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

Diff of /code/trunk/ChangeLog

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

revision 657 by ph10, Mon Aug 15 17:39:09 2011 UTC revision 1684 by ph10, Wed Feb 22 17:17:49 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.13 02-Aug-2011  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 8.41
8    ------------
9    
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
12    
13    2. Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14    mode with --only-matching matched several lines, it restarted scanning at the
15    next line instead of moving on to the end of the matched string, which can be
16    several lines after the start.
17    
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
19    
20    4.  A (?# style comment is now ignored between a basic quantifier and a
21    following '+' or '?' (example: /X+(?#comment)?Y/.
22    
23    5.  Avoid use of a potentially overflowing buffer in pcregrep (patch by Petr
24    Pisar).
25    
26    
27    Version 8.40 11-January-2017
28    ----------------------------
29    
30    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
31        the match extended over a line boundary.
32    
33    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
34        MSVC static compilation, putting the first patch under a new option.
35    
36    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
37    
38    4.  Ignore "show all captures" (/=) for DFA matching.
39    
40    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
41    
42    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
43        without PCRE_UCP set, a negative character type such as \D in a positive
44        class should cause all characters greater than 255 to match, whatever else
45        is in the class. There was a bug that caused this not to happen if a
46        Unicode property item was added to such a class, for example [\D\P{Nd}] or
47        [\W\pL].
48    
49    7.  When pcretest was outputing information from a callout, the caret indicator
50        for the current position in the subject line was incorrect if it was after
51        an escape sequence for a character whose code point was greater than
52        \x{ff}.
53    
54    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
55        the conditional was interpreted as a reference to capturing group 1 instead
56        of a test for recursion. Any group whose name began with R was
57        misinterpreted in this way. (The reference interpretation should only
58        happen if the group's name is precisely "R".)
59    
60    9.  A number of bugs have been mended relating to match start-up optimizations
61        when the first thing in a pattern is a positive lookahead. These all
62        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
63    
64        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
65            both an initial 'X' and a following 'X'.
66        (b) Some patterns starting with an assertion that started with .* were
67            incorrectly optimized as having to match at the start of the subject or
68            after a newline. There are cases where this is not true, for example,
69            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
70            start with spaces. Starting .* in an assertion is no longer taken as an
71            indication of matching at the start (or after a newline).
72    
73    
74    Version 8.39 14-June-2016
75    -------------------------
76    
77    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
78        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
79        misbehaved. This bug was found by the LLVM fuzzer.
80    
81    2.  Similar to the above, if an isolated \E was present between an item and its
82        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
83        bug was found by the LLVM fuzzer.
84    
85    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
86        working correctly in UCP mode.
87    
88    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
89        was set when the pmatch argument was NULL. It now returns REG_INVARG.
90    
91    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
92    
93    6.  An empty \Q\E sequence between an item and its qualifier caused
94        pcre_compile() to misbehave when auto callouts were enabled. This bug was
95        found by the LLVM fuzzer.
96    
97    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
98        space or a #-type comment that was followed by (?-x), which turns off
99        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
100        pcre_compile() assumed that (?-x) applied to the whole pattern and
101        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
102    
103    8.  A call of pcre_copy_named_substring() for a named substring whose number
104        was greater than the space in the ovector could cause a crash.
105    
106    9.  Yet another buffer overflow bug involved duplicate named groups with a
107        group that reset capture numbers (compare 8.38/7 below). Once again, I have
108        just allowed for more memory, even if not needed. (A proper fix is
109        implemented in PCRE2, but it involves a lot of refactoring.)
110    
111    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
112        start of the match to be earlier than the end.
113    
114    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
115    
116    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
117        assertion, caused pcretest to generate incorrect output, and also to read
118        uninitialized memory (detected by ASAN or valgrind).
119    
120    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
121        nested set of parentheses of sufficient size caused an overflow of the
122        compiling workspace (which was diagnosed, but of course is not desirable).
123    
124    14. And yet another buffer overflow bug involving duplicate named groups, this
125        time nested, with a nested back reference. Yet again, I have just allowed
126        for more memory, because anything more needs all the refactoring that has
127        been done for PCRE2. An example pattern that provoked this bug is:
128        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
129        registered as CVE-2016-1283.
130    
131    15. pcretest went into a loop if global matching was requested with an ovector
132        size less than 2. It now gives an error message. This bug was found by
133        afl-fuzz.
134    
135    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
136        ("assertion expected") when (?(?C) was not followed by an opening
137        parenthesis.
138    
139    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
140        actually affect anything, by sheer luck.
141    
142    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
143        static compilation.
144    
145    19. Modified the RunTest script to incorporate a valgrind suppressions file so
146        that certain errors, provoked by the SSE2 instruction set when JIT is used,
147        are ignored.
148    
149    20. A racing condition is fixed in JIT reported by Mozilla.
150    
151    21. Minor code refactor to avoid "array subscript is below array bounds"
152        compiler warning.
153    
154    22. Minor code refactor to avoid "left shift of negative number" warning.
155    
156    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
157        UCP support.
158    
159    24. Refactor to avoid compiler warnings in pcrecpp.cc.
160    
161    25. Refactor to fix a typo in pcre_jit_test.c
162    
163    26. Patch to support compiling pcrecpp.cc with Intel compiler.
164    
165    
166    Version 8.38 23-November-2015
167    -----------------------------
168    
169    1.  If a group that contained a recursive back reference also contained a
170        forward reference subroutine call followed by a non-forward-reference
171        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
172        compile correct code, leading to undefined behaviour or an internally
173        detected error. This bug was discovered by the LLVM fuzzer.
174    
175    2.  Quantification of certain items (e.g. atomic back references) could cause
176        incorrect code to be compiled when recursive forward references were
177        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
178        This bug was discovered by the LLVM fuzzer.
179    
180    3.  A repeated conditional group whose condition was a reference by name caused
181        a buffer overflow if there was more than one group with the given name.
182        This bug was discovered by the LLVM fuzzer.
183    
184    4.  A recursive back reference by name within a group that had the same name as
185        another group caused a buffer overflow. For example:
186        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
187    
188    5.  A forward reference by name to a group whose number is the same as the
189        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
190        a buffer overflow at compile time. This bug was discovered by the LLVM
191        fuzzer.
192    
193    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
194        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
195    
196    7.  Another buffer overflow bug involved duplicate named groups with a
197        reference between their definition, with a group that reset capture
198        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
199        fixed by always allowing for more memory, even if not needed. (A proper fix
200        is implemented in PCRE2, but it involves more refactoring.)
201    
202    8.  There was no check for integer overflow in subroutine calls such as (?123).
203    
204    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
205        being treated as a literal 'l' instead of causing an error.
206    
207    10. There was a buffer overflow if pcre_exec() was called with an ovector of
208        size 1. This bug was found by american fuzzy lop.
209    
210    11. If a non-capturing group containing a conditional group that could match
211        an empty string was repeated, it was not identified as matching an empty
212        string itself. For example: /^(?:(?(1)x|)+)+$()/.
213    
214    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
215        \a and \e in test subject lines.
216    
217    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
218        instead of the EBCDIC value.
219    
220    14. The handling of \c in an EBCDIC environment has been revised so that it is
221        now compatible with the specification in Perl's perlebcdic page.
222    
223    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
224        ASCII/Unicode. This has now been added to the list of characters that are
225        recognized as white space in EBCDIC.
226    
227    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
228        error (correctly) when used outside a class, but did not give an error
229        within a class.
230    
231    17. \h within a class was incorrectly compiled in EBCDIC environments.
232    
233    18. A pattern with an unmatched closing parenthesis that contained a backward
234        assertion which itself contained a forward reference caused buffer
235        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
236    
237    19. JIT should return with error when the compiled pattern requires more stack
238        space than the maximum.
