ViewVC logotype

Diff of /code/trunk/ChangeLog

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

revision 740 by zherczeg, Mon Oct 31 06:10:14 2011 UTC revision 1662 by zherczeg, Fri Jul 15 09:57:53 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4  Version 8.21  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.
7    Version 8.40 17-June-2016
8    -------------------------
10    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
11        the match extended over a line boundary.
13    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
14        MSVC static compilation, putting the first patch under a new option.
16    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
18    4.  Ignore "show all captures" (/=) for DFA matching.
20    5.  Fix unaligned accesses on x86. Patch by Marc Mutz.
23    Version 8.39 14-June-2016
24    -------------------------
26    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
27        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
28        misbehaved. This bug was found by the LLVM fuzzer.
30    2.  Similar to the above, if an isolated \E was present between an item and its
31        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
32        bug was found by the LLVM fuzzer.
34    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
35        working correctly in UCP mode.
37    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
38        was set when the pmatch argument was NULL. It now returns REG_INVARG.
40    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
42    6.  An empty \Q\E sequence between an item and its qualifier caused
43        pcre_compile() to misbehave when auto callouts were enabled. This bug was
44        found by the LLVM fuzzer.
46    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
47        space or a #-type comment that was followed by (?-x), which turns off
48        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
49        pcre_compile() assumed that (?-x) applied to the whole pattern and
50        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
52    8.  A call of pcre_copy_named_substring() for a named substring whose number
53        was greater than the space in the ovector could cause a crash.
55    9.  Yet another buffer overflow bug involved duplicate named groups with a
56        group that reset capture numbers (compare 8.38/7 below). Once again, I have
57        just allowed for more memory, even if not needed. (A proper fix is
58        implemented in PCRE2, but it involves a lot of refactoring.)
60    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
61        start of the match to be earlier than the end.
63    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
65    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
66        assertion, caused pcretest to generate incorrect output, and also to read
67        uninitialized memory (detected by ASAN or valgrind).
69    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
70        nested set of parentheses of sufficient size caused an overflow of the
71        compiling workspace (which was diagnosed, but of course is not desirable).
73    14. And yet another buffer overflow bug involving duplicate named groups, this
74        time nested, with a nested back reference. Yet again, I have just allowed
75        for more memory, because anything more needs all the refactoring that has
76        been done for PCRE2. An example pattern that provoked this bug is:
77        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
78        registered as CVE-2016-1283.
80    15. pcretest went into a loop if global matching was requested with an ovector
81        size less than 2. It now gives an error message. This bug was found by
82        afl-fuzz.
84    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
85        ("assertion expected") when (?(?C) was not followed by an opening
86        parenthesis.
88    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
89        actually affect anything, by sheer luck.
91    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
92        static compilation.
94    19. Modified the RunTest script to incorporate a valgrind suppressions file so
95        that certain errors, provoked by the SSE2 instruction set when JIT is used,
96        are ignored.
98    20. A racing condition is fixed in JIT reported by Mozilla.
100    21. Minor code refactor to avoid "array subscript is below array bounds"
101        compiler warning.
103    22. Minor code refactor to avoid "left shift of negative number" warning.
105    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
106        UCP support.
108    24. Refactor to avoid compiler warnings in pcrecpp.cc.
110    25. Refactor to fix a typo in pcre_jit_test.c
112    26. Patch to support compiling pcrecpp.cc with Intel compiler.
115    Version 8.38 23-November-2015
116    -----------------------------
118    1.  If a group that contained a recursive back reference also contained a
119        forward reference subroutine call followed by a non-forward-reference
120        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
121        compile correct code, leading to undefined behaviour or an internally
122        detected error. This bug was discovered by the LLVM fuzzer.
124    2.  Quantification of certain items (e.g. atomic back references) could cause
125        incorrect code to be compiled when recursive forward references were
126        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
127        This bug was discovered by the LLVM fuzzer.
129    3.  A repeated conditional group whose condition was a reference by name caused
130        a buffer overflow if there was more than one group with the given name.
131        This bug was discovered by the LLVM fuzzer.
133    4.  A recursive back reference by name within a group that had the same name as
134        another group caused a buffer overflow. For example:
135        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
137    5.  A forward reference by name to a group whose number is the same as the
138        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
139        a buffer overflow at compile time. This bug was discovered by the LLVM
140        fuzzer.
142    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
143        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
145    7.  Another buffer overflow bug involved duplicate named groups with a
146        reference between their definition, with a group that reset capture
147        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
148        fixed by always allowing for more memory, even if not needed. (A proper fix
149        is implemented in PCRE2, but it involves more refactoring.)
151    8.  There was no check for integer overflow in subroutine calls such as (?123).
153    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
154        being treated as a literal 'l' instead of causing an error.
156    10. There was a buffer overflow if pcre_exec() was called with an ovector of
157        size 1. This bug was found by american fuzzy lop.
159    11. If a non-capturing group containing a conditional group that could match
160        an empty string was repeated, it was not identified as matching an empty
161        string itself. For example: /^(?:(?(1)x|)+)+$()/.
163    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
164        \a and \e in test subject lines.
166    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
167        instead of the EBCDIC value.
169    14. The handling of \c in an EBCDIC environment has been revised so that it is
170        now compatible with the specification in Perl's perlebcdic page.
172    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
173        ASCII/Unicode. This has now been added to the list of characters that are
174        recognized as white space in EBCDIC.
176    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
177        error (correctly) when used outside a class, but did not give an error
178        within a class.
180    17. \h within a class was incorrectly compiled in EBCDIC environments.
182    18. A pattern with an unmatched closing parenthesis that contained a backward
183        assertion which itself contained a forward reference caused buffer
184        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
186    19. JIT should return with error when the compiled pattern requires more stack
187        space than the maximum.
189    20. A possessively repeated conditional group that could match an empty string,
190        for example, /(?(R))*+/, was incorrectly compiled.
192    21. Fix infinite recursion in the JIT compiler when certain patterns such as
193        /(?:|a|){100}x/ are analysed.
195    22. Some patterns with character classes involving [: and \\ were incorrectly
196        compiled and could cause reading from uninitialized memory or an incorrect
197        error diagnosis.
199    23. Pathological patterns containing many nested occurrences of [: caused
200        pcre_compile() to run for a very long time.
