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

Diff of /code/trunk/ChangeLog

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

revision 491 by ph10, Mon Mar 1 17:45:08 2010 UTC revision 1615 by ph10, Sun Nov 29 17:46:23 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.02 01-Mar-2010  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 8.39 xx-xxxxxx-201x
8    ---------------------------
9    
10    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12        misbehaved. This bug was found by the LLVM fuzzer.
13    
14    2.  Similar to the above, if an isolated \E was present between an item and its
15        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16        bug was found by the LLVM fuzzer.
17    
18    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19        working correctly in UCP mode.
20    
21    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22        was set when the pmatch argument was NULL. It now returns REG_INVARG.
23    
24    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25    
26    
27    Version 8.38 23-November-2015
28    -----------------------------
29    
30    1.  If a group that contained a recursive back reference also contained a
31        forward reference subroutine call followed by a non-forward-reference
32        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
33        compile correct code, leading to undefined behaviour or an internally
34        detected error. This bug was discovered by the LLVM fuzzer.
35    
36    2.  Quantification of certain items (e.g. atomic back references) could cause
37        incorrect code to be compiled when recursive forward references were
38        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
39        This bug was discovered by the LLVM fuzzer.
40    
41    3.  A repeated conditional group whose condition was a reference by name caused
42        a buffer overflow if there was more than one group with the given name.
43        This bug was discovered by the LLVM fuzzer.
44    
45    4.  A recursive back reference by name within a group that had the same name as
46        another group caused a buffer overflow. For example:
47        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
48    
49    5.  A forward reference by name to a group whose number is the same as the
50        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
51        a buffer overflow at compile time. This bug was discovered by the LLVM
52        fuzzer.
53    
54    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
55        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
56    
57    7.  Another buffer overflow bug involved duplicate named groups with a
58        reference between their definition, with a group that reset capture
59        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
60        fixed by always allowing for more memory, even if not needed. (A proper fix
61        is implemented in PCRE2, but it involves more refactoring.)
62    
63    8.  There was no check for integer overflow in subroutine calls such as (?123).
64    
65    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
66        being treated as a literal 'l' instead of causing an error.
67    
68    10. There was a buffer overflow if pcre_exec() was called with an ovector of
69        size 1. This bug was found by american fuzzy lop.
70    
71    11. If a non-capturing group containing a conditional group that could match
72        an empty string was repeated, it was not identified as matching an empty
73        string itself. For example: /^(?:(?(1)x|)+)+$()/.
74    
75    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
76        \a and \e in test subject lines.
77    
78    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
79        instead of the EBCDIC value.
80    
81    14. The handling of \c in an EBCDIC environment has been revised so that it is
82        now compatible with the specification in Perl's perlebcdic page.
83    
84    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
85        ASCII/Unicode. This has now been added to the list of characters that are
86        recognized as white space in EBCDIC.
87    
88    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
89        error (correctly) when used outside a class, but did not give an error
90        within a class.
91    
92    17. \h within a class was incorrectly compiled in EBCDIC environments.
93    
94    18. A pattern with an unmatched closing parenthesis that contained a backward
95        assertion which itself contained a forward reference caused buffer
96        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
97    
98    19. JIT should return with error when the compiled pattern requires more stack
99        space than the maximum.
100    
101    20. A possessively repeated conditional group that could match an empty string,
102        for example, /(?(R))*+/, was incorrectly compiled.
103    
104    21. Fix infinite recursion in the JIT compiler when certain patterns such as
105        /(?:|a|){100}x/ are analysed.
106    
107    22. Some patterns with character classes involving [: and \\ were incorrectly
108        compiled and could cause reading from uninitialized memory or an incorrect
109        error diagnosis.
110    
111    23. Pathological patterns containing many nested occurrences of [: caused
112        pcre_compile() to run for a very long time.
113    
114    24. A conditional group with only one branch has an implicit empty alternative
115        branch and must therefore be treated as potentially matching an empty
116        string.
117    
118    25. If (?R was followed by - or + incorrect behaviour happened instead of a
119        diagnostic.
120    
121    26. Arrange to give up on finding the minimum matching length for overly
122        complex patterns.
123    
124    27. Similar to (4) above: in a pattern with duplicated named groups and an
125        occurrence of (?| it is possible for an apparently non-recursive back
126        reference to become recursive if a later named group with the relevant
127        number is encountered. This could lead to a buffer overflow. Wen Guanxing
128        from Venustech ADLAB discovered this bug.
129    
130    28. If pcregrep was given the -q option with -c or -l, or when handling a
131        binary file, it incorrectly wrote output to stdout.
132    
133    29. The JIT compiler did not restore the control verb head in case of *THEN
134        control verbs. This issue was found by Karl Skomski with a custom LLVM
135        fuzzer.
136    
137    30. Error messages for syntax errors following \g and \k were giving inaccurate
138        offsets in the pattern.
139    
140    31. Added a check for integer overflow in conditions (?(<digits>) and
141        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
142        fuzzer.
143    
144    32. Handling recursive references such as (?2) when the reference is to a group
145        later in the pattern uses code that is very hacked about and error-prone.
146        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
147        give an internal error if it is obvious that compiling has gone wrong.
148    
149    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
150        This issue was found by Karl Skomski with a custom LLVM fuzzer.
151    
152    34. The JIT compiler should restore the control chain for empty possessive
153        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
154    
155    35. Match limit check added to JIT recursion. This issue was found by Karl
156        Skomski with a custom LLVM fuzzer.
157    
158    36. Yet another case similar to 27 above has been circumvented by an
159        unconditional allocation of extra memory. This issue is fixed "properly" in
160        PCRE2 by refactoring the way references are handled. Wen Guanxing
161        from Venustech ADLAB discovered this bug.