239    
240    20. A possessively repeated conditional group that could match an empty string,
241        for example, /(?(R))*+/, was incorrectly compiled.
242    
243    21. Fix infinite recursion in the JIT compiler when certain patterns such as
244        /(?:|a|){100}x/ are analysed.
245    
246    22. Some patterns with character classes involving [: and \\ were incorrectly
247        compiled and could cause reading from uninitialized memory or an incorrect
248        error diagnosis.
249    
250    23. Pathological patterns containing many nested occurrences of [: caused
251        pcre_compile() to run for a very long time.
252    
253    24. A conditional group with only one branch has an implicit empty alternative
254        branch and must therefore be treated as potentially matching an empty
255        string.
256    
257    25. If (?R was followed by - or + incorrect behaviour happened instead of a
258        diagnostic.
259    
260    26. Arrange to give up on finding the minimum matching length for overly
261        complex patterns.
262    
263    27. Similar to (4) above: in a pattern with duplicated named groups and an
264        occurrence of (?| it is possible for an apparently non-recursive back
265        reference to become recursive if a later named group with the relevant
266        number is encountered. This could lead to a buffer overflow. Wen Guanxing
267        from Venustech ADLAB discovered this bug.
268    
269    28. If pcregrep was given the -q option with -c or -l, or when handling a
270        binary file, it incorrectly wrote output to stdout.
271    
272    29. The JIT compiler did not restore the control verb head in case of *THEN
273        control verbs. This issue was found by Karl Skomski with a custom LLVM
274        fuzzer.
275    
276    30. Error messages for syntax errors following \g and \k were giving inaccurate
277        offsets in the pattern.
278    
279    31. Added a check for integer overflow in conditions (?(<digits>) and
280        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
281        fuzzer.
282    
283    32. Handling recursive references such as (?2) when the reference is to a group
284        later in the pattern uses code that is very hacked about and error-prone.
285        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
286        give an internal error if it is obvious that compiling has gone wrong.
287    
288    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
289        This issue was found by Karl Skomski with a custom LLVM fuzzer.
290    
291    34. The JIT compiler should restore the control chain for empty possessive
292        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
293    
294    35. Match limit check added to JIT recursion. This issue was found by Karl
295        Skomski with a custom LLVM fuzzer.
296    
297    36. Yet another case similar to 27 above has been circumvented by an
298        unconditional allocation of extra memory. This issue is fixed "properly" in
299        PCRE2 by refactoring the way references are handled. Wen Guanxing
300        from Venustech ADLAB discovered this bug.
301    
302    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
303        with a custom LLVM fuzzer.
304    
305    38. Fixed a corner case of range optimization in JIT.
306    
307    39. An incorrect error "overran compiling workspace" was given if there were
308        exactly enough group forward references such that the last one extended
309        into the workspace safety margin. The next one would have expanded the
310        workspace. The test for overflow was not including the safety margin.
311    
312    40. A match limit issue is fixed in JIT which was found by Karl Skomski
313        with a custom LLVM fuzzer.
314    
315    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
316        work under Windows. (Why has it taken so long for anyone to notice?)
317    
318    42. In a character class such as [\W\p{Any}] where both a negative-type escape
319        ("not a word character") and a property escape were present, the property
320        escape was being ignored.
321    
322    43. Fix crash caused by very long (*MARK) or (*THEN) names.
323    
324    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
325        by a single ASCII character in a class item, was incorrectly compiled in
326        UCP mode. The POSIX class got lost, but only if the single character
327        followed it.
328    
329    45. [:punct:] in UCP mode was matching some characters in the range 128-255
330        that should not have been matched.
331    
332    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
333        class, all characters with code points greater than 255 are in the class.
334        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
335        such as \w are turned into Unicode properties), wide characters were not
336        correctly handled, and could fail to match.
337    
338    
339    Version 8.37 28-April-2015
340    --------------------------
341    
342    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
343        for those parentheses to be closed with whatever has been captured so far.
344        However, it was failing to mark any other groups between the hightest
345        capture so far and the currrent group as "unset". Thus, the ovector for
346        those groups contained whatever was previously there. An example is the
347        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
348    
349    2.  If an assertion condition was quantified with a minimum of zero (an odd
350        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
351    
352    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
353        unrecognized modifier, a crash could occur.
354    
355    4.  An attempt to do global matching in pcretest with a zero-length ovector
356        caused a crash.
357    
358    5.  Fixed a memory leak during matching that could occur for a subpattern
359        subroutine call (recursive or otherwise) if the number of captured groups
360        that had to be saved was greater than ten.
361    
362    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
363        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
364        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
365    
366    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
367        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
368    
369    8. When a pattern is compiled, it remembers the highest back reference so that
370       when matching, if the ovector is too small, extra memory can be obtained to
371       use instead. A conditional subpattern whose condition is a check on a
372       capture having happened, such as, for example in the pattern
373       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
374       setting the highest backreference number. This mattered only if pcre_exec()
375       was called with an ovector that was too small to hold the capture, and there
376       was no other kind of back reference (a situation which is probably quite
377       rare). The effect of the bug was that the condition was always treated as
378       FALSE when the capture could not be consulted, leading to a incorrect
379       behaviour by pcre_exec(). This bug has been fixed.
380    
381    9. A reference to a duplicated named group (either a back reference or a test
382       for being set in a conditional) that occurred in a part of the pattern where
383       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
384       to be incorrectly calculated, leading to overwriting.
385    
386    10. A mutually recursive set of back references such as (\2)(\1) caused a
387        segfault at study time (while trying to find the minimum matching length).
388        The infinite loop is now broken (with the minimum length unset, that is,
389        zero).
390    
391    11. If an assertion that was used as a condition was quantified with a minimum
392        of zero, matching went wrong. In particular, if the whole group had
393        unlimited repetition and could match an empty string, a segfault was
394        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
395        assertions to be quantified, but not if they are being used as conditions,
396        so the above pattern is faulted by Perl. PCRE has now been changed so that
397        it also rejects such patterns.
398    
399    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
400        failed to allow the zero-repeat case if pcre2_exec() was called with an
401        ovector too small to capture the group.
402    
403    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
404        Red Hat Product Security:
405    
406        (a) A crash if /K and /F were both set with the option to save the compiled
407        pattern.
408    
409        (b) Another crash if the option to print captured substrings in a callout
410        was combined with setting a null ovector, for example \O\C+ as a subject
411        string.
412    
413    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
414        forward reference repeated a large (but limited) number of times within a
415        repeated outer group that has a zero minimum quantifier, caused incorrect
416        code to be compiled, leading to the error "internal error:
417        previously-checked referenced subpattern not found" when an incorrect
418        memory address was read. This bug was reported as "heap overflow",
419        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
420        CVE-2015-2325.
421    
422    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
423        call within a group that also contained a recursive back reference caused
424        incorrect code to be compiled. This bug was reported as "heap overflow",
425        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
426        number CVE-2015-2326.
427    
428    24. Computing the size of the JIT read-only data in advance has been a source
429        of various issues, and new ones are still appear unfortunately. To fix
430        existing and future issues, size computation is eliminated from the code,
431        and replaced by on-demand memory allocation.
432    
433    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
434        adjacent to the end of the range, and the range contained characters with
435        more than one other case, caused incorrect behaviour when compiled in UTF
436        mode. In that example, the range a-j was left out of the class.
437    
438    26. Fix JIT compilation of conditional blocks, which assertion
439        is converted to (*FAIL). E.g: /(?(?!))/.
440    
441    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
442        discovered by the LLVM fuzzer.
443    
444    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
445        when this assertion was used as a condition, for example (?(?!)a|b). In
446        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
447        error about an unsupported item.
448    
449    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
450        possessification code could take exponential time to complete. A recursion
451        depth limit of 1000 has been imposed to limit the resources used by this
452        optimization.