202    24. A conditional group with only one branch has an implicit empty alternative
203        branch and must therefore be treated as potentially matching an empty
204        string.
206    25. If (?R was followed by - or + incorrect behaviour happened instead of a
207        diagnostic.
209    26. Arrange to give up on finding the minimum matching length for overly
210        complex patterns.
212    27. Similar to (4) above: in a pattern with duplicated named groups and an
213        occurrence of (?| it is possible for an apparently non-recursive back
214        reference to become recursive if a later named group with the relevant
215        number is encountered. This could lead to a buffer overflow. Wen Guanxing
216        from Venustech ADLAB discovered this bug.
218    28. If pcregrep was given the -q option with -c or -l, or when handling a
219        binary file, it incorrectly wrote output to stdout.
221    29. The JIT compiler did not restore the control verb head in case of *THEN
222        control verbs. This issue was found by Karl Skomski with a custom LLVM
223        fuzzer.
225    30. Error messages for syntax errors following \g and \k were giving inaccurate
226        offsets in the pattern.
228    31. Added a check for integer overflow in conditions (?(<digits>) and
229        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
230        fuzzer.
232    32. Handling recursive references such as (?2) when the reference is to a group
233        later in the pattern uses code that is very hacked about and error-prone.
234        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
235        give an internal error if it is obvious that compiling has gone wrong.
237    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
238        This issue was found by Karl Skomski with a custom LLVM fuzzer.
240    34. The JIT compiler should restore the control chain for empty possessive
241        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
243    35. Match limit check added to JIT recursion. This issue was found by Karl
244        Skomski with a custom LLVM fuzzer.
246    36. Yet another case similar to 27 above has been circumvented by an
247        unconditional allocation of extra memory. This issue is fixed "properly" in
248        PCRE2 by refactoring the way references are handled. Wen Guanxing
249        from Venustech ADLAB discovered this bug.
251    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
252        with a custom LLVM fuzzer.
254    38. Fixed a corner case of range optimization in JIT.
256    39. An incorrect error "overran compiling workspace" was given if there were
257        exactly enough group forward references such that the last one extended
258        into the workspace safety margin. The next one would have expanded the
259        workspace. The test for overflow was not including the safety margin.
261    40. A match limit issue is fixed in JIT which was found by Karl Skomski
262        with a custom LLVM fuzzer.
264    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
265        work under Windows. (Why has it taken so long for anyone to notice?)
267    42. In a character class such as [\W\p{Any}] where both a negative-type escape
268        ("not a word character") and a property escape were present, the property
269        escape was being ignored.
271    43. Fix crash caused by very long (*MARK) or (*THEN) names.
273    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
274        by a single ASCII character in a class item, was incorrectly compiled in
275        UCP mode. The POSIX class got lost, but only if the single character
276        followed it.
278    45. [:punct:] in UCP mode was matching some characters in the range 128-255
279        that should not have been matched.
281    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
282        class, all characters with code points greater than 255 are in the class.
283        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
284        such as \w are turned into Unicode properties), wide characters were not
285        correctly handled, and could fail to match.
288    Version 8.37 28-April-2015
289    --------------------------
291    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
292        for those parentheses to be closed with whatever has been captured so far.
293        However, it was failing to mark any other groups between the hightest
294        capture so far and the currrent group as "unset". Thus, the ovector for
295        those groups contained whatever was previously there. An example is the
296        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
298    2.  If an assertion condition was quantified with a minimum of zero (an odd
299        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
301    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
302        unrecognized modifier, a crash could occur.
304    4.  An attempt to do global matching in pcretest with a zero-length ovector
305        caused a crash.
307    5.  Fixed a memory leak during matching that could occur for a subpattern
308        subroutine call (recursive or otherwise) if the number of captured groups
309        that had to be saved was greater than ten.
311    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
312        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
313        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
315    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
316        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
318    8. When a pattern is compiled, it remembers the highest back reference so that
319       when matching, if the ovector is too small, extra memory can be obtained to
320       use instead. A conditional subpattern whose condition is a check on a
321       capture having happened, such as, for example in the pattern
322       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
323       setting the highest backreference number. This mattered only if pcre_exec()
324       was called with an ovector that was too small to hold the capture, and there
325       was no other kind of back reference (a situation which is probably quite
326       rare). The effect of the bug was that the condition was always treated as
327       FALSE when the capture could not be consulted, leading to a incorrect
328       behaviour by pcre_exec(). This bug has been fixed.
330    9. A reference to a duplicated named group (either a back reference or a test
331       for being set in a conditional) that occurred in a part of the pattern where
332       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
333       to be incorrectly calculated, leading to overwriting.
335    10. A mutually recursive set of back references such as (\2)(\1) caused a
336        segfault at study time (while trying to find the minimum matching length).
337        The infinite loop is now broken (with the minimum length unset, that is,
338        zero).
340    11. If an assertion that was used as a condition was quantified with a minimum
341        of zero, matching went wrong. In particular, if the whole group had
342        unlimited repetition and could match an empty string, a segfault was
343        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
344        assertions to be quantified, but not if they are being used as conditions,
345        so the above pattern is faulted by Perl. PCRE has now been changed so that
346        it also rejects such patterns.
348    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
349        failed to allow the zero-repeat case if pcre2_exec() was called with an
350        ovector too small to capture the group.
352    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
353        Red Hat Product Security:
355        (a) A crash if /K and /F were both set with the option to save the compiled
356        pattern.
358        (b) Another crash if the option to print captured substrings in a callout
359        was combined with setting a null ovector, for example \O\C+ as a subject
360        string.
362    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
363        forward reference repeated a large (but limited) number of times within a
364        repeated outer group that has a zero minimum quantifier, caused incorrect
365        code to be compiled, leading to the error "internal error:
366        previously-checked referenced subpattern not found" when an incorrect
367        memory address was read. This bug was reported as "heap overflow",
368        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
369        CVE-2015-2325.
371    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
372        call within a group that also contained a recursive back reference caused
373        incorrect code to be compiled. This bug was reported as "heap overflow",
374        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
375        number CVE-2015-2326.
377    24. Computing the size of the JIT read-only data in advance has been a source
378        of various issues, and new ones are still appear unfortunately. To fix
379        existing and future issues, size computation is eliminated from the code,
380        and replaced by on-demand memory allocation.