162    
163    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
164        with a custom LLVM fuzzer.
165    
166    38. Fixed a corner case of range optimization in JIT.
167    
168    39. An incorrect error "overran compiling workspace" was given if there were
169        exactly enough group forward references such that the last one extended
170        into the workspace safety margin. The next one would have expanded the
171        workspace. The test for overflow was not including the safety margin.
172    
173    40. A match limit issue is fixed in JIT which was found by Karl Skomski
174        with a custom LLVM fuzzer.
175    
176    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
177        work under Windows. (Why has it taken so long for anyone to notice?)
178    
179    42. In a character class such as [\W\p{Any}] where both a negative-type escape
180        ("not a word character") and a property escape were present, the property
181        escape was being ignored.
182    
183    43. Fix crash caused by very long (*MARK) or (*THEN) names.
184    
185    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
186        by a single ASCII character in a class item, was incorrectly compiled in
187        UCP mode. The POSIX class got lost, but only if the single character
188        followed it.
189    
190    45. [:punct:] in UCP mode was matching some characters in the range 128-255
191        that should not have been matched.
192    
193    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
194        class, all characters with code points greater than 255 are in the class.
195        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
196        such as \w are turned into Unicode properties), wide characters were not
197        correctly handled, and could fail to match.
198    
199    
200    Version 8.37 28-April-2015
201    --------------------------
202    
203    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
204        for those parentheses to be closed with whatever has been captured so far.
205        However, it was failing to mark any other groups between the hightest
206        capture so far and the currrent group as "unset". Thus, the ovector for
207        those groups contained whatever was previously there. An example is the
208        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
209    
210    2.  If an assertion condition was quantified with a minimum of zero (an odd
211        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
212    
213    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
214        unrecognized modifier, a crash could occur.
215    
216    4.  An attempt to do global matching in pcretest with a zero-length ovector
217        caused a crash.
218    
219    5.  Fixed a memory leak during matching that could occur for a subpattern
220        subroutine call (recursive or otherwise) if the number of captured groups
221        that had to be saved was greater than ten.
222    
223    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
224        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
225        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
226    
227    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
228        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
229    
230    8. When a pattern is compiled, it remembers the highest back reference so that
231       when matching, if the ovector is too small, extra memory can be obtained to
232       use instead. A conditional subpattern whose condition is a check on a
233       capture having happened, such as, for example in the pattern
234       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
235       setting the highest backreference number. This mattered only if pcre_exec()
236       was called with an ovector that was too small to hold the capture, and there
237       was no other kind of back reference (a situation which is probably quite
238       rare). The effect of the bug was that the condition was always treated as
239       FALSE when the capture could not be consulted, leading to a incorrect
240       behaviour by pcre_exec(). This bug has been fixed.
241    
242    9. A reference to a duplicated named group (either a back reference or a test
243       for being set in a conditional) that occurred in a part of the pattern where
244       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
245       to be incorrectly calculated, leading to overwriting.
246    
247    10. A mutually recursive set of back references such as (\2)(\1) caused a
248        segfault at study time (while trying to find the minimum matching length).
249        The infinite loop is now broken (with the minimum length unset, that is,
250        zero).
251    
252    11. If an assertion that was used as a condition was quantified with a minimum
253        of zero, matching went wrong. In particular, if the whole group had
254        unlimited repetition and could match an empty string, a segfault was
255        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
256        assertions to be quantified, but not if they are being used as conditions,
257        so the above pattern is faulted by Perl. PCRE has now been changed so that
258        it also rejects such patterns.
259    
260    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
261        failed to allow the zero-repeat case if pcre2_exec() was called with an
262        ovector too small to capture the group.
263    
264    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
265        Red Hat Product Security:
266    
267        (a) A crash if /K and /F were both set with the option to save the compiled
268        pattern.
269    
270        (b) Another crash if the option to print captured substrings in a callout
271        was combined with setting a null ovector, for example \O\C+ as a subject
272        string.
273    
274    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
275        forward reference repeated a large (but limited) number of times within a
276        repeated outer group that has a zero minimum quantifier, caused incorrect
277        code to be compiled, leading to the error "internal error:
278        previously-checked referenced subpattern not found" when an incorrect
279        memory address was read. This bug was reported as "heap overflow",
280        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
281        CVE-2015-2325.
282    
283    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
284        call within a group that also contained a recursive back reference caused
285        incorrect code to be compiled. This bug was reported as "heap overflow",
286        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
287        number CVE-2015-2326.
288    
289    24. Computing the size of the JIT read-only data in advance has been a source
290        of various issues, and new ones are still appear unfortunately. To fix
291        existing and future issues, size computation is eliminated from the code,
292        and replaced by on-demand memory allocation.
293    
294    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
295        adjacent to the end of the range, and the range contained characters with
296        more than one other case, caused incorrect behaviour when compiled in UTF
297        mode. In that example, the range a-j was left out of the class.
298    
299    26. Fix JIT compilation of conditional blocks, which assertion
300        is converted to (*FAIL). E.g: /(?(?!))/.
301    
302    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
303        discovered by the LLVM fuzzer.
304    
305    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
306        when this assertion was used as a condition, for example (?(?!)a|b). In
307        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
308        error about an unsupported item.