453    
454    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
455        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
456        because \S ensures they are all in the class. The code for doing this was
457        interacting badly with the code for computing the amount of space needed to
458        compile the pattern, leading to a buffer overflow. This bug was discovered
459        by the LLVM fuzzer.
460    
461    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
462        other kinds of group caused stack overflow at compile time. This bug was
463        discovered by the LLVM fuzzer.
464    
465    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
466        between a subroutine call and its quantifier was incorrectly compiled,
467        leading to buffer overflow or other errors. This bug was discovered by the
468        LLVM fuzzer.
469    
470    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
471        assertion after (?(. The code was failing to check the character after
472        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
473        was discovered by the LLVM fuzzer.
474    
475    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
476        a fixed maximum following a group that contains a subroutine reference was
477        incorrectly compiled and could trigger buffer overflow. This bug was
478        discovered by the LLVM fuzzer.
479    
480    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
481        caused a stack overflow instead of the diagnosis of a non-fixed length
482        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
483    
484    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
485        (e.g. /(?<=\Ka)/) could make pcregrep loop.
486    
487    37. There was a similar problem to 36 in pcretest for global matches.
488    
489    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
490        and a subsequent item in the pattern caused a non-match, backtracking over
491        the repeated \X did not stop, but carried on past the start of the subject,
492        causing reference to random memory and/or a segfault. There were also some
493        other cases where backtracking after \C could crash. This set of bugs was
494        discovered by the LLVM fuzzer.
495    
496    39. The function for finding the minimum length of a matching string could take
497        a very long time if mutual recursion was present many times in a pattern,
498        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
499        method has been implemented. This infelicity was discovered by the LLVM
500        fuzzer.
501    
502    40. Static linking against the PCRE library using the pkg-config module was
503        failing on missing pthread symbols.
504    
505    
506    Version 8.36 26-September-2014
507    ------------------------------
508    
509    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
510        -Wmissing-field-initializers and -Wunused-parameter.
511    
512    2.  The tests for quantifiers being too big (greater than 65535) were being
513        applied after reading the number, and stupidly assuming that integer
514        overflow would give a negative number. The tests are now applied as the
515        numbers are read.
516    
517    3.  Tidy code in pcre_exec.c where two branches that used to be different are
518        now the same.
519    
520    4.  The JIT compiler did not generate match limit checks for certain
521        bracketed expressions with quantifiers. This may lead to exponential
522        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
523        issue should be resolved now.
524    
525    5.  Fixed an issue, which occures when nested alternatives are optimized
526        with table jumps.
527    
528    6.  Inserted two casts and changed some ints to size_t in the light of some
529        reported 64-bit compiler warnings (Bugzilla 1477).
530    
531    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
532        an empty string, which sometimes were behaving incorrectly in the
533        interpreter (though correctly in the JIT matcher). This pcretest input is
534        an example:
535    
536          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
537          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
538    
539        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
540        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
541        for an empty string was breaking the inner loop and carrying on at a lower
542        level, when possessive repeated groups should always return to a higher
543        level as they have no backtrack points in them. The empty string test now
544        occurs at the outer level.
545    
546    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
547        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
548    
549    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
550        doesn't).
551    
552    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
553        to hold a pattern to be incorrectly computed (too small) when there were
554        named back references to duplicated names. This could cause "internal
555        error: code overflow" or "double free or corruption" or other memory
556        handling errors.
557    
558    11. When named subpatterns had the same prefixes, back references could be
559        confused. For example, in this pattern:
560    
561          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
562    
563        the reference to 'Name' was incorrectly treated as a reference to a
564        duplicate name.
565    
566    12. A pattern such as /^s?c/mi8 where the optional character has more than
567        one "other case" was incorrectly compiled such that it would only try to
568        match starting at "c".
569    
570    13. When a pattern starting with \s was studied, VT was not included in the
571        list of possible starting characters; this should have been part of the
572        8.34/18 patch.
573    
574    14. If a character class started [\Qx]... where x is any character, the class
575        was incorrectly terminated at the ].
576    
577    15. If a pattern that started with a caseless match for a character with more
578        than one "other case" was studied, PCRE did not set up the starting code
579        unit bit map for the list of possible characters. Now it does. This is an
580        optimization improvement, not a bug fix.
581    
582    16. The Unicode data tables have been updated to Unicode 7.0.0.
583    
584    17. Fixed a number of memory leaks in pcregrep.
585    
586    18. Avoid a compiler warning (from some compilers) for a function call with
587        a cast that removes "const" from an lvalue by using an intermediate
588        variable (to which the compiler does not object).
589    
590    19. Incorrect code was compiled if a group that contained an internal recursive
591        back reference was optional (had quantifier with a minimum of zero). This
592        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
593        caused segmentation faults because of stack overflows at compile time.
594    
595    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
596        group that is quantified with an indefinite repeat, caused a compile-time
597        loop which used up all the system stack and provoked a segmentation fault.
598        This was not the same bug as 19 above.
599    
600    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
601        Patch by Mike Frysinger.
602    
603    
604    Version 8.35 04-April-2014
605    --------------------------
606    
607    1.  A new flag is set, when property checks are present in an XCLASS.
608        When this flag is not set, PCRE can perform certain optimizations
609        such as studying these XCLASS-es.
610    
611    2.  The auto-possessification of character sets were improved: a normal
612        and an extended character set can be compared now. Furthermore
613        the JIT compiler optimizes more character set checks.
614    
615    3.  Got rid of some compiler warnings for potentially uninitialized variables
616        that show up only when compiled with -O2.
617    
618    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
619        of a match later then the end of the match. The pcretest program was not
620        handling the case sensibly - it was outputting from the start to the next
621        binary zero. It now reports this situation in a message, and outputs the
622        text from the end to the start.
623    
624    5.  Fast forward search is improved in JIT. Instead of the first three
625        characters, any three characters with fixed position can be searched.
626        Search order: first, last, middle.
627    
628    6.  Improve character range checks in JIT. Characters are read by an inprecise
629        function now, which returns with an unknown value if the character code is
630        above a certain threshold (e.g: 256). The only limitation is that the value
631        must be bigger than the threshold as well. This function is useful when
632        the characters above the threshold are handled in the same way.
633    
634    7.  The macros whose names start with RAWUCHAR are placeholders for a future
635        mode in which only the bottom 21 bits of 32-bit data items are used. To
636        make this more memorable for those maintaining the code, the names have
637        been changed to start with UCHAR21, and an extensive comment has been added
638        to their definition.
639    
640    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
641        to the export list in Makefile.am (they were accidentally omitted from the
642        8.34 tarball).
643    
644    9.  The informational output from pcretest used the phrase "starting byte set"
645        which is inappropriate for the 16-bit and 32-bit libraries. As the output
646        for "first char" and "need char" really means "non-UTF-char", I've changed
647        "byte" to "char", and slightly reworded the output. The documentation about
648        these values has also been (I hope) clarified.
649    
650    10. Another JIT related optimization: use table jumps for selecting the correct
651        backtracking path, when more than four alternatives are present inside a
652        bracket.
653    
654    11. Empty match is not possible, when the minimum length is greater than zero,
655        and there is no \K in the pattern. JIT should avoid empty match checks in
656        such cases.
657    
658    12. In a caseless character class with UCP support, when a character with more
659        than one alternative case was not the first character of a range, not all
660        the alternative cases were added to the class. For example, s and \x{17f}
661        are both alternative cases for S: the class [RST] was handled correctly,
662        but [R-T] was not.
663    
664    13. The configure.ac file always checked for pthread support when JIT was
665        enabled. This is not used in Windows, so I have put this test inside a
666        check for the presence of windows.h (which was already tested for).
667    
668    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
669        The algorithm provides a way to skip certain starting offsets, and usually
670        faster than linear prefix searches.