382    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
383        adjacent to the end of the range, and the range contained characters with
384        more than one other case, caused incorrect behaviour when compiled in UTF
385        mode. In that example, the range a-j was left out of the class.
387    26. Fix JIT compilation of conditional blocks, which assertion
388        is converted to (*FAIL). E.g: /(?(?!))/.
390    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
391        discovered by the LLVM fuzzer.
393    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
394        when this assertion was used as a condition, for example (?(?!)a|b). In
395        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
396        error about an unsupported item.
398    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
399        possessification code could take exponential time to complete. A recursion
400        depth limit of 1000 has been imposed to limit the resources used by this
401        optimization.
403    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
404        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
405        because \S ensures they are all in the class. The code for doing this was
406        interacting badly with the code for computing the amount of space needed to
407        compile the pattern, leading to a buffer overflow. This bug was discovered
408        by the LLVM fuzzer.
410    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
411        other kinds of group caused stack overflow at compile time. This bug was
412        discovered by the LLVM fuzzer.
414    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
415        between a subroutine call and its quantifier was incorrectly compiled,
416        leading to buffer overflow or other errors. This bug was discovered by the
417        LLVM fuzzer.
419    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
420        assertion after (?(. The code was failing to check the character after
421        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
422        was discovered by the LLVM fuzzer.
424    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
425        a fixed maximum following a group that contains a subroutine reference was
426        incorrectly compiled and could trigger buffer overflow. This bug was
427        discovered by the LLVM fuzzer.
429    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
430        caused a stack overflow instead of the diagnosis of a non-fixed length
431        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
433    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
434        (e.g. /(?<=\Ka)/) could make pcregrep loop.
436    37. There was a similar problem to 36 in pcretest for global matches.
438    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
439        and a subsequent item in the pattern caused a non-match, backtracking over
440        the repeated \X did not stop, but carried on past the start of the subject,
441        causing reference to random memory and/or a segfault. There were also some
442        other cases where backtracking after \C could crash. This set of bugs was
443        discovered by the LLVM fuzzer.
445    39. The function for finding the minimum length of a matching string could take
446        a very long time if mutual recursion was present many times in a pattern,
447        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
448        method has been implemented. This infelicity was discovered by the LLVM
449        fuzzer.
451    40. Static linking against the PCRE library using the pkg-config module was
452        failing on missing pthread symbols.
455    Version 8.36 26-September-2014
456    ------------------------------
458    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
459        -Wmissing-field-initializers and -Wunused-parameter.
461    2.  The tests for quantifiers being too big (greater than 65535) were being
462        applied after reading the number, and stupidly assuming that integer
463        overflow would give a negative number. The tests are now applied as the
464        numbers are read.
466    3.  Tidy code in pcre_exec.c where two branches that used to be different are
467        now the same.
469    4.  The JIT compiler did not generate match limit checks for certain
470        bracketed expressions with quantifiers. This may lead to exponential
471        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
472        issue should be resolved now.
474    5.  Fixed an issue, which occures when nested alternatives are optimized
475        with table jumps.
477    6.  Inserted two casts and changed some ints to size_t in the light of some
478        reported 64-bit compiler warnings (Bugzilla 1477).
480    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
481        an empty string, which sometimes were behaving incorrectly in the
482        interpreter (though correctly in the JIT matcher). This pcretest input is
483        an example:
485          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
488        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
489        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
490        for an empty string was breaking the inner loop and carrying on at a lower
491        level, when possessive repeated groups should always return to a higher
492        level as they have no backtrack points in them. The empty string test now
493        occurs at the outer level.
495    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
496        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
498    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
499        doesn't).
501    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
502        to hold a pattern to be incorrectly computed (too small) when there were
503        named back references to duplicated names. This could cause "internal
504        error: code overflow" or "double free or corruption" or other memory
505        handling errors.
507    11. When named subpatterns had the same prefixes, back references could be
508        confused. For example, in this pattern:
510          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
512        the reference to 'Name' was incorrectly treated as a reference to a
513        duplicate name.
515    12. A pattern such as /^s?c/mi8 where the optional character has more than
516        one "other case" was incorrectly compiled such that it would only try to
517        match starting at "c".
519    13. When a pattern starting with \s was studied, VT was not included in the
520        list of possible starting characters; this should have been part of the
521        8.34/18 patch.
523    14. If a character class started [\Qx]... where x is any character, the class
524        was incorrectly terminated at the ].
526    15. If a pattern that started with a caseless match for a character with more
527        than one "other case" was studied, PCRE did not set up the starting code
528        unit bit map for the list of possible characters. Now it does. This is an
529        optimization improvement, not a bug fix.
531    16. The Unicode data tables have been updated to Unicode 7.0.0.
533    17. Fixed a number of memory leaks in pcregrep.
535    18. Avoid a compiler warning (from some compilers) for a function call with
536        a cast that removes "const" from an lvalue by using an intermediate
537        variable (to which the compiler does not object).
539    19. Incorrect code was compiled if a group that contained an internal recursive
540        back reference was optional (had quantifier with a minimum of zero). This
541        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
542        caused segmentation faults because of stack overflows at compile time.
544    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
545        group that is quantified with an indefinite repeat, caused a compile-time
546        loop which used up all the system stack and provoked a segmentation fault.
547        This was not the same bug as 19 above.
549    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
550        Patch by Mike Frysinger.
553    Version 8.35 04-April-2014
554    --------------------------
556    1.  A new flag is set, when property checks are present in an XCLASS.
557        When this flag is not set, PCRE can perform certain optimizations
558        such as studying these XCLASS-es.
560    2.  The auto-possessification of character sets were improved: a normal
561        and an extended character set can be compared now. Furthermore
562        the JIT compiler optimizes more character set checks.
564    3.  Got rid of some compiler warnings for potentially uninitialized variables
565        that show up only when compiled with -O2.
567    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
568        of a match later then the end of the match. The pcretest program was not
569        handling the case sensibly - it was outputting from the start to the next
570        binary zero. It now reports this situation in a message, and outputs the
571        text from the end to the start.
573    5.  Fast forward search is improved in JIT. Instead of the first three
574        characters, any three characters with fixed position can be searched.