309    
310    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
311        possessification code could take exponential time to complete. A recursion
312        depth limit of 1000 has been imposed to limit the resources used by this
313        optimization.
314    
315    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
316        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
317        because \S ensures they are all in the class. The code for doing this was
318        interacting badly with the code for computing the amount of space needed to
319        compile the pattern, leading to a buffer overflow. This bug was discovered
320        by the LLVM fuzzer.
321    
322    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
323        other kinds of group caused stack overflow at compile time. This bug was
324        discovered by the LLVM fuzzer.
325    
326    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
327        between a subroutine call and its quantifier was incorrectly compiled,
328        leading to buffer overflow or other errors. This bug was discovered by the
329        LLVM fuzzer.
330    
331    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
332        assertion after (?(. The code was failing to check the character after
333        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
334        was discovered by the LLVM fuzzer.
335    
336    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
337        a fixed maximum following a group that contains a subroutine reference was
338        incorrectly compiled and could trigger buffer overflow. This bug was
339        discovered by the LLVM fuzzer.
340    
341    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
342        caused a stack overflow instead of the diagnosis of a non-fixed length
343        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
344    
345    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
346        (e.g. /(?<=\Ka)/) could make pcregrep loop.
347    
348    37. There was a similar problem to 36 in pcretest for global matches.
349    
350    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
351        and a subsequent item in the pattern caused a non-match, backtracking over
352        the repeated \X did not stop, but carried on past the start of the subject,
353        causing reference to random memory and/or a segfault. There were also some
354        other cases where backtracking after \C could crash. This set of bugs was
355        discovered by the LLVM fuzzer.
356    
357    39. The function for finding the minimum length of a matching string could take
358        a very long time if mutual recursion was present many times in a pattern,
359        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
360        method has been implemented. This infelicity was discovered by the LLVM
361        fuzzer.
362    
363    40. Static linking against the PCRE library using the pkg-config module was
364        failing on missing pthread symbols.
365    
366    
367    Version 8.36 26-September-2014
368    ------------------------------
369    
370    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
371        -Wmissing-field-initializers and -Wunused-parameter.
372    
373    2.  The tests for quantifiers being too big (greater than 65535) were being
374        applied after reading the number, and stupidly assuming that integer
375        overflow would give a negative number. The tests are now applied as the
376        numbers are read.
377    
378    3.  Tidy code in pcre_exec.c where two branches that used to be different are
379        now the same.
380    
381    4.  The JIT compiler did not generate match limit checks for certain
382        bracketed expressions with quantifiers. This may lead to exponential
383        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
384        issue should be resolved now.
385    
386    5.  Fixed an issue, which occures when nested alternatives are optimized
387        with table jumps.
388    
389    6.  Inserted two casts and changed some ints to size_t in the light of some
390        reported 64-bit compiler warnings (Bugzilla 1477).
391    
392    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
393        an empty string, which sometimes were behaving incorrectly in the
394        interpreter (though correctly in the JIT matcher). This pcretest input is
395        an example:
396    
397          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
398          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
399    
400        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
401        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
402        for an empty string was breaking the inner loop and carrying on at a lower
403        level, when possessive repeated groups should always return to a higher
404        level as they have no backtrack points in them. The empty string test now
405        occurs at the outer level.
406    
407    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
408        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
409    
410    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
411        doesn't).
412    
413    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
414        to hold a pattern to be incorrectly computed (too small) when there were
415        named back references to duplicated names. This could cause "internal
416        error: code overflow" or "double free or corruption" or other memory
417        handling errors.
418    
419    11. When named subpatterns had the same prefixes, back references could be
420        confused. For example, in this pattern:
421    
422          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
423    
424        the reference to 'Name' was incorrectly treated as a reference to a
425        duplicate name.
426    
427    12. A pattern such as /^s?c/mi8 where the optional character has more than
428        one "other case" was incorrectly compiled such that it would only try to
429        match starting at "c".
430    
431    13. When a pattern starting with \s was studied, VT was not included in the
432        list of possible starting characters; this should have been part of the
433        8.34/18 patch.
434    
435    14. If a character class started [\Qx]... where x is any character, the class
436        was incorrectly terminated at the ].
437    
438    15. If a pattern that started with a caseless match for a character with more
439        than one "other case" was studied, PCRE did not set up the starting code
440        unit bit map for the list of possible characters. Now it does. This is an
441        optimization improvement, not a bug fix.
442    
443    16. The Unicode data tables have been updated to Unicode 7.0.0.
444    
445    17. Fixed a number of memory leaks in pcregrep.
446    
447    18. Avoid a compiler warning (from some compilers) for a function call with
448        a cast that removes "const" from an lvalue by using an intermediate
449        variable (to which the compiler does not object).
450    
451    19. Incorrect code was compiled if a group that contained an internal recursive
452        back reference was optional (had quantifier with a minimum of zero). This
453        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
454        caused segmentation faults because of stack overflows at compile time.
455    
456    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
457        group that is quantified with an indefinite repeat, caused a compile-time
458        loop which used up all the system stack and provoked a segmentation fault.
459        This was not the same bug as 19 above.
460    
461    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
462        Patch by Mike Frysinger.
463    
464    
465    Version 8.35 04-April-2014
466    --------------------------
467    
468    1.  A new flag is set, when property checks are present in an XCLASS.
469        When this flag is not set, PCRE can perform certain optimizations
470        such as studying these XCLASS-es.
471    
472    2.  The auto-possessification of character sets were improved: a normal
473        and an extended character set can be compared now. Furthermore
474        the JIT compiler optimizes more character set checks.