671    
672    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
673        as for 'fr_FR' and 'french'. For some reason, however, it then used the
674        Windows-specific input and output files, which have 'french' screwed in.
675        So this could never have worked. One of the problems with locales is that
676        they aren't always the same. I have now updated RunTest so that it checks
677        the output of the locale test (test 3) against three different output
678        files, and it allows the test to pass if any one of them matches. With luck
679        this should make the test pass on some versions of Solaris where it was
680        failing. Because of the uncertainty, the script did not used to stop if
681        test 3 failed; it now does. If further versions of a French locale ever
682        come to light, they can now easily be added.
683    
684    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
685        there was a message during ./configure, but it did not stop. This now
686        provokes an error. The invalid example in README has been corrected.
687        If a value less than the minimum is given, the minimum value has always
688        been used, but now a warning is given.
689    
690    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
691        was a bug in the test system, which is now fixed. Also, the list of various
692        configurations that are tested for each release did not have one with both
693        16/32 bits and --enable-bar-anycrlf. It now does.
694    
695    18. pcretest was missing "-C bsr" for displaying the \R default setting.
696    
697    19. Little endian PowerPC systems are supported now by the JIT compiler.
698    
699    20. The fast forward newline mechanism could enter to an infinite loop on
700        certain invalid UTF-8 input. Although we don't support these cases
701        this issue can be fixed by a performance optimization.
702    
703    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
704        not take account if existing stack usage. There is now a new global
705        variable called pcre_stack_guard that can be set to point to an external
706        function to check stack availability. It is called at the start of
707        processing every parenthesized group.
708    
709    22. A typo in the code meant that in ungreedy mode the max/min qualifier
710        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
711        not match "ab".
712    
713    23. When UTF was disabled, the JIT program reported some incorrect compile
714        errors. These messages are silenced now.
715    
716    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
717        compiler.
718    
719    25. Change all the temporary files used in RunGrepTest to be different to those
720        used by RunTest so that the tests can be run simultaneously, for example by
721        "make -j check".
722    
723    
724    Version 8.34 15-December-2013
725    -----------------------------
726    
727    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
728        executable memory. Patch inspired by Carsten Klein.
729    
730    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
731        this macro is never tested and has no effect, because the work to support
732        coverage involves only compiling and linking options and special targets in
733        the Makefile. The comment in config.h implied that defining the macro would
734        enable coverage support, which is totally false. There was also support for
735        setting this macro in the CMake files (my fault, I just copied it from
736        configure). SUPPORT_GCOV has now been removed.
737    
738    3.  Make a small performance improvement in strlen16() and strlen32() in
739        pcretest.
740    
741    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
742        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
743        these cases). There was also one in pcretest.c.
744    
745    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
746    
747    6.  In UTF mode, the code for checking whether a group could match an empty
748        string (which is used for indefinitely repeated groups to allow for
749        breaking an infinite loop) was broken when the group contained a repeated
750        negated single-character class with a character that occupied more than one
751        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
752        UTF-8 mode). The effect was undefined: the group might or might not be
753        deemed as matching an empty string, or the program might have crashed.
754    
755    7.  The code for checking whether a group could match an empty string was not
756        recognizing that \h, \H, \v, \V, and \R must match a character.
757    
758    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
759        an empty string. If it can, pcretest shows this in its information output.
760    
761    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
762        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
763        matched by pcre_exec() without using JIT:
764    
765        (a) If the rest of the pattern did not match after a maximal run of
766            grapheme clusters, the code for backing up to try with fewer of them
767            did not always back up over a full grapheme when characters that do not
768            have the modifier quality were involved, e.g. Hangul syllables.
769    
770        (b) If the match point in a subject started with modifier character, and
771            there was no match, the code could incorrectly back up beyond the match
772            point, and potentially beyond the first character in the subject,
773            leading to a segfault or an incorrect match result.
774    
775    10. A conditional group with an assertion condition could lead to PCRE
776        recording an incorrect first data item for a match if no other first data
777        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
778        first data item, and therefore matched "ca" after "c" instead of at the
779        start.
780    
781    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
782        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
783    
784    12. The source of pcregrep now includes z/OS-specific code so that it can be
785        compiled for z/OS as part of the special z/OS distribution.
786    
787    13. Added the -T and -TM options to pcretest.
788    
789    14. The code in pcre_compile.c for creating the table of named capturing groups
790        has been refactored. Instead of creating the table dynamically during the
791        actual compiling pass, the information is remembered during the pre-compile
792        pass (on the stack unless there are more than 20 named groups, in which
793        case malloc() is used) and the whole table is created before the actual
794        compile happens. This has simplified the code (it is now nearly 150 lines
795        shorter) and prepared the way for better handling of references to groups
796        with duplicate names.
797    
798    15. A back reference to a named subpattern when there is more than one of the
799        same name now checks them in the order in which they appear in the pattern.
800        The first one that is set is used for the reference. Previously only the
801        first one was inspected. This change makes PCRE more compatible with Perl.
802    
803    16. Unicode character properties were updated from Unicode 6.3.0.
804    
805    17. The compile-time code for auto-possessification has been refactored, based
806        on a patch by Zoltan Herczeg. It now happens after instead of during
807        compilation. The code is cleaner, and more cases are handled. The option
808        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
809        options in pcretest are provided to set it. It can also be set by
810        (*NO_AUTO_POSSESS) at the start of a pattern.
811    
812    18. The character VT has been added to the default ("C" locale) set of
813        characters that match \s and are generally treated as white space,
814        following this same change in Perl 5.18. There is now no difference between
815        "Perl space" and "POSIX space". Whether VT is treated as white space in
816        other locales depends on the locale.
817    
818    19. The code for checking named groups as conditions, either for being set or
819        for being recursed, has been refactored (this is related to 14 and 15
820        above). Processing unduplicated named groups should now be as fast at
821        numerical groups, and processing duplicated groups should be faster than
822        before.
823    
824    20. Two patches to the CMake build system, by Alexander Barkov:
825    
826          (1) Replace the "source" command by "." in CMakeLists.txt because
827              "source" is a bash-ism.
828    
829          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
830              without these the CMake build does not work on Solaris.
831    
832    21. Perl has changed its handling of \8 and \9. If there is no previously
833        encountered capturing group of those numbers, they are treated as the
834        literal characters 8 and 9 instead of a binary zero followed by the
835        literals. PCRE now does the same.
836    
837    22. Following Perl, added \o{} to specify codepoints in octal, making it
838        possible to specify values greater than 0777 and also making them
839        unambiguous.
840    
841    23. Perl now gives an error for missing closing braces after \x{... instead of
842        treating the string as literal. PCRE now does the same.
843    
844    24. RunTest used to grumble if an inappropriate test was selected explicitly,
845        but just skip it when running all tests. This make it awkward to run ranges
846        of tests when one of them was inappropriate. Now it just skips any
847        inappropriate tests, as it always did when running all tests.
848    
849    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
850        character types such as \d or \w, too many callouts were inserted, and the
851        data that they returned was rubbish.
852    
853    26. In UCP mode, \s was not matching two of the characters that Perl matches,
854        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
855        were matched by \h. The code has now been refactored so that the lists of
856        the horizontal and vertical whitespace characters used for \h and \v (which
857        are defined only in one place) are now also used for \s.
858    
859    27. Add JIT support for the 64 bit TileGX architecture.
860        Patch by Jiong Wang (Tilera Corporation).
861    
862    28. Possessive quantifiers for classes (both explicit and automatically
863        generated) now use special opcodes instead of wrapping in ONCE brackets.
864    
865    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
866        (because it's meaningless), this was not happening when PCRE_CASELESS was
867        set. Not wrong, but inefficient.
868    
869    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
870        properties for \w, \d, etc) is present in a test regex. Otherwise if the
871        test contains no characters greater than 255, Perl doesn't realise it
872        should be using Unicode semantics.
873    
874    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
875        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
876        does in Unicode mode.