575        Search order: first, last, middle.
577    6.  Improve character range checks in JIT. Characters are read by an inprecise
578        function now, which returns with an unknown value if the character code is
579        above a certain threshold (e.g: 256). The only limitation is that the value
580        must be bigger than the threshold as well. This function is useful when
581        the characters above the threshold are handled in the same way.
583    7.  The macros whose names start with RAWUCHAR are placeholders for a future
584        mode in which only the bottom 21 bits of 32-bit data items are used. To
585        make this more memorable for those maintaining the code, the names have
586        been changed to start with UCHAR21, and an extensive comment has been added
587        to their definition.
589    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
590        to the export list in Makefile.am (they were accidentally omitted from the
591        8.34 tarball).
593    9.  The informational output from pcretest used the phrase "starting byte set"
594        which is inappropriate for the 16-bit and 32-bit libraries. As the output
595        for "first char" and "need char" really means "non-UTF-char", I've changed
596        "byte" to "char", and slightly reworded the output. The documentation about
597        these values has also been (I hope) clarified.
599    10. Another JIT related optimization: use table jumps for selecting the correct
600        backtracking path, when more than four alternatives are present inside a
601        bracket.
603    11. Empty match is not possible, when the minimum length is greater than zero,
604        and there is no \K in the pattern. JIT should avoid empty match checks in
605        such cases.
607    12. In a caseless character class with UCP support, when a character with more
608        than one alternative case was not the first character of a range, not all
609        the alternative cases were added to the class. For example, s and \x{17f}
610        are both alternative cases for S: the class [RST] was handled correctly,
611        but [R-T] was not.
613    13. The configure.ac file always checked for pthread support when JIT was
614        enabled. This is not used in Windows, so I have put this test inside a
615        check for the presence of windows.h (which was already tested for).
617    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
618        The algorithm provides a way to skip certain starting offsets, and usually
619        faster than linear prefix searches.
621    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
622        as for 'fr_FR' and 'french'. For some reason, however, it then used the
623        Windows-specific input and output files, which have 'french' screwed in.
624        So this could never have worked. One of the problems with locales is that
625        they aren't always the same. I have now updated RunTest so that it checks
626        the output of the locale test (test 3) against three different output
627        files, and it allows the test to pass if any one of them matches. With luck
628        this should make the test pass on some versions of Solaris where it was
629        failing. Because of the uncertainty, the script did not used to stop if
630        test 3 failed; it now does. If further versions of a French locale ever
631        come to light, they can now easily be added.
633    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
634        there was a message during ./configure, but it did not stop. This now
635        provokes an error. The invalid example in README has been corrected.
636        If a value less than the minimum is given, the minimum value has always
637        been used, but now a warning is given.
639    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
640        was a bug in the test system, which is now fixed. Also, the list of various
641        configurations that are tested for each release did not have one with both
642        16/32 bits and --enable-bar-anycrlf. It now does.
644    18. pcretest was missing "-C bsr" for displaying the \R default setting.
646    19. Little endian PowerPC systems are supported now by the JIT compiler.
648    20. The fast forward newline mechanism could enter to an infinite loop on
649        certain invalid UTF-8 input. Although we don't support these cases
650        this issue can be fixed by a performance optimization.
652    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
653        not take account if existing stack usage. There is now a new global
654        variable called pcre_stack_guard that can be set to point to an external
655        function to check stack availability. It is called at the start of
656        processing every parenthesized group.
658    22. A typo in the code meant that in ungreedy mode the max/min qualifier
659        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
660        not match "ab".
662    23. When UTF was disabled, the JIT program reported some incorrect compile
663        errors. These messages are silenced now.
665    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
666        compiler.
668    25. Change all the temporary files used in RunGrepTest to be different to those
669        used by RunTest so that the tests can be run simultaneously, for example by
670        "make -j check".
673    Version 8.34 15-December-2013
674    -----------------------------
676    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
677        executable memory. Patch inspired by Carsten Klein.
679    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
680        this macro is never tested and has no effect, because the work to support
681        coverage involves only compiling and linking options and special targets in
682        the Makefile. The comment in config.h implied that defining the macro would
683        enable coverage support, which is totally false. There was also support for
684        setting this macro in the CMake files (my fault, I just copied it from
685        configure). SUPPORT_GCOV has now been removed.
687    3.  Make a small performance improvement in strlen16() and strlen32() in
688        pcretest.
690    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
691        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
692        these cases). There was also one in pcretest.c.
694    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
696    6.  In UTF mode, the code for checking whether a group could match an empty
697        string (which is used for indefinitely repeated groups to allow for
698        breaking an infinite loop) was broken when the group contained a repeated
699        negated single-character class with a character that occupied more than one
700        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
701        UTF-8 mode). The effect was undefined: the group might or might not be
702        deemed as matching an empty string, or the program might have crashed.
704    7.  The code for checking whether a group could match an empty string was not
705        recognizing that \h, \H, \v, \V, and \R must match a character.
707    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
708        an empty string. If it can, pcretest shows this in its information output.
710    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
711        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
712        matched by pcre_exec() without using JIT:
714        (a) If the rest of the pattern did not match after a maximal run of
715            grapheme clusters, the code for backing up to try with fewer of them
716            did not always back up over a full grapheme when characters that do not
717            have the modifier quality were involved, e.g. Hangul syllables.
719        (b) If the match point in a subject started with modifier character, and
720            there was no match, the code could incorrectly back up beyond the match
721            point, and potentially beyond the first character in the subject,
722            leading to a segfault or an incorrect match result.
724    10. A conditional group with an assertion condition could lead to PCRE
725        recording an incorrect first data item for a match if no other first data
726        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
727        first data item, and therefore matched "ca" after "c" instead of at the
728        start.
730    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
731        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
733    12. The source of pcregrep now includes z/OS-specific code so that it can be
734        compiled for z/OS as part of the special z/OS distribution.
736    13. Added the -T and -TM options to pcretest.
738    14. The code in pcre_compile.c for creating the table of named capturing groups
739        has been refactored. Instead of creating the table dynamically during the
740        actual compiling pass, the information is remembered during the pre-compile
741        pass (on the stack unless there are more than 20 named groups, in which
742        case malloc() is used) and the whole table is created before the actual
743        compile happens. This has simplified the code (it is now nearly 150 lines
744        shorter) and prepared the way for better handling of references to groups
745        with duplicate names.