475    
476    3.  Got rid of some compiler warnings for potentially uninitialized variables
477        that show up only when compiled with -O2.
478    
479    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
480        of a match later then the end of the match. The pcretest program was not
481        handling the case sensibly - it was outputting from the start to the next
482        binary zero. It now reports this situation in a message, and outputs the
483        text from the end to the start.
484    
485    5.  Fast forward search is improved in JIT. Instead of the first three
486        characters, any three characters with fixed position can be searched.
487        Search order: first, last, middle.
488    
489    6.  Improve character range checks in JIT. Characters are read by an inprecise
490        function now, which returns with an unknown value if the character code is
491        above a certain threshold (e.g: 256). The only limitation is that the value
492        must be bigger than the threshold as well. This function is useful when
493        the characters above the threshold are handled in the same way.
494    
495    7.  The macros whose names start with RAWUCHAR are placeholders for a future
496        mode in which only the bottom 21 bits of 32-bit data items are used. To
497        make this more memorable for those maintaining the code, the names have
498        been changed to start with UCHAR21, and an extensive comment has been added
499        to their definition.
500    
501    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
502        to the export list in Makefile.am (they were accidentally omitted from the
503        8.34 tarball).
504    
505    9.  The informational output from pcretest used the phrase "starting byte set"
506        which is inappropriate for the 16-bit and 32-bit libraries. As the output
507        for "first char" and "need char" really means "non-UTF-char", I've changed
508        "byte" to "char", and slightly reworded the output. The documentation about
509        these values has also been (I hope) clarified.
510    
511    10. Another JIT related optimization: use table jumps for selecting the correct
512        backtracking path, when more than four alternatives are present inside a
513        bracket.
514    
515    11. Empty match is not possible, when the minimum length is greater than zero,
516        and there is no \K in the pattern. JIT should avoid empty match checks in
517        such cases.
518    
519    12. In a caseless character class with UCP support, when a character with more
520        than one alternative case was not the first character of a range, not all
521        the alternative cases were added to the class. For example, s and \x{17f}
522        are both alternative cases for S: the class [RST] was handled correctly,
523        but [R-T] was not.
524    
525    13. The configure.ac file always checked for pthread support when JIT was
526        enabled. This is not used in Windows, so I have put this test inside a
527        check for the presence of windows.h (which was already tested for).
528    
529    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
530        The algorithm provides a way to skip certain starting offsets, and usually
531        faster than linear prefix searches.
532    
533    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
534        as for 'fr_FR' and 'french'. For some reason, however, it then used the
535        Windows-specific input and output files, which have 'french' screwed in.
536        So this could never have worked. One of the problems with locales is that
537        they aren't always the same. I have now updated RunTest so that it checks
538        the output of the locale test (test 3) against three different output
539        files, and it allows the test to pass if any one of them matches. With luck
540        this should make the test pass on some versions of Solaris where it was
541        failing. Because of the uncertainty, the script did not used to stop if
542        test 3 failed; it now does. If further versions of a French locale ever
543        come to light, they can now easily be added.
544    
545    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
546        there was a message during ./configure, but it did not stop. This now
547        provokes an error. The invalid example in README has been corrected.
548        If a value less than the minimum is given, the minimum value has always
549        been used, but now a warning is given.
550    
551    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
552        was a bug in the test system, which is now fixed. Also, the list of various
553        configurations that are tested for each release did not have one with both
554        16/32 bits and --enable-bar-anycrlf. It now does.
555    
556    18. pcretest was missing "-C bsr" for displaying the \R default setting.
557    
558    19. Little endian PowerPC systems are supported now by the JIT compiler.
559    
560    20. The fast forward newline mechanism could enter to an infinite loop on
561        certain invalid UTF-8 input. Although we don't support these cases
562        this issue can be fixed by a performance optimization.
563    
564    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
565        not take account if existing stack usage. There is now a new global
566        variable called pcre_stack_guard that can be set to point to an external
567        function to check stack availability. It is called at the start of
568        processing every parenthesized group.
569    
570    22. A typo in the code meant that in ungreedy mode the max/min qualifier
571        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
572        not match "ab".
573    
574    23. When UTF was disabled, the JIT program reported some incorrect compile
575        errors. These messages are silenced now.
576    
577    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
578        compiler.
579    
580    25. Change all the temporary files used in RunGrepTest to be different to those
581        used by RunTest so that the tests can be run simultaneously, for example by
582        "make -j check".
583    
584    
585    Version 8.34 15-December-2013
586    -----------------------------
587    
588    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
589        executable memory. Patch inspired by Carsten Klein.
590    
591    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
592        this macro is never tested and has no effect, because the work to support
593        coverage involves only compiling and linking options and special targets in
594        the Makefile. The comment in config.h implied that defining the macro would
595        enable coverage support, which is totally false. There was also support for
596        setting this macro in the CMake files (my fault, I just copied it from
597        configure). SUPPORT_GCOV has now been removed.
598    
599    3.  Make a small performance improvement in strlen16() and strlen32() in
600        pcretest.
601    
602    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
603        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
604        these cases). There was also one in pcretest.c.
605    
606    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
607    
608    6.  In UTF mode, the code for checking whether a group could match an empty
609        string (which is used for indefinitely repeated groups to allow for
610        breaking an infinite loop) was broken when the group contained a repeated
611        negated single-character class with a character that occupied more than one
612        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
613        UTF-8 mode). The effect was undefined: the group might or might not be
614        deemed as matching an empty string, or the program might have crashed.