877    
878    32. Added the "forbid" facility to pcretest so that putting tests into the
879        wrong test files can sometimes be quickly detected.
880    
881    33. There is now a limit (default 250) on the depth of nesting of parentheses.
882        This limit is imposed to control the amount of system stack used at compile
883        time. It can be changed at build time by --with-parens-nest-limit=xxx or
884        the equivalent in CMake.
885    
886    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
887        errors. Perl warns for these when in warning mode, but PCRE has no facility
888        for giving warnings.
889    
890    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
891        However, this was not working for (?!) because it is optimized to (*FAIL),
892        for which PCRE does not allow quantifiers. The optimization is now disabled
893        when a quantifier follows (?!). I can't see any use for this, but it makes
894        things uniform.
895    
896    36. Perl no longer allows group names to start with digits, so I have made this
897        change also in PCRE. It simplifies the code a bit.
898    
899    37. In extended mode, Perl ignores spaces before a + that indicates a
900        possessive quantifier. PCRE allowed a space before the quantifier, but not
901        before the possessive +. It now does.
902    
903    38. The use of \K (reset reported match start) within a repeated possessive
904        group such as (a\Kb)*+ was not working.
905    
906    40. Document that the same character tables must be used at compile time and
907        run time, and that the facility to pass tables to pcre_exec() and
908        pcre_dfa_exec() is for use only with saved/restored patterns.
909    
910    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
911        features for Builds with MSVC:
912    
913        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
914           builds). The .rc files can be used to set FileDescription and many other
915           attributes.
916    
917        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
918           This allows higher-level build scripts which want .pdb files to avoid
919           hard-coding the exact files needed."
920    
921    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
922        mean "start of word" and "end of word", respectively, as a transition aid.
923    
924    43. A minimizing repeat of a class containing codepoints greater than 255 in
925        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
926        compiled to use the heap for recursion.
927    
928    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
929        is configured.
930    
931    
932    Version 8.33 28-May-2013
933    ------------------------
934    
935    1.  Added 'U' to some constants that are compared to unsigned integers, to
936        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
937        variables that are added to signed variables, to ensure the result is
938        signed and can be negated.
939    
940    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
941        CMake config files.
942    
943    3.  Revise the creation of config.h.generic so that all boolean macros are
944        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
945        overriding via -D on the command line possible.
946    
947    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
948        to unsigned int is reported to make a quite noticeable speed difference in
949        a specific Windows environment. Testing on Linux did also appear to show
950        some benefit (and it is clearly not harmful). Also fixed the definition of
951        Xop which should be unsigned.
952    
953    5.  Related to (4), changing the definition of the intermediate variable cc
954        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
955        improvements.
956    
957    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
958        unnecessary spaces.
959    
960    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
961        and later.
962    
963    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
964    
965    9.  Optimizing fast_forward_start_bits in JIT.
966    
967    10. Adding support for callouts in JIT, and fixing some issues revealed
968        during this work. Namely:
969    
970        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
971    
972        (b) Minimum length was not checked before the matching is started.
973    
974    11. The value of capture_last that is passed to callouts was incorrect in some
975        cases when there was a capture on one path that was subsequently abandoned
976        after a backtrack. Also, the capture_last value is now reset after a
977        recursion, since all captures are also reset in this case.
978    
979    12. The interpreter no longer returns the "too many substrings" error in the
980        case when an overflowing capture is in a branch that is subsequently
981        abandoned after a backtrack.
982    
983    13. In the pathological case when an offset vector of size 2 is used, pcretest
984        now prints out the matched string after a yield of 0 or 1.
985    
986    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
987        Only supported by the JIT compiler at the moment.
988    
989    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
990    
991    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
992        offset of the starting point of the matching process, provided the offsets
993        vector is large enough.
994    
995    17. The \A escape now records a lookbehind value of 1, though its execution
996        does not actually inspect the previous character. This is to ensure that,
997        in partial multi-segment matching, at least one character from the old
998        segment is retained when a new segment is processed. Otherwise, if there
999        are no lookbehinds in the pattern, \A might match incorrectly at the start
1000        of a new segment.
1001    
1002    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1003    
1004    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1005        gives some modest performance improvement in 8-bit mode.
1006    
1007    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1008        be expressed in certain programming languages using Universal Character
1009        Names.
1010    
1011    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1012        which points out that "non characters" are not "characters that may not
1013        appear in Unicode strings" but rather "characters that are reserved for
1014        internal use and have only local meaning".
1015    
1016    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1017        there was a conditional group that depended on an assertion, if the
1018        assertion was false, the callout that immediately followed the alternation
1019        in the condition was skipped when pcre_exec() was used for matching.
1020    
1021    23. Allow an explicit callout to be inserted before an assertion that is the
1022        condition for a conditional group, for compatibility with automatic
1023        callouts, which always insert a callout at this point.
1024    
1025    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1026        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1027    
1028    25. (*PRUNE) is now supported by the JIT compiler.
1029    
1030    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1031    
1032    27. Fix the case where there are two or more SKIPs with arguments that may be
1033        ignored.
1034    
1035    28. (*SKIP) is now supported by the JIT compiler.
1036    
1037    29. (*THEN) is now supported by the JIT compiler.
1038    
1039    30. Update RunTest with additional test selector options.
1040    
1041    31. The way PCRE handles backtracking verbs has been changed in two ways.
1042    
1043        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1044        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1045        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1046        rather obscure rules do not always do the same thing.
1047    
1048        (2) Previously, backtracking verbs were confined within assertions. This is
1049        no longer the case for positive assertions, except for (*ACCEPT). Again,
1050        this sometimes improves Perl compatibility, and sometimes does not.
1051    
1052    32. A number of tests that were in test 2 because Perl did things differently
1053        have been moved to test 1, because either Perl or PCRE has changed, and
1054        these tests are now compatible.
1055    
1056    32. Backtracking control verbs are now handled in the same way in JIT and
1057        interpreter.
1058    
1059    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1060        contained a forward subroutine reference caused a compile error.
1061    
1062    34. Auto-detect and optimize limited repetitions in JIT.
1063    
1064    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1065        blocking (*UTF) etc.
1066    
1067    36. In the interpreter, maximizing pattern repetitions for characters and
1068        character types now use tail recursion, which reduces stack usage.
1069    
1070    37. The value of the max lookbehind was not correctly preserved if a compiled
1071        and saved regex was reloaded on a host of different endianness.
1072    
1073    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1074        of the compiled pattern block, expand the flags field from 16 to 32 bits
1075        because it was almost full.
1076    
1077    39. Try madvise first before posix_madvise.
1078    
1079    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1080        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1081        original change locked out all matches of empty strings. This has been
1082        changed so that one match of an empty string per line is recognized.
1083        Subsequent searches on the same line (for colouring or for --only-matching,
1084        for example) do not recognize empty strings.
1085    
1086    41. Applied a user patch to fix a number of spelling mistakes in comments.
1087    
1088    42. Data lines longer than 65536 caused pcretest to crash.
1089    
1090    43. Clarified the data type for length and startoffset arguments for pcre_exec
1091        and pcre_dfa_exec in the function-specific man pages, where they were
1092        explicitly stated to be in bytes, never having been updated. I also added
1093        some clarification to the pcreapi man page.
1094    
1095    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1096        a segmentation fault.
1097    
1098    
1099    Version 8.32 30-November-2012
1100    -----------------------------
1101    
1102    1.  Improved JIT compiler optimizations for first character search and single
1103        character iterators.
1104    
1105    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1106        Patch by Daniel Richard G.
1107    
1108    3.  Single character iterator optimizations in the JIT compiler.
1109    
1110    4.  Improved JIT compiler optimizations for character ranges.
1111    
1112    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1113        Reported by Giuseppe D'Angelo.
1114    
1115    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1116        of a line, was being set incorrectly in cases where .* appeared inside
1117        atomic brackets at the start of a pattern, or where there was a subsequent
1118        *PRUNE or *SKIP.