747    15. A back reference to a named subpattern when there is more than one of the
748        same name now checks them in the order in which they appear in the pattern.
749        The first one that is set is used for the reference. Previously only the
750        first one was inspected. This change makes PCRE more compatible with Perl.
752    16. Unicode character properties were updated from Unicode 6.3.0.
754    17. The compile-time code for auto-possessification has been refactored, based
755        on a patch by Zoltan Herczeg. It now happens after instead of during
756        compilation. The code is cleaner, and more cases are handled. The option
757        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
758        options in pcretest are provided to set it. It can also be set by
759        (*NO_AUTO_POSSESS) at the start of a pattern.
761    18. The character VT has been added to the default ("C" locale) set of
762        characters that match \s and are generally treated as white space,
763        following this same change in Perl 5.18. There is now no difference between
764        "Perl space" and "POSIX space". Whether VT is treated as white space in
765        other locales depends on the locale.
767    19. The code for checking named groups as conditions, either for being set or
768        for being recursed, has been refactored (this is related to 14 and 15
769        above). Processing unduplicated named groups should now be as fast at
770        numerical groups, and processing duplicated groups should be faster than
771        before.
773    20. Two patches to the CMake build system, by Alexander Barkov:
775          (1) Replace the "source" command by "." in CMakeLists.txt because
776              "source" is a bash-ism.
778          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
779              without these the CMake build does not work on Solaris.
781    21. Perl has changed its handling of \8 and \9. If there is no previously
782        encountered capturing group of those numbers, they are treated as the
783        literal characters 8 and 9 instead of a binary zero followed by the
784        literals. PCRE now does the same.
786    22. Following Perl, added \o{} to specify codepoints in octal, making it
787        possible to specify values greater than 0777 and also making them
788        unambiguous.
790    23. Perl now gives an error for missing closing braces after \x{... instead of
791        treating the string as literal. PCRE now does the same.
793    24. RunTest used to grumble if an inappropriate test was selected explicitly,
794        but just skip it when running all tests. This make it awkward to run ranges
795        of tests when one of them was inappropriate. Now it just skips any
796        inappropriate tests, as it always did when running all tests.
798    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
799        character types such as \d or \w, too many callouts were inserted, and the
800        data that they returned was rubbish.
802    26. In UCP mode, \s was not matching two of the characters that Perl matches,
803        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
804        were matched by \h. The code has now been refactored so that the lists of
805        the horizontal and vertical whitespace characters used for \h and \v (which
806        are defined only in one place) are now also used for \s.
808    27. Add JIT support for the 64 bit TileGX architecture.
809        Patch by Jiong Wang (Tilera Corporation).
811    28. Possessive quantifiers for classes (both explicit and automatically
812        generated) now use special opcodes instead of wrapping in ONCE brackets.
814    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
815        (because it's meaningless), this was not happening when PCRE_CASELESS was
816        set. Not wrong, but inefficient.
818    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
819        properties for \w, \d, etc) is present in a test regex. Otherwise if the
820        test contains no characters greater than 255, Perl doesn't realise it
821        should be using Unicode semantics.
823    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
824        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
825        does in Unicode mode.
827    32. Added the "forbid" facility to pcretest so that putting tests into the
828        wrong test files can sometimes be quickly detected.
830    33. There is now a limit (default 250) on the depth of nesting of parentheses.
831        This limit is imposed to control the amount of system stack used at compile
832        time. It can be changed at build time by --with-parens-nest-limit=xxx or
833        the equivalent in CMake.
835    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
836        errors. Perl warns for these when in warning mode, but PCRE has no facility
837        for giving warnings.
839    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
840        However, this was not working for (?!) because it is optimized to (*FAIL),
841        for which PCRE does not allow quantifiers. The optimization is now disabled
842        when a quantifier follows (?!). I can't see any use for this, but it makes
843        things uniform.
845    36. Perl no longer allows group names to start with digits, so I have made this
846        change also in PCRE. It simplifies the code a bit.
848    37. In extended mode, Perl ignores spaces before a + that indicates a
849        possessive quantifier. PCRE allowed a space before the quantifier, but not
850        before the possessive +. It now does.
852    38. The use of \K (reset reported match start) within a repeated possessive
853        group such as (a\Kb)*+ was not working.
855    40. Document that the same character tables must be used at compile time and
856        run time, and that the facility to pass tables to pcre_exec() and
857        pcre_dfa_exec() is for use only with saved/restored patterns.
859    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
860        features for Builds with MSVC:
862        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
863           builds). The .rc files can be used to set FileDescription and many other
864           attributes.
866        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
867           This allows higher-level build scripts which want .pdb files to avoid
868           hard-coding the exact files needed."
870    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
871        mean "start of word" and "end of word", respectively, as a transition aid.
873    43. A minimizing repeat of a class containing codepoints greater than 255 in
874        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
875        compiled to use the heap for recursion.
877    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
878        is configured.
881    Version 8.33 28-May-2013
882    ------------------------
884    1.  Added 'U' to some constants that are compared to unsigned integers, to
885        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
886        variables that are added to signed variables, to ensure the result is
887        signed and can be negated.
889    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
890        CMake config files.
892    3.  Revise the creation of config.h.generic so that all boolean macros are
893        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
894        overriding via -D on the command line possible.
896    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
897        to unsigned int is reported to make a quite noticeable speed difference in
898        a specific Windows environment. Testing on Linux did also appear to show
899        some benefit (and it is clearly not harmful). Also fixed the definition of
900        Xop which should be unsigned.
902    5.  Related to (4), changing the definition of the intermediate variable cc
903        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
904        improvements.
906    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
907        unnecessary spaces.
909    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
910        and later.
912    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
914    9.  Optimizing fast_forward_start_bits in JIT.
916    10. Adding support for callouts in JIT, and fixing some issues revealed
917        during this work. Namely:
919        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
921        (b) Minimum length was not checked before the matching is started.
923    11. The value of capture_last that is passed to callouts was incorrect in some
924        cases when there was a capture on one path that was subsequently abandoned
925        after a backtrack. Also, the capture_last value is now reset after a
926        recursion, since all captures are also reset in this case.
928    12. The interpreter no longer returns the "too many substrings" error in the
929        case when an overflowing capture is in a branch that is subsequently
930        abandoned after a backtrack.