615    
616    7.  The code for checking whether a group could match an empty string was not
617        recognizing that \h, \H, \v, \V, and \R must match a character.
618    
619    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
620        an empty string. If it can, pcretest shows this in its information output.
621    
622    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
623        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
624        matched by pcre_exec() without using JIT:
625    
626        (a) If the rest of the pattern did not match after a maximal run of
627            grapheme clusters, the code for backing up to try with fewer of them
628            did not always back up over a full grapheme when characters that do not
629            have the modifier quality were involved, e.g. Hangul syllables.
630    
631        (b) If the match point in a subject started with modifier character, and
632            there was no match, the code could incorrectly back up beyond the match
633            point, and potentially beyond the first character in the subject,
634            leading to a segfault or an incorrect match result.
635    
636    10. A conditional group with an assertion condition could lead to PCRE
637        recording an incorrect first data item for a match if no other first data
638        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
639        first data item, and therefore matched "ca" after "c" instead of at the
640        start.
641    
642    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
643        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
644    
645    12. The source of pcregrep now includes z/OS-specific code so that it can be
646        compiled for z/OS as part of the special z/OS distribution.
647    
648    13. Added the -T and -TM options to pcretest.
649    
650    14. The code in pcre_compile.c for creating the table of named capturing groups
651        has been refactored. Instead of creating the table dynamically during the
652        actual compiling pass, the information is remembered during the pre-compile
653        pass (on the stack unless there are more than 20 named groups, in which
654        case malloc() is used) and the whole table is created before the actual
655        compile happens. This has simplified the code (it is now nearly 150 lines
656        shorter) and prepared the way for better handling of references to groups
657        with duplicate names.
658    
659    15. A back reference to a named subpattern when there is more than one of the
660        same name now checks them in the order in which they appear in the pattern.
661        The first one that is set is used for the reference. Previously only the
662        first one was inspected. This change makes PCRE more compatible with Perl.
663    
664    16. Unicode character properties were updated from Unicode 6.3.0.
665    
666    17. The compile-time code for auto-possessification has been refactored, based
667        on a patch by Zoltan Herczeg. It now happens after instead of during
668        compilation. The code is cleaner, and more cases are handled. The option
669        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
670        options in pcretest are provided to set it. It can also be set by
671        (*NO_AUTO_POSSESS) at the start of a pattern.
672    
673    18. The character VT has been added to the default ("C" locale) set of
674        characters that match \s and are generally treated as white space,
675        following this same change in Perl 5.18. There is now no difference between
676        "Perl space" and "POSIX space". Whether VT is treated as white space in
677        other locales depends on the locale.
678    
679    19. The code for checking named groups as conditions, either for being set or
680        for being recursed, has been refactored (this is related to 14 and 15
681        above). Processing unduplicated named groups should now be as fast at
682        numerical groups, and processing duplicated groups should be faster than
683        before.
684    
685    20. Two patches to the CMake build system, by Alexander Barkov:
686    
687          (1) Replace the "source" command by "." in CMakeLists.txt because
688              "source" is a bash-ism.
689    
690          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
691              without these the CMake build does not work on Solaris.
692    
693    21. Perl has changed its handling of \8 and \9. If there is no previously
694        encountered capturing group of those numbers, they are treated as the
695        literal characters 8 and 9 instead of a binary zero followed by the
696        literals. PCRE now does the same.
697    
698    22. Following Perl, added \o{} to specify codepoints in octal, making it
699        possible to specify values greater than 0777 and also making them
700        unambiguous.
701    
702    23. Perl now gives an error for missing closing braces after \x{... instead of
703        treating the string as literal. PCRE now does the same.
704    
705    24. RunTest used to grumble if an inappropriate test was selected explicitly,
706        but just skip it when running all tests. This make it awkward to run ranges
707        of tests when one of them was inappropriate. Now it just skips any
708        inappropriate tests, as it always did when running all tests.
709    
710    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
711        character types such as \d or \w, too many callouts were inserted, and the
712        data that they returned was rubbish.
713    
714    26. In UCP mode, \s was not matching two of the characters that Perl matches,
715        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
716        were matched by \h. The code has now been refactored so that the lists of
717        the horizontal and vertical whitespace characters used for \h and \v (which
718        are defined only in one place) are now also used for \s.
719    
720    27. Add JIT support for the 64 bit TileGX architecture.
721        Patch by Jiong Wang (Tilera Corporation).
722    
723    28. Possessive quantifiers for classes (both explicit and automatically
724        generated) now use special opcodes instead of wrapping in ONCE brackets.
725    
726    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
727        (because it's meaningless), this was not happening when PCRE_CASELESS was
728        set. Not wrong, but inefficient.
729    
730    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
731        properties for \w, \d, etc) is present in a test regex. Otherwise if the
732        test contains no characters greater than 255, Perl doesn't realise it
733        should be using Unicode semantics.
734    
735    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
736        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
737        does in Unicode mode.
738    
739    32. Added the "forbid" facility to pcretest so that putting tests into the
740        wrong test files can sometimes be quickly detected.
741    
742    33. There is now a limit (default 250) on the depth of nesting of parentheses.
743        This limit is imposed to control the amount of system stack used at compile
744        time. It can be changed at build time by --with-parens-nest-limit=xxx or
745        the equivalent in CMake.
746    
747    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
748        errors. Perl warns for these when in warning mode, but PCRE has no facility
749        for giving warnings.