1119    
1120    7.  Improved instruction cache flush for POWER/PowerPC.
1121        Patch by Daniel Richard G.
1122    
1123    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1124        grep:
1125    
1126        (a) There is now no limit to the number of patterns to be matched.
1127    
1128        (b) An error is given if a pattern is too long.
1129    
1130        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1131            are now supported.
1132    
1133        (d) --exclude-from and --include-from (multiple use) have been added.
1134    
1135        (e) Exclusions and inclusions now apply to all files and directories, not
1136            just to those obtained from scanning a directory recursively.
1137    
1138        (f) Multiple uses of -f and --file-list are now supported.
1139    
1140        (g) In a Windows environment, the default for -d has been changed from
1141            "read" (the GNU grep default) to "skip", because otherwise the presence
1142            of a directory in the file list provokes an error.
1143    
1144        (h) The documentation has been revised and clarified in places.
1145    
1146    9.  Improve the matching speed of capturing brackets.
1147    
1148    10. Changed the meaning of \X so that it now matches a Unicode extended
1149        grapheme cluster.
1150    
1151    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1152        out POSIX threads when JIT support is configured.
1153    
1154    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1155    
1156    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1157        structure could go wrong in environments where size_t is not the same size
1158        as int.
1159    
1160    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1161        set.
1162    
1163    15. The EBCDIC support had decayed; later updates to the code had included
1164        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1165        general tidy up of EBCDIC-related issues, and the documentation was also
1166        not quite right. There is now a test that can be run on ASCII systems to
1167        check some of the EBCDIC-related things (but is it not a full test).
1168    
1169    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1170        in a small tidy to the code.
1171    
1172    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1173    
1174    18. If the --only-matching (-o) option in pcregrep is specified multiple
1175        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1176        substrings matched by the 1st and 2nd capturing parentheses. A separating
1177        string can be specified by --om-separator (default empty).
1178    
1179    19. Improving the first n character searches.
1180    
1181    20. Turn case lists for horizontal and vertical white space into macros so that
1182        they are defined only once.
1183    
1184    21. This set of changes together give more compatible Unicode case-folding
1185        behaviour for characters that have more than one other case when UCP
1186        support is available.
1187    
1188        (a) The Unicode property table now has offsets into a new table of sets of
1189            three or more characters that are case-equivalent. The MultiStage2.py
1190            script that generates these tables (the pcre_ucd.c file) now scans
1191            CaseFolding.txt instead of UnicodeData.txt for character case
1192            information.
1193    
1194        (b) The code for adding characters or ranges of characters to a character
1195            class has been abstracted into a generalized function that also handles
1196            case-independence. In UTF-mode with UCP support, this uses the new data
1197            to handle characters with more than one other case.
1198    
1199        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1200            whose other case is greater than 256 are now correctly matched
1201            caselessly. Previously, the high codepoint matched the low one, but not
1202            vice versa.
1203    
1204        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1205            of the new class addition function, using character lists defined as
1206            macros alongside the case definitions of 20 above.
1207    
1208        (e) Caseless back references now work with characters that have more than
1209            one other case.
1210    
1211        (f) General caseless matching of characters with more than one other case
1212            is supported.
1213    
1214    22. Unicode character properties were updated from Unicode 6.2.0
1215    
1216    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1217    
1218    24. Add support for 32-bit character strings, and UTF-32
1219    
1220    25. Major JIT compiler update (code refactoring and bugfixing).
1221        Experimental Sparc 32 support is added.
1222    
1223    26. Applied a modified version of Daniel Richard G's patch to create
1224        pcre.h.generic and config.h.generic by "make" instead of in the
1225        PrepareRelease script.
1226    
1227    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1228        pcre_compile.c when checking for a zero character.
1229    
1230    28. Introducing a native interface for JIT. Through this interface, the compiled
1231        machine code can be directly executed. The purpose of this interface is to
1232        provide fast pattern matching, so several sanity checks are not performed.
1233        However, feature tests are still performed. The new interface provides
1234        1.4x speedup compared to the old one.
1235    
1236    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1237        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1238        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1239    
1240    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1241        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1242        the "old" RFC 2279). Instead, it ended up passing a negative length to
1243        pcre_exec().
1244    
1245    31. Add support for GCC's visibility feature to hide internal functions.
1246    
1247    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1248        "unknown -C option" after outputting 0 or 1.
1249    
1250    33. There is now support for generating a code coverage report for the test
1251        suite in environments where gcc is the compiler and lcov is installed. This
1252        is mainly for the benefit of the developers.
1253    
1254    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1255        unaddressable using valgrind annotations, allowing valgrind to detect
1256        invalid memory accesses. This is mainly for the benefit of the developers.
1257    
1258    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1259    
1260    26. Give configure error if --enable-cpp but no C++ compiler found.
1261    
1262    
1263    Version 8.31 06-July-2012
1264    -------------------------
1265    
1266    1.  Fixing a wrong JIT test case and some compiler warnings.
1267    
1268    2.  Removed a bashism from the RunTest script.
1269    
1270    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1271        to unsigned type, result still unsigned" that was given by an MS compiler
1272        on encountering the code "-sizeof(xxx)".
1273    
1274    4.  Partial matching support is added to the JIT compiler.
1275    
1276    5.  Fixed several bugs concerned with partial matching of items that consist
1277        of more than one character:
1278    
1279        (a) /^(..)\1/ did not partially match "aba" because checking references was
1280            done on an "all or nothing" basis. This also applied to repeated
1281            references.
1282    
1283        (b) \R did not give a hard partial match if \r was found at the end of the
1284            subject.
1285    
1286        (c) \X did not give a hard partial match after matching one or more
1287            characters at the end of the subject.
1288    
1289        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1290            a partial match for the string "\r".
1291    
1292        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1293            a partial match for a CR character at the end of the subject string.
1294    
1295    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1296        running pcretest, the text "(JIT)" added to the output whenever JIT is
1297        actually used to run the match.
1298    
1299    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1300        or /S+[+] with a digit between 1 and 7.
1301    
1302    8.  OP_NOT now supports any UTF character not just single-byte ones.
1303    
1304    9.  (*MARK) control verb is now supported by the JIT compiler.
1305    
1306    10. The command "./RunTest list" lists the available tests without actually
1307        running any of them. (Because I keep forgetting what they all are.)
1308    
1309    11. Add PCRE_INFO_MAXLOOKBEHIND.
1310    
1311    12. Applied a (slightly modified) user-supplied patch that improves performance
1312        when the heap is used for recursion (compiled with --disable-stack-for-
1313        recursion). Instead of malloc and free for each heap frame each time a
1314        logical recursion happens, frames are retained on a chain and re-used where
1315        possible. This sometimes gives as much as 30% improvement.
1316    
1317    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1318        call.
1319    
1320    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1321    
1322    15. It is now possible to link pcretest with libedit as an alternative to
1323        libreadline.
1324    
1325    16. (*COMMIT) control verb is now supported by the JIT compiler.
1326    
1327    17. The Unicode data tables have been updated to Unicode 6.1.0.
1328    
1329    18. Added --file-list option to pcregrep.
1330    
1331    19. Added binary file support to pcregrep, including the -a, --binary-files,
1332        -I, and --text options.
1333    
1334    20. The madvise function is renamed for posix_madvise for QNX compatibility
1335        reasons. Fixed by Giuseppe D'Angelo.
1336    
1337    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1338    
1339    22. Changed the option for creating symbolic links for 16-bit man pages from
1340        -s to -sf so that re-installing does not cause issues.
1341    
1342    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1343    
1344    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1345        matches in certain environments (the workspace was not being correctly
1346        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1347        some of the workspace data at the beginning of a restart.
1348    
1349    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1350        was not doing so when it should - probably a typo introduced by SVN 528
1351        (change 8.10/14).
