have
2786        moved the internal flags into a new 16-bit field to free up more option
2787        bits.
2788    
2789    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2790        but did not set the internal JCHANGED flag - either of these is enough to
2791        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2792        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2793        start sets both bits.
2794    
2795    4.  Added options (at build time, compile time, exec time) to change \R from
2796        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2797    
2798    5.  doc/pcresyntax.html was missing from the distribution.
2799    
2800    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2801        compatibility, even though it is no longer used.
2802    
2803    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2804        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2805        windows.h file is present (where different names are used). [This was
2806        reversed later after testing - see 16 below.]
2807    
2808    8.  Changed all #include <config.h> to #include "config.h". There were also
2809        some further <pcre.h> cases that I changed to "pcre.h".
2810    
2811    9.  When pcregrep was used with the --colour option, it missed the line ending
2812        sequence off the lines that it output.
2813    
2814    10. It was pointed out to me that arrays of string pointers cause lots of
2815        relocations when a shared library is dynamically loaded. A technique of
2816        using a single long string with a table of offsets can drastically reduce
2817        these. I have refactored PCRE in four places to do this. The result is
2818        dramatic:
2819    
2820          Originally:                          290
2821          After changing UCP table:            187
2822          After changing error message table:   43
2823          After changing table of "verbs"       36
2824          After changing table of Posix names   22
2825    
2826        Thanks to the folks working on Gregex for glib for this insight.
2827    
2828    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2829        unicode-properties was also set.
2830    
2831    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2832    
2833    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2834        checked only for CRLF.
2835    
2836    14. Added casts to pcretest.c to avoid compiler warnings.
2837    
2838    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2839    
2840    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2841        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2842        entirely. This removes changes made in 7 above.
2843    
2844    17. The CMake files have been updated, and there is now more information about
2845        building with CMake in the NON-UNIX-USE document.
2846    
2847    
2848    Version 7.3 28-Aug-07
2849  ---------------------  ---------------------
2850    
2851   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 2873  Version 7.3 05-Jul-07
2873   4. Debugging: adjusted the names of opcodes for different kinds of parentheses   4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2874      in debug output.      in debug output.
2875    
2876   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
2877      unittest when running under MinGW.      long printing in the pcrecpp unittest when running under MinGW.
2878    
2879   6. ESC_K was left out of the EBCDIC table.   6. ESC_K was left out of the EBCDIC table.
2880    
# Line 53  Version 7.3 05-Jul-07 Line 2897  Version 7.3 05-Jul-07
2897      pcreposix and pcrecpp parts of the library. These were overlooked when this      pcreposix and pcrecpp parts of the library. These were overlooked when this
2898      problem was solved for the main library.      problem was solved for the main library.
2899    
2900     9. There were some crude static tests to avoid integer overflow when computing
2901        the size of patterns that contain repeated groups with explicit upper
2902        limits. As the maximum quantifier is 65535, the maximum group length was
2903        set at 30,000 so that the product of these two numbers did not overflow a
2904        32-bit integer. However, it turns out that people want to use groups that
2905        are longer than 30,000 bytes (though not repeat them that many times).
2906        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2907        made it possible to implement the integer overflow checks in a much more
2908        dynamic way, which I have now done. The artificial limitation on group
2909        length has been removed - we now have only the limit on the total length of
2910        the compiled pattern, which depends on the LINK_SIZE setting.
2911    
2912    10. Fixed a bug in the documentation for get/copy named substring when
2913        duplicate names are permitted. If none of the named substrings are set, the
2914        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2915        empty string.
2916    
2917    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2918        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2919        because the ] is interpreted as the first data character and the
2920        terminating ] is not found. PCRE has been made compatible with Perl in this
2921        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2922        cause memory overwriting.
2923    
2924    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2925        string has been matched (to stop an infinite loop). It was not recognizing
2926        a conditional subpattern that could match an empty string if that
2927        subpattern was within another subpattern. For example, it looped when
2928        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2929        condition was not nested. This bug has been fixed.
2930    
2931    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2932        past the start of the subject in the presence of bytes with the top bit
2933        set, for example "\x8aBCD".
2934    
2935    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2936        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2937    
2938    14. Optimized (?!) to (*FAIL).
2939    
2940    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2941        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2942        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2943        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2944        does: it's just the validity check that is more restrictive.
2945    
2946    16. Inserted checks for integer overflows during escape sequence (backslash)
2947        processing, and also fixed erroneous offset values for syntax errors during
2948        backslash processing.
2949    
2950    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2951        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2952    
2953    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2954        caused an overrun.
2955    
2956    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2957        something other than just ASCII characters) inside a group that had an
2958        unlimited repeat caused a loop at compile time (while checking to see
2959        whether the group could match an empty string).
2960    
2961    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2962        [\P{Any}] did so. (Error in the code for printing property names.)
2963    
2964    21. An orphan \E inside a character class could cause a crash.
2965    
2966    22. A repeated capturing bracket such as (A)? could cause a wild memory
2967        reference during compilation.
2968    
2969    23. There are several functions in pcre_compile() that scan along a compiled
2970        expression for various reasons (e.g. to see if it's fixed length for look
2971        behind). There were bugs in these functions when a repeated \p or \P was
2972        present in the pattern. These operators have additional parameters compared
2973        with \d, etc, and these were not being taken into account when moving along
2974        the compiled data. Specifically:
2975    
2976        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2977            length.
2978    
2979        (b) An item such as \pL+ within a repeated group could cause crashes or
2980            loops.
2981    
2982        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2983            "reference to non-existent subpattern" error.
2984    
2985        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2986    
2987    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2988        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2989    
2990    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2991    
2992    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2993        character were causing crashes (broken optimization).
2994    
2995    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2996        \p or \P) caused a compile-time loop.
2997    
2998    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2999        break. For example, the unstudied pattern [\r\n]A does not match the string
3000        "\r\nA" because change 7.0/46 below moves the current point on by two
3001        characters after failing to match at the start. However, the pattern \nA
3002        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3003        the same is true. There doesn't seem any very clean way out of this, but
3004        what I have chosen to do makes the common cases work: PCRE now takes note
3005        of whether there can be an explicit match for \r or \n anywhere in the
3006        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3007        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3008        pattern has explicit CR or LF references.
3009    
3010    29. Added (*CR) etc for changing newline setting at start of pattern.
3011    
3012    
3013  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3014  ---------------------  ---------------------

Legend:
Removed from v.200  
changed lines
  Added in v.1563

  ViewVC Help
Powered by ViewVC 1.1.5