750    
751    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
752        However, this was not working for (?!) because it is optimized to (*FAIL),
753        for which PCRE does not allow quantifiers. The optimization is now disabled
754        when a quantifier follows (?!). I can't see any use for this, but it makes
755        things uniform.
756    
757    36. Perl no longer allows group names to start with digits, so I have made this
758        change also in PCRE. It simplifies the code a bit.
759    
760    37. In extended mode, Perl ignores spaces before a + that indicates a
761        possessive quantifier. PCRE allowed a space before the quantifier, but not
762        before the possessive +. It now does.
763    
764    38. The use of \K (reset reported match start) within a repeated possessive
765        group such as (a\Kb)*+ was not working.
766    
767    40. Document that the same character tables must be used at compile time and
768        run time, and that the facility to pass tables to pcre_exec() and
769        pcre_dfa_exec() is for use only with saved/restored patterns.
770    
771    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
772        features for Builds with MSVC:
773    
774        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
775           builds). The .rc files can be used to set FileDescription and many other
776           attributes.
777    
778        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
779           This allows higher-level build scripts which want .pdb files to avoid
780           hard-coding the exact files needed."
781    
782    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
783        mean "start of word" and "end of word", respectively, as a transition aid.
784    
785    43. A minimizing repeat of a class containing codepoints greater than 255 in
786        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
787        compiled to use the heap for recursion.
788    
789    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
790        is configured.
791    
792    
793    Version 8.33 28-May-2013
794    ------------------------
795    
796    1.  Added 'U' to some constants that are compared to unsigned integers, to
797        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
798        variables that are added to signed variables, to ensure the result is
799        signed and can be negated.
800    
801    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
802        CMake config files.
803    
804    3.  Revise the creation of config.h.generic so that all boolean macros are
805        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
806        overriding via -D on the command line possible.
807    
808    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
809        to unsigned int is reported to make a quite noticeable speed difference in
810        a specific Windows environment. Testing on Linux did also appear to show
811        some benefit (and it is clearly not harmful). Also fixed the definition of
812        Xop which should be unsigned.
813    
814    5.  Related to (4), changing the definition of the intermediate variable cc
815        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
816        improvements.
817    
818    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
819        unnecessary spaces.
820    
821    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
822        and later.
823    
824    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
825    
826    9.  Optimizing fast_forward_start_bits in JIT.
827    
828    10. Adding support for callouts in JIT, and fixing some issues revealed
829        during this work. Namely:
830    
831        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
832    
833        (b) Minimum length was not checked before the matching is started.
834    
835    11. The value of capture_last that is passed to callouts was incorrect in some
836        cases when there was a capture on one path that was subsequently abandoned
837        after a backtrack. Also, the capture_last value is now reset after a
838        recursion, since all captures are also reset in this case.
839    
840    12. The interpreter no longer returns the "too many substrings" error in the
841        case when an overflowing capture is in a branch that is subsequently
842        abandoned after a backtrack.
843    
844    13. In the pathological case when an offset vector of size 2 is used, pcretest
845        now prints out the matched string after a yield of 0 or 1.
846    
847    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
848        Only supported by the JIT compiler at the moment.
849    
850    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
851    
852    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
853        offset of the starting point of the matching process, provided the offsets
854        vector is large enough.
855    
856    17. The \A escape now records a lookbehind value of 1, though its execution
857        does not actually inspect the previous character. This is to ensure that,
858        in partial multi-segment matching, at least one character from the old
859        segment is retained when a new segment is processed. Otherwise, if there
860        are no lookbehinds in the pattern, \A might match incorrectly at the start
861        of a new segment.
862    
863    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
864    
865    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
866        gives some modest performance improvement in 8-bit mode.
867    
868    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
869        be expressed in certain programming languages using Universal Character
870        Names.
871    
872    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
873        which points out that "non characters" are not "characters that may not
874        appear in Unicode strings" but rather "characters that are reserved for
875        internal use and have only local meaning".
876    
877    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
878        there was a conditional group that depended on an assertion, if the
879        assertion was false, the callout that immediately followed the alternation
880        in the condition was skipped when pcre_exec() was used for matching.
881    
882    23. Allow an explicit callout to be inserted before an assertion that is the
883        condition for a conditional group, for compatibility with automatic
884        callouts, which always insert a callout at this point.
885    
886    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
887        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
888    
889    25. (*PRUNE) is now supported by the JIT compiler.
890    
891    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
892    
893    27. Fix the case where there are two or more SKIPs with arguments that may be
894        ignored.
895    
896    28. (*SKIP) is now supported by the JIT compiler.
897    
898    29. (*THEN) is now supported by the JIT compiler.
899    
900    30. Update RunTest with additional test selector options.
901    
902    31. The way PCRE handles backtracking verbs has been changed in two ways.
903    
904        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
905        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
906        backtracking. In some cases this makes it more Perl-compatible, but Perl's
907        rather obscure rules do not always do the same thing.
908    
909        (2) Previously, backtracking verbs were confined within assertions. This is
910        no longer the case for positive assertions, except for (*ACCEPT). Again,
911        this sometimes improves Perl compatibility, and sometimes does not.
912    
913    32. A number of tests that were in test 2 because Perl did things differently
914        have been moved to test 1, because either Perl or PCRE has changed, and
915        these tests are now compatible.
916    
917    32. Backtracking control verbs are now handled in the same way in JIT and
918        interpreter.
919    
920    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
921        contained a forward subroutine reference caused a compile error.
922    
923    34. Auto-detect and optimize limited repetitions in JIT.
924    
925    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
926        blocking (*UTF) etc.