1352    
1353    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1354        \w+ when the character tables indicated that \x{c4} was a word character.
1355        There were several related cases, all because the tests for doing a table
1356        lookup were testing for characters less than 127 instead of 255.
1357    
1358    27. If a pattern contains capturing parentheses that are not used in a match,
1359        their slots in the ovector are set to -1. For those that are higher than
1360        any matched groups, this happens at the end of processing. In the case when
1361        there were back references that the ovector was too small to contain
1362        (causing temporary malloc'd memory to be used during matching), and the
1363        highest capturing number was not used, memory off the end of the ovector
1364        was incorrectly being set to -1. (It was using the size of the temporary
1365        memory instead of the true size.)
1366    
1367    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1368        ovector size, it uses memory at the end of the block that it has got.
1369    
1370    29. Check for an overlong MARK name and give an error at compile time. The
1371        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1372    
1373    30. JIT compiler update.
1374    
1375    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1376        Rill for the patch.
1377    
1378    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1379    
1380    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1381    
1382    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1383        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1384        was enabled.
1385    
1386    35. Improve JIT code generation for greedy plus quantifier.
1387    
1388    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1389        1 to "aa" instead of to an empty string. The bug affected repeated groups
1390        that could potentially match an empty string.
1391    
1392    37. Optimizing single character iterators in JIT.
1393    
1394    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1395        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1396        codepoints that are too big for the mode are faulted, and in a UTF mode,
1397        disallowed codepoints are also faulted.
1398    
1399    39. If PCRE was compiled with UTF support, in three places in the DFA
1400        matcher there was code that should only have been obeyed in UTF mode, but
1401        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1402        processing when bytes with values greater than 127 were present. In 16-bit
1403        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1404        both cases the values are those that cannot be the first data item in a UTF
1405        character. The three items that might have provoked this were recursions,
1406        possessively repeated groups, and atomic groups.
1407    
1408    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1409        and pcregrep, because some OS require shared objects to be explicitly
1410        passed to ld, causing the link step to fail if they are not.
1411    
1412    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1413        mode, patterns that started with \h* or \R* might be incorrectly matched.
1414    
1415    
1416    Version 8.30 04-February-2012
1417    -----------------------------
1418    
1419    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1420        name is defined in ctype.h.
1421    
1422    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1423        only in quite long subpatterns.
1424    
1425    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1426        since it was replaced by pcre_fullinfo() in February 2000.
1427    
1428    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1429        match a (*MARK), and the match failed at the start of the subject, a
1430        reference to memory before the start of the subject could occur. This bug
1431        was introduced by fix 17 of release 8.21.
1432    
1433    5.  A reference to an unset group with zero minimum repetition was giving
1434        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1435        /(another)?(\1?)test/ matched against "hello world test". This bug was
1436        introduced in release 8.13.
1437    
1438    6.  Add support for 16-bit character strings (a large amount of work involving
1439        many changes and refactorings).
1440    
1441    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1442        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1443        from a file.
1444    
1445    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1446        rounding is not applied in this particular case).
1447    
1448    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1449        if they appear, or are escaped, in patterns.
1450    
1451    10. Get rid of a number of -Wunused-but-set-variable warnings.
1452    
1453    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1454        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1455        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1456        also returns the mark "x". This bug applied to capturing parentheses,
1457        non-capturing parentheses, and atomic parentheses. It also applied to some
1458        assertions.
1459    
1460    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1461        information out of configure.ac instead of relying on pcre.h.generic, which
1462        is not stored in the repository.
1463    
1464    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1465        -lreadline.
1466    
1467    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1468    
1469    15. Applied Graycode's patch to put the top-level frame on the stack rather
1470        than the heap when not using the stack for recursion. This gives a
1471        performance improvement in many cases when recursion is not deep.
1472    
1473    16. Experimental code added to "pcretest -C" to output the stack frame size.
1474    
1475    
1476    Version 8.21 12-Dec-2011
1477    ------------------------
1478    
1479    1.  Updating the JIT compiler.
1480    
1481    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1482        are added as well.
1483    
1484    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1485        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1486        calling _pcre_jit_exec. Some extra comments are added.
1487    
1488    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1489        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1490        was introduced by change 18 for 8.20.
1491    
1492    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1493        ECMA-262 standard.
1494    
1495    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1496        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1497        This bug was probably introduced by change 9 of 8.13.
1498    
1499    7.  While fixing 6 above, I noticed that a number of other items were being
1500        incorrectly rejected as "not fixed length". This arose partly because newer
1501        opcodes had not been added to the fixed-length checking code. I have (a)
1502        corrected the bug and added tests for these items, and (b) arranged for an
1503        error to occur if an unknown opcode is encountered while checking for fixed
1504        length instead of just assuming "not fixed length". The items that were
1505        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1506        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1507        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1508    
1509    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1510        being incorrectly compiled and would have given unpredicatble results.
1511    
1512    9.  A possessively repeated subpattern with minimum repeat count greater than
1513        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1514        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1515        the first (A) could occur when it should not.
1516    
1517    10. Add a cast and remove a redundant test from the code.
1518    
1519    11. JIT should use pcre_malloc/pcre_free for allocation.
1520    
1521    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1522        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1523        is anything other than /usr, -L is still shown).
1524    
1525    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1526    
1527    14. Perl does not support \N without a following name in a [] class; PCRE now
1528        also gives an error.
1529    
1530    15. If a forward reference was repeated with an upper limit of around 2000,
1531        it caused the error "internal error: overran compiling workspace". The
1532        maximum number of forward references (including repeats) was limited by the
1533        internal workspace, and dependent on the LINK_SIZE. The code has been
1534        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1535        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1536        of around 200,000 forward references. While doing this, I also speeded up
1537        the filling in of repeated forward references.
1538    
1539    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1540        incorrectly expecting the subject to contain another "a" after the start.
1541    
1542    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1543        in the match, the SKIP should be ignored. This was not happening; instead
1544        the SKIP was being treated as NOMATCH. For patterns such as
1545        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1546        tested.
1547    
1548    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1549        now much more compatible with Perl, in particular in cases where the result
1550        is a non-match for a non-anchored pattern. For example, if
1551        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1552        "m", where previously it did not return a name. A side effect of this
1553        change is that for partial matches, the last encountered mark name is
1554        returned, as for non matches. A number of tests that were previously not
1555        Perl-compatible have been moved into the Perl-compatible test files. The
1556        refactoring has had the pleasing side effect of removing one argument from
1557        the match() function, thus reducing its stack requirements.
1558    
1559    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1560        subsequent uses of /S (without +) incorrectly behaved like /S+.
1561    
1562    21. Retrieve executable code size support for the JIT compiler and fixing
1563        some warnings.
1564    
1565    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1566        not work when the shorter character appeared right at the end of the
1567        subject string.
1568    
1569    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1570        systems.
1571    
1572    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1573        output it when the /M option is used in pcretest.
1574    
1575    25. The CheckMan script was not being included in the distribution. Also, added
1576        an explicit "perl" to run Perl scripts from the PrepareRelease script
1577        because this is reportedly needed in Windows.
1578    
1579    26. If study data was being save in a file and studying had not found a set of
1580        "starts with" bytes for the pattern, the data written to the file (though
1581        never used) was taken from uninitialized memory and so caused valgrind to
1582        complain.
1583    
1584    27. Updated RunTest.bat as provided by Sheri Pierce.
1585    
1586    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1587    
1588    29. Computation of memory usage for the table of capturing group names was
1589        giving an unnecessarily large value.
1590    
1591    
1592    Version 8.20 21-Oct-2011
1593    ------------------------
1594    
1595    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1596        a POSIX class. After further experiments with Perl, which convinced me that
1597        Perl has bugs and confusions, a closing square bracket is no longer allowed
1598        in a POSIX name. This bug also affected patterns with classes that started
1599        with full stops.