932    13. In the pathological case when an offset vector of size 2 is used, pcretest
933        now prints out the matched string after a yield of 0 or 1.
935    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
936        Only supported by the JIT compiler at the moment.
938    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
940    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
941        offset of the starting point of the matching process, provided the offsets
942        vector is large enough.
944    17. The \A escape now records a lookbehind value of 1, though its execution
945        does not actually inspect the previous character. This is to ensure that,
946        in partial multi-segment matching, at least one character from the old
947        segment is retained when a new segment is processed. Otherwise, if there
948        are no lookbehinds in the pattern, \A might match incorrectly at the start
949        of a new segment.
951    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
953    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
954        gives some modest performance improvement in 8-bit mode.
956    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
957        be expressed in certain programming languages using Universal Character
958        Names.
960    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
961        which points out that "non characters" are not "characters that may not
962        appear in Unicode strings" but rather "characters that are reserved for
963        internal use and have only local meaning".
965    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
966        there was a conditional group that depended on an assertion, if the
967        assertion was false, the callout that immediately followed the alternation
968        in the condition was skipped when pcre_exec() was used for matching.
970    23. Allow an explicit callout to be inserted before an assertion that is the
971        condition for a conditional group, for compatibility with automatic
972        callouts, which always insert a callout at this point.
974    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
975        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
977    25. (*PRUNE) is now supported by the JIT compiler.
979    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
981    27. Fix the case where there are two or more SKIPs with arguments that may be
982        ignored.
984    28. (*SKIP) is now supported by the JIT compiler.
986    29. (*THEN) is now supported by the JIT compiler.
988    30. Update RunTest with additional test selector options.
990    31. The way PCRE handles backtracking verbs has been changed in two ways.
992        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
993        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
994        backtracking. In some cases this makes it more Perl-compatible, but Perl's
995        rather obscure rules do not always do the same thing.
997        (2) Previously, backtracking verbs were confined within assertions. This is
998        no longer the case for positive assertions, except for (*ACCEPT). Again,
999        this sometimes improves Perl compatibility, and sometimes does not.
1001    32. A number of tests that were in test 2 because Perl did things differently
1002        have been moved to test 1, because either Perl or PCRE has changed, and
1003        these tests are now compatible.
1005    32. Backtracking control verbs are now handled in the same way in JIT and
1006        interpreter.
1008    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1009        contained a forward subroutine reference caused a compile error.
1011    34. Auto-detect and optimize limited repetitions in JIT.
1013    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1014        blocking (*UTF) etc.
1016    36. In the interpreter, maximizing pattern repetitions for characters and
1017        character types now use tail recursion, which reduces stack usage.
1019    37. The value of the max lookbehind was not correctly preserved if a compiled
1020        and saved regex was reloaded on a host of different endianness.
1022    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1023        of the compiled pattern block, expand the flags field from 16 to 32 bits
1024        because it was almost full.
1026    39. Try madvise first before posix_madvise.
1028    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1029        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1030        original change locked out all matches of empty strings. This has been
1031        changed so that one match of an empty string per line is recognized.
1032        Subsequent searches on the same line (for colouring or for --only-matching,
1033        for example) do not recognize empty strings.
1035    41. Applied a user patch to fix a number of spelling mistakes in comments.
1037    42. Data lines longer than 65536 caused pcretest to crash.
1039    43. Clarified the data type for length and startoffset arguments for pcre_exec
1040        and pcre_dfa_exec in the function-specific man pages, where they were
1041        explicitly stated to be in bytes, never having been updated. I also added
1042        some clarification to the pcreapi man page.
1044    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1045        a segmentation fault.
1048    Version 8.32 30-November-2012
1049    -----------------------------
1051    1.  Improved JIT compiler optimizations for first character search and single
1052        character iterators.
1054    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1055        Patch by Daniel Richard G.
1057    3.  Single character iterator optimizations in the JIT compiler.
1059    4.  Improved JIT compiler optimizations for character ranges.
1061    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1062        Reported by Giuseppe D'Angelo.
1064    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1065        of a line, was being set incorrectly in cases where .* appeared inside
1066        atomic brackets at the start of a pattern, or where there was a subsequent
1067        *PRUNE or *SKIP.
1069    7.  Improved instruction cache flush for POWER/PowerPC.
1070        Patch by Daniel Richard G.
1072    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1073        grep:
1075        (a) There is now no limit to the number of patterns to be matched.
1077        (b) An error is given if a pattern is too long.
1079        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1080            are now supported.
1082        (d) --exclude-from and --include-from (multiple use) have been added.
1084        (e) Exclusions and inclusions now apply to all files and directories, not
1085            just to those obtained from scanning a directory recursively.
1087        (f) Multiple uses of -f and --file-list are now supported.
1089        (g) In a Windows environment, the default for -d has been changed from
1090            "read" (the GNU grep default) to "skip", because otherwise the presence
1091            of a directory in the file list provokes an error.
1093        (h) The documentation has been revised and clarified in places.
1095    9.  Improve the matching speed of capturing brackets.
1097    10. Changed the meaning of \X so that it now matches a Unicode extended
1098        grapheme cluster.
1100    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1101        out POSIX threads when JIT support is configured.
1103    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1105    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1106        structure could go wrong in environments where size_t is not the same size
1107        as int.
1109    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1110        set.
1112    15. The EBCDIC support had decayed; later updates to the code had included
1113        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1114        general tidy up of EBCDIC-related issues, and the documentation was also
1115        not quite right. There is now a test that can be run on ASCII systems to
1116        check some of the EBCDIC-related things (but is it not a full test).
1118    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1119        in a small tidy to the code.
1121    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1123    18. If the --only-matching (-o) option in pcregrep is specified multiple
1124        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1125        substrings matched by the 1st and 2nd capturing parentheses. A separating
1126        string can be specified by --om-separator (default empty).
1128    19. Improving the first n character searches.
1130    20. Turn case lists for horizontal and vertical white space into macros so that
1131        they are defined only once.
1133    21. This set of changes together give more compatible Unicode case-folding
1134        behaviour for characters that have more than one other case when UCP
1135        support is available.