927    
928    36. In the interpreter, maximizing pattern repetitions for characters and
929        character types now use tail recursion, which reduces stack usage.
930    
931    37. The value of the max lookbehind was not correctly preserved if a compiled
932        and saved regex was reloaded on a host of different endianness.
933    
934    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
935        of the compiled pattern block, expand the flags field from 16 to 32 bits
936        because it was almost full.
937    
938    39. Try madvise first before posix_madvise.
939    
940    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
941        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
942        original change locked out all matches of empty strings. This has been
943        changed so that one match of an empty string per line is recognized.
944        Subsequent searches on the same line (for colouring or for --only-matching,
945        for example) do not recognize empty strings.
946    
947    41. Applied a user patch to fix a number of spelling mistakes in comments.
948    
949    42. Data lines longer than 65536 caused pcretest to crash.
950    
951    43. Clarified the data type for length and startoffset arguments for pcre_exec
952        and pcre_dfa_exec in the function-specific man pages, where they were
953        explicitly stated to be in bytes, never having been updated. I also added
954        some clarification to the pcreapi man page.
955    
956    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
957        a segmentation fault.
958    
959    
960    Version 8.32 30-November-2012
961    -----------------------------
962    
963    1.  Improved JIT compiler optimizations for first character search and single
964        character iterators.
965    
966    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
967        Patch by Daniel Richard G.
968    
969    3.  Single character iterator optimizations in the JIT compiler.
970    
971    4.  Improved JIT compiler optimizations for character ranges.
972    
973    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
974        Reported by Giuseppe D'Angelo.
975    
976    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
977        of a line, was being set incorrectly in cases where .* appeared inside
978        atomic brackets at the start of a pattern, or where there was a subsequent
979        *PRUNE or *SKIP.
980    
981    7.  Improved instruction cache flush for POWER/PowerPC.
982        Patch by Daniel Richard G.
983    
984    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
985        grep:
986    
987        (a) There is now no limit to the number of patterns to be matched.
988    
989        (b) An error is given if a pattern is too long.
990    
991        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
992            are now supported.
993    
994        (d) --exclude-from and --include-from (multiple use) have been added.
995    
996        (e) Exclusions and inclusions now apply to all files and directories, not
997            just to those obtained from scanning a directory recursively.
998    
999        (f) Multiple uses of -f and --file-list are now supported.
1000    
1001        (g) In a Windows environment, the default for -d has been changed from
1002            "read" (the GNU grep default) to "skip", because otherwise the presence
1003            of a directory in the file list provokes an error.
1004    
1005        (h) The documentation has been revised and clarified in places.
1006    
1007    9.  Improve the matching speed of capturing brackets.
1008    
1009    10. Changed the meaning of \X so that it now matches a Unicode extended
1010        grapheme cluster.
1011    
1012    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1013        out POSIX threads when JIT support is configured.
1014    
1015    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1016    
1017    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1018        structure could go wrong in environments where size_t is not the same size
1019        as int.
1020    
1021    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1022        set.
1023    
1024    15. The EBCDIC support had decayed; later updates to the code had included
1025        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1026        general tidy up of EBCDIC-related issues, and the documentation was also
1027        not quite right. There is now a test that can be run on ASCII systems to
1028        check some of the EBCDIC-related things (but is it not a full test).
1029    
1030    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1031        in a small tidy to the code.
1032    
1033    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1034    
1035    18. If the --only-matching (-o) option in pcregrep is specified multiple
1036        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1037        substrings matched by the 1st and 2nd capturing parentheses. A separating
1038        string can be specified by --om-separator (default empty).
1039    
1040    19. Improving the first n character searches.
1041    
1042    20. Turn case lists for horizontal and vertical white space into macros so that
1043        they are defined only once.
1044    
1045    21. This set of changes together give more compatible Unicode case-folding
1046        behaviour for characters that have more than one other case when UCP
1047        support is available.
1048    
1049        (a) The Unicode property table now has offsets into a new table of sets of
1050            three or more characters that are case-equivalent. The MultiStage2.py
1051            script that generates these tables (the pcre_ucd.c file) now scans
1052            CaseFolding.txt instead of UnicodeData.txt for character case
1053            information.
1054    
1055        (b) The code for adding characters or ranges of characters to a character
1056            class has been abstracted into a generalized function that also handles
1057            case-independence. In UTF-mode with UCP support, this uses the new data
1058            to handle characters with more than one other case.
1059    
1060        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1061            whose other case is greater than 256 are now correctly matched
1062            caselessly. Previously, the high codepoint matched the low one, but not
1063            vice versa.
1064    
1065        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1066            of the new class addition function, using character lists defined as
1067            macros alongside the case definitions of 20 above.
1068    
1069        (e) Caseless back references now work with characters that have more than
1070            one other case.
1071    
1072        (f) General caseless matching of characters with more than one other case
1073            is supported.
1074    
1075    22. Unicode character properties were updated from Unicode 6.2.0
1076    
1077    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1078    
1079    24. Add support for 32-bit character strings, and UTF-32
1080    
1081    25. Major JIT compiler update (code refactoring and bugfixing).
1082        Experimental Sparc 32 support is added.
1083    
1084    26. Applied a modified version of Daniel Richard G's patch to create
1085        pcre.h.generic and config.h.generic by "make" instead of in the
1086        PrepareRelease script.
1087    
1088    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1089        pcre_compile.c when checking for a zero character.