1600    
1601    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1602        captured substring, but while checking the failing first alternative,
1603        substring 1 is temporarily captured. If the output vector supplied to
1604        pcre_exec() was not big enough for this capture, the yield of the function
1605        was still zero ("insufficient space for captured substrings"). This cannot
1606        be totally fixed without adding another stack variable, which seems a lot
1607        of expense for a edge case. However, I have improved the situation in cases
1608        such as /(a)(b)x|abc/ matched against "abc", where the return code
1609        indicates that fewer than the maximum number of slots in the ovector have
1610        been set.
1611    
1612    3.  Related to (2) above: when there are more back references in a pattern than
1613        slots in the output vector, pcre_exec() uses temporary memory during
1614        matching, and copies in the captures as far as possible afterwards. It was
1615        using the entire output vector, but this conflicts with the specification
1616        that only 2/3 is used for passing back captured substrings. Now it uses
1617        only the first 2/3, for compatibility. This is, of course, another edge
1618        case.
1619    
1620    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1621        main code base, and can be used by building with --enable-jit. When this is
1622        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1623        runtime --no-jit option is given.
1624    
1625    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1626        ovector, the return from the function was zero, implying that there were
1627        other matches that did not fit. The correct "exactly full" value is now
1628        returned.
1629    
1630    6.  If a subpattern that was called recursively or as a subroutine contained
1631        (*PRUNE) or any other control that caused it to give a non-standard return,
1632        invalid errors such as "Error -26 (nested recursion at the same subject
1633        position)" or even infinite loops could occur.
1634    
1635    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1636        computing the minimum length on reaching *ACCEPT, and so ended up with the
1637        wrong value of 1 rather than 0. Further investigation indicates that
1638        computing a minimum subject length in the presence of *ACCEPT is difficult
1639        (think back references, subroutine calls), and so I have changed the code
1640        so that no minimum is registered for a pattern that contains *ACCEPT.
1641    
1642    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1643        it was not handled as intended. [But see 16 below.]
1644    
1645    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1646        Sheri Pierce.
1647    
1648    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1649        the first byte in a match must be "a".
1650    
1651    11. Change 17 for 8.13 increased the recursion depth for patterns like
1652        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1653        pattern contains any instances of (*THEN). If it does not, the old
1654        optimizations are restored. It would be nice to do this on a per-group
1655        basis, but at the moment that is not feasible.
1656    
1657    12. In some environments, the output of pcretest -C is CRLF terminated. This
1658        broke RunTest's code that checks for the link size. A single white space
1659        character after the value is now allowed for.
1660    
1661    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1662        For "fr", it uses the Windows-specific input and output files.
1663    
1664    14. If (*THEN) appeared in a group that was called recursively or as a
1665        subroutine, it did not work as intended. [But see next item.]
1666    
1667    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1668        pattern fragments (but not containing any | characters). If A and B are
1669        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1670        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1671        D. In other words, Perl considers parentheses that do not contain any |
1672        characters to be part of a surrounding alternative, whereas PCRE was
1673        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1674        differently. PCRE now behaves in the same way as Perl, except in the case
1675        of subroutine/recursion calls such as (?1) which have in any case always
1676        been different (but PCRE had them first :-).
1677    
1678    16. Related to 15 above: Perl does not treat the | in a conditional group as
1679        creating alternatives. Such a group is treated in the same way as an
1680        ordinary group without any | characters when processing (*THEN). PCRE has
1681        been changed to match Perl's behaviour.
1682    
1683    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1684        RunGrepTest script failed.
1685    
1686    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1687        inevitable for groups that contain captures, but it can lead to a lot of
1688        stack use in large patterns. The old behaviour has been restored for atomic
1689        groups that do not contain any capturing parentheses.
1690    
1691    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1692        suppress the check for a minimum subject length at run time. (If it was
1693        given to pcre_exec() or pcre_dfa_exec() it did work.)
1694    
1695    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1696        fail to work when decoding hex characters in data strings in EBCDIC
1697        environments.
1698    
1699    21. It appears that in at least one Mac OS environment, the isxdigit() function
1700        is implemented as a macro that evaluates to its argument more than once,
1701        contravening the C 90 Standard (I haven't checked a later standard). There
1702        was an instance in pcretest which caused it to go wrong when processing
1703        \x{...} escapes in subject strings. The has been rewritten to avoid using
1704        things like p++ in the argument of isxdigit().
1705    
1706    
1707    Version 8.13 16-Aug-2011
1708  ------------------------  ------------------------
1709    
1710  1.  The Unicode data tables have been updated to Unicode 6.0.0.  1.  The Unicode data tables have been updated to Unicode 6.0.0.
# Line 98  Version 8.13 02-Aug-2011 Line 1801  Version 8.13 02-Aug-2011
1801      tail recursion to cut down on stack usage. Unfortunately, now that there is      tail recursion to cut down on stack usage. Unfortunately, now that there is
1802      the possibility of (*THEN) occurring in these branches, tail recursion is      the possibility of (*THEN) occurring in these branches, tail recursion is
1803      no longer possible because the return has to be checked for (*THEN). These      no longer possible because the return has to be checked for (*THEN). These
1804      two optimizations have therefore been removed.      two optimizations have therefore been removed. [But see 8.20/11 above.]
1805    
1806  18. If a pattern containing \R was studied, it was assumed that \R always  18. If a pattern containing \R was studied, it was assumed that \R always
1807      matched two bytes, thus causing the minimum subject length to be      matched two bytes, thus causing the minimum subject length to be
# Line 189  Version 8.13 02-Aug-2011 Line 1892  Version 8.13 02-Aug-2011
1892      For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,      For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1893      unescaped square brackets may also appear as part of class names. For      unescaped square brackets may also appear as part of class names. For
1894      example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves      example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1895      more like Perl.      more like Perl. (But see 8.20/1 above.)
1896    
1897  38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this  38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1898      was because it thought it was \N{name}, which is not supported).      was because it thought it was \N{name}, which is not supported).
# Line 224  Version 8.13 02-Aug-2011 Line 1927  Version 8.13 02-Aug-2011
1927      (a) The default value of the buffer size parameter has been increased from      (a) The default value of the buffer size parameter has been increased from
1928          8K to 20K. (The actual buffer used is three times this size.)          8K to 20K. (The actual buffer used is three times this size.)
1929    
1930      (b) The default can be changed by ./configure --with-pcregrep-bufsiz when      (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1931          PCRE is built.          PCRE is built.
1932    
1933      (c) A --buffer-size=n option has been added to pcregrep, to allow the size      (c) A --buffer-size=n option has been added to pcregrep, to allow the size
# Line 245  Version 8.13 02-Aug-2011 Line 1948  Version 8.13 02-Aug-2011
1948  47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a  47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1949      complete match instead of a partial match. This bug was dependent on both      complete match instead of a partial match. This bug was dependent on both
1950      the PCRE_UTF8 and PCRE_DOTALL options being set.      the PCRE_UTF8 and PCRE_DOTALL options being set.
1951    
1952  48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the  48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1953      starting byte set, because \b was not being ignored.      starting byte set, because \b was not being ignored.
1954    
1955    
1956  Version 8.12 15-Jan-2011  Version 8.12 15-Jan-2011
# Line 900  Version 7.9 11-Apr-09 Line 2603  Version 7.9 11-Apr-09
2603  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
2604      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
2605      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2606      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2607        8.33.]
2608    
2609  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
2610      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 1143  Version 7.7 07-May-08 Line 2847  Version 7.7 07-May-08
2847      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2848      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2849      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2850      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2851      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2852      was subsequently set up correctly.)      was subsequently set up correctly.)
2853    

Legend:
Removed from v.657  
changed lines
  Added in v.1684

  ViewVC Help
Powered by ViewVC 1.1.5