1137        (a) The Unicode property table now has offsets into a new table of sets of
1138            three or more characters that are case-equivalent. The MultiStage2.py
1139            script that generates these tables (the pcre_ucd.c file) now scans
1140            CaseFolding.txt instead of UnicodeData.txt for character case
1141            information.
1143        (b) The code for adding characters or ranges of characters to a character
1144            class has been abstracted into a generalized function that also handles
1145            case-independence. In UTF-mode with UCP support, this uses the new data
1146            to handle characters with more than one other case.
1148        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1149            whose other case is greater than 256 are now correctly matched
1150            caselessly. Previously, the high codepoint matched the low one, but not
1151            vice versa.
1153        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1154            of the new class addition function, using character lists defined as
1155            macros alongside the case definitions of 20 above.
1157        (e) Caseless back references now work with characters that have more than
1158            one other case.
1160        (f) General caseless matching of characters with more than one other case
1161            is supported.
1163    22. Unicode character properties were updated from Unicode 6.2.0
1165    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1167    24. Add support for 32-bit character strings, and UTF-32
1169    25. Major JIT compiler update (code refactoring and bugfixing).
1170        Experimental Sparc 32 support is added.
1172    26. Applied a modified version of Daniel Richard G's patch to create
1173        pcre.h.generic and config.h.generic by "make" instead of in the
1174        PrepareRelease script.
1176    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1177        pcre_compile.c when checking for a zero character.
1179    28. Introducing a native interface for JIT. Through this interface, the compiled
1180        machine code can be directly executed. The purpose of this interface is to
1181        provide fast pattern matching, so several sanity checks are not performed.
1182        However, feature tests are still performed. The new interface provides
1183        1.4x speedup compared to the old one.
1185    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1186        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1187        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1189    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1190        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1191        the "old" RFC 2279). Instead, it ended up passing a negative length to
1192        pcre_exec().
1194    31. Add support for GCC's visibility feature to hide internal functions.
1196    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1197        "unknown -C option" after outputting 0 or 1.
1199    33. There is now support for generating a code coverage report for the test
1200        suite in environments where gcc is the compiler and lcov is installed. This
1201        is mainly for the benefit of the developers.
1203    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1204        unaddressable using valgrind annotations, allowing valgrind to detect
1205        invalid memory accesses. This is mainly for the benefit of the developers.
1207    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1209    26. Give configure error if --enable-cpp but no C++ compiler found.
1212    Version 8.31 06-July-2012
1213    -------------------------
1215    1.  Fixing a wrong JIT test case and some compiler warnings.
1217    2.  Removed a bashism from the RunTest script.
1219    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1220        to unsigned type, result still unsigned" that was given by an MS compiler
1221        on encountering the code "-sizeof(xxx)".
1223    4.  Partial matching support is added to the JIT compiler.
1225    5.  Fixed several bugs concerned with partial matching of items that consist
1226        of more than one character:
1228        (a) /^(..)\1/ did not partially match "aba" because checking references was
1229            done on an "all or nothing" basis. This also applied to repeated
1230            references.
1232        (b) \R did not give a hard partial match if \r was found at the end of the
1233            subject.
1235        (c) \X did not give a hard partial match after matching one or more
1236            characters at the end of the subject.
1238        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1239            a partial match for the string "\r".
1241        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1242            a partial match for a CR character at the end of the subject string.
1244    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1245        running pcretest, the text "(JIT)" added to the output whenever JIT is
1246        actually used to run the match.
1248    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1249        or /S+[+] with a digit between 1 and 7.
1251    8.  OP_NOT now supports any UTF character not just single-byte ones.
1253    9.  (*MARK) control verb is now supported by the JIT compiler.
1255    10. The command "./RunTest list" lists the available tests without actually
1256        running any of them. (Because I keep forgetting what they all are.)
1260    12. Applied a (slightly modified) user-supplied patch that improves performance
1261        when the heap is used for recursion (compiled with --disable-stack-for-
1262        recursion). Instead of malloc and free for each heap frame each time a
1263        logical recursion happens, frames are retained on a chain and re-used where
1264        possible. This sometimes gives as much as 30% improvement.
1266    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1267        call.
1269    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1271    15. It is now possible to link pcretest with libedit as an alternative to
1272        libreadline.
1274    16. (*COMMIT) control verb is now supported by the JIT compiler.
1276    17. The Unicode data tables have been updated to Unicode 6.1.0.
1278    18. Added --file-list option to pcregrep.
1280    19. Added binary file support to pcregrep, including the -a, --binary-files,
1281        -I, and --text options.
1283    20. The madvise function is renamed for posix_madvise for QNX compatibility
1284        reasons. Fixed by Giuseppe D'Angelo.
1286    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1288    22. Changed the option for creating symbolic links for 16-bit man pages from
1289        -s to -sf so that re-installing does not cause issues.
1291    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1293    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1294        matches in certain environments (the workspace was not being correctly
1295        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1296        some of the workspace data at the beginning of a restart.
1298    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1299        was not doing so when it should - probably a typo introduced by SVN 528
1300        (change 8.10/14).
1302    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1303        \w+ when the character tables indicated that \x{c4} was a word character.
1304        There were several related cases, all because the tests for doing a table
1305        lookup were testing for characters less than 127 instead of 255.
1307    27. If a pattern contains capturing parentheses that are not used in a match,
1308        their slots in the ovector are set to -1. For those that are higher than
1309        any matched groups, this happens at the end of processing. In the case when
1310        there were back references that the ovector was too small to contain
1311        (causing temporary malloc'd memory to be used during matching), and the
1312        highest capturing number was not used, memory off the end of the ovector
1313        was incorrectly being set to -1. (It was using the size of the temporary
1314        memory instead of the true size.)
1316    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1317        ovector size, it uses memory at the end of the block that it has got.
1319    29. Check for an overlong MARK name and give an error at compile time. The
1320        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1322    30. JIT compiler update.
1324    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1325        Rill for the patch.
1327    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1329    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1331    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1332        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1333        was enabled.
1335    35. Improve JIT code generation for greedy plus quantifier.
1337    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1338        1 to "aa" instead of to an empty string. The bug affected repeated groups
1339        that could potentially match an empty string.
1341    37. Optimizing single character iterators in JIT.
1343    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1344        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1345        codepoints that are too big for the mode are faulted, and in a UTF mode,
1346        disallowed codepoints are also faulted.