1090    
1091    28. Introducing a native interface for JIT. Through this interface, the compiled
1092        machine code can be directly executed. The purpose of this interface is to
1093        provide fast pattern matching, so several sanity checks are not performed.
1094        However, feature tests are still performed. The new interface provides
1095        1.4x speedup compared to the old one.
1096    
1097    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1098        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1099        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1100    
1101    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1102        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1103        the "old" RFC 2279). Instead, it ended up passing a negative length to
1104        pcre_exec().
1105    
1106    31. Add support for GCC's visibility feature to hide internal functions.
1107    
1108    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1109        "unknown -C option" after outputting 0 or 1.
1110    
1111    33. There is now support for generating a code coverage report for the test
1112        suite in environments where gcc is the compiler and lcov is installed. This
1113        is mainly for the benefit of the developers.
1114    
1115    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1116        unaddressable using valgrind annotations, allowing valgrind to detect
1117        invalid memory accesses. This is mainly for the benefit of the developers.
1118    
1119    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1120    
1121    26. Give configure error if --enable-cpp but no C++ compiler found.
1122    
1123    
1124    Version 8.31 06-July-2012
1125    -------------------------
1126    
1127    1.  Fixing a wrong JIT test case and some compiler warnings.
1128    
1129    2.  Removed a bashism from the RunTest script.
1130    
1131    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1132        to unsigned type, result still unsigned" that was given by an MS compiler
1133        on encountering the code "-sizeof(xxx)".
1134    
1135    4.  Partial matching support is added to the JIT compiler.
1136    
1137    5.  Fixed several bugs concerned with partial matching of items that consist
1138        of more than one character:
1139    
1140        (a) /^(..)\1/ did not partially match "aba" because checking references was
1141            done on an "all or nothing" basis. This also applied to repeated
1142            references.
1143    
1144        (b) \R did not give a hard partial match if \r was found at the end of the
1145            subject.
1146    
1147        (c) \X did not give a hard partial match after matching one or more
1148            characters at the end of the subject.
1149    
1150        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1151            a partial match for the string "\r".
1152    
1153        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1154            a partial match for a CR character at the end of the subject string.
1155    
1156    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1157        running pcretest, the text "(JIT)" added to the output whenever JIT is
1158        actually used to run the match.
1159    
1160    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1161        or /S+[+] with a digit between 1 and 7.
1162    
1163    8.  OP_NOT now supports any UTF character not just single-byte ones.
1164    
1165    9.  (*MARK) control verb is now supported by the JIT compiler.
1166    
1167    10. The command "./RunTest list" lists the available tests without actually
1168        running any of them. (Because I keep forgetting what they all are.)
1169    
1170    11. Add PCRE_INFO_MAXLOOKBEHIND.
1171    
1172    12. Applied a (slightly modified) user-supplied patch that improves performance
1173        when the heap is used for recursion (compiled with --disable-stack-for-
1174        recursion). Instead of malloc and free for each heap frame each time a
1175        logical recursion happens, frames are retained on a chain and re-used where
1176        possible. This sometimes gives as much as 30% improvement.
1177    
1178    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1179        call.
1180    
1181    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1182    
1183    15. It is now possible to link pcretest with libedit as an alternative to
1184        libreadline.
1185    
1186    16. (*COMMIT) control verb is now supported by the JIT compiler.
1187    
1188    17. The Unicode data tables have been updated to Unicode 6.1.0.
1189    
1190    18. Added --file-list option to pcregrep.
1191    
1192    19. Added binary file support to pcregrep, including the -a, --binary-files,
1193        -I, and --text options.
1194    
1195    20. The madvise function is renamed for posix_madvise for QNX compatibility
1196        reasons. Fixed by Giuseppe D'Angelo.
1197    
1198    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1199    
1200    22. Changed the option for creating symbolic links for 16-bit man pages from
1201        -s to -sf so that re-installing does not cause issues.
1202    
1203    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1204    
1205    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1206        matches in certain environments (the workspace was not being correctly
1207        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1208        some of the workspace data at the beginning of a restart.
1209    
1210    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1211        was not doing so when it should - probably a typo introduced by SVN 528
1212        (change 8.10/14).
1213    
1214    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1215        \w+ when the character tables indicated that \x{c4} was a word character.
1216        There were several related cases, all because the tests for doing a table
1217        lookup were testing for characters less than 127 instead of 255.
1218    
1219    27. If a pattern contains capturing parentheses that are not used in a match,
1220        their slots in the ovector are set to -1. For those that are higher than
1221        any matched groups, this happens at the end of processing. In the case when
1222        there were back references that the ovector was too small to contain
1223        (causing temporary malloc'd memory to be used during matching), and the
1224        highest capturing number was not used, memory off the end of the ovector
1225        was incorrectly being set to -1. (It was using the size of the temporary
1226        memory instead of the true size.)
1227    
1228    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1229        ovector size, it uses memory at the end of the block that it has got.
1230    
1231    29. Check for an overlong MARK name and give an error at compile time. The
1232        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1233    
1234    30. JIT compiler update.
1235    
1236    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1237        Rill for the patch.
1238    
1239    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1240    
1241    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1242    
1243    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1244        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1245        was enabled.
1246    
1247    35. Improve JIT code generation for greedy plus quantifier.
1248    
1249    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1250        1 to "aa" instead of to an empty string. The bug affected repeated groups
1251        that could potentially match an empty string.
1252    
1253    37. Optimizing single character iterators in JIT.
1254    
1255    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1256        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1257        codepoints that are too big for the mode are faulted, and in a UTF mode,
1258        disallowed codepoints are also faulted.