1348    39. If PCRE was compiled with UTF support, in three places in the DFA
1349        matcher there was code that should only have been obeyed in UTF mode, but
1350        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1351        processing when bytes with values greater than 127 were present. In 16-bit
1352        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1353        both cases the values are those that cannot be the first data item in a UTF
1354        character. The three items that might have provoked this were recursions,
1355        possessively repeated groups, and atomic groups.
1357    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1358        and pcregrep, because some OS require shared objects to be explicitly
1359        passed to ld, causing the link step to fail if they are not.
1361    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1362        mode, patterns that started with \h* or \R* might be incorrectly matched.
1365    Version 8.30 04-February-2012
1366    -----------------------------
1368    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1369        name is defined in ctype.h.
1371    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1372        only in quite long subpatterns.
1374    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1375        since it was replaced by pcre_fullinfo() in February 2000.
1377    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1378        match a (*MARK), and the match failed at the start of the subject, a
1379        reference to memory before the start of the subject could occur. This bug
1380        was introduced by fix 17 of release 8.21.
1382    5.  A reference to an unset group with zero minimum repetition was giving
1383        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1384        /(another)?(\1?)test/ matched against "hello world test". This bug was
1385        introduced in release 8.13.
1387    6.  Add support for 16-bit character strings (a large amount of work involving
1388        many changes and refactorings).
1390    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1391        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1392        from a file.
1394    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1395        rounding is not applied in this particular case).
1397    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1398        if they appear, or are escaped, in patterns.
1400    10. Get rid of a number of -Wunused-but-set-variable warnings.
1402    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1403        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1404        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1405        also returns the mark "x". This bug applied to capturing parentheses,
1406        non-capturing parentheses, and atomic parentheses. It also applied to some
1407        assertions.
1409    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1410        information out of configure.ac instead of relying on pcre.h.generic, which
1411        is not stored in the repository.
1413    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1414        -lreadline.
1416    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1418    15. Applied Graycode's patch to put the top-level frame on the stack rather
1419        than the heap when not using the stack for recursion. This gives a
1420        performance improvement in many cases when recursion is not deep.
1422    16. Experimental code added to "pcretest -C" to output the stack frame size.
1425    Version 8.21 12-Dec-2011
1426    ------------------------
1428  1.  Updating the JIT compiler.  1.  Updating the JIT compiler.
1430    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1431        are added as well.
1433    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1434        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1435        calling _pcre_jit_exec. Some extra comments are added.
1437    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1438        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1439        was introduced by change 18 for 8.20.
1441    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1442        ECMA-262 standard.
1444    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1445        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1446        This bug was probably introduced by change 9 of 8.13.
1448    7.  While fixing 6 above, I noticed that a number of other items were being
1449        incorrectly rejected as "not fixed length". This arose partly because newer
1450        opcodes had not been added to the fixed-length checking code. I have (a)
1451        corrected the bug and added tests for these items, and (b) arranged for an
1452        error to occur if an unknown opcode is encountered while checking for fixed
1453        length instead of just assuming "not fixed length". The items that were
1454        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1455        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1456        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1458    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1459        being incorrectly compiled and would have given unpredicatble results.
1461    9.  A possessively repeated subpattern with minimum repeat count greater than
1462        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1463        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1464        the first (A) could occur when it should not.
1466    10. Add a cast and remove a redundant test from the code.
1468    11. JIT should use pcre_malloc/pcre_free for allocation.
1470    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1471        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1472        is anything other than /usr, -L is still shown).
1474    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1476    14. Perl does not support \N without a following name in a [] class; PCRE now
1477        also gives an error.
1479    15. If a forward reference was repeated with an upper limit of around 2000,
1480        it caused the error "internal error: overran compiling workspace". The
1481        maximum number of forward references (including repeats) was limited by the
1482        internal workspace, and dependent on the LINK_SIZE. The code has been
1483        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1484        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1485        of around 200,000 forward references. While doing this, I also speeded up
1486        the filling in of repeated forward references.
1488    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1489        incorrectly expecting the subject to contain another "a" after the start.
1491    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1492        in the match, the SKIP should be ignored. This was not happening; instead
1493        the SKIP was being treated as NOMATCH. For patterns such as
1494        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1495        tested.
1497    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1498        now much more compatible with Perl, in particular in cases where the result
1499        is a non-match for a non-anchored pattern. For example, if
1500        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1501        "m", where previously it did not return a name. A side effect of this
1502        change is that for partial matches, the last encountered mark name is
1503        returned, as for non matches. A number of tests that were previously not
1504        Perl-compatible have been moved into the Perl-compatible test files. The
1505        refactoring has had the pleasing side effect of removing one argument from
1506        the match() function, thus reducing its stack requirements.
1508    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1509        subsequent uses of /S (without +) incorrectly behaved like /S+.
1511    21. Retrieve executable code size support for the JIT compiler and fixing
1512        some warnings.
1514    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1515        not work when the shorter character appeared right at the end of the
1516        subject string.
1518    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1519        systems.
1521    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1522        output it when the /M option is used in pcretest.
1524    25. The CheckMan script was not being included in the distribution. Also, added
1525        an explicit "perl" to run Perl scripts from the PrepareRelease script
1526        because this is reportedly needed in Windows.
1528    26. If study data was being save in a file and studying had not found a set of
1529        "starts with" bytes for the pattern, the data written to the file (though
1530        never used) was taken from uninitialized memory and so caused valgrind to
1531        complain.
1533    27. Updated RunTest.bat as provided by Sheri Pierce.
1535    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1537    29. Computation of memory usage for the table of capturing group names was
1538        giving an unnecessarily large value.
1541  Version 8.20 21-Oct-2011  Version 8.20 21-Oct-2011
1542  ------------------------  ------------------------
# Line 1021  Version 7.9 11-Apr-09 Line 2552  Version 7.9 11-Apr-09
2552  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
2553      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
2554      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2555      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2556        8.33.]
2558  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2559      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 1264  Version 7.7 07-May-08 Line 2796  Version 7.7 07-May-08
2796      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2797      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2798      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2799      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2800      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2801      was subsequently set up correctly.)      was subsequently set up correctly.)

Removed from v.740  
changed lines
  Added in v.1662

  ViewVC Help
Powered by ViewVC 1.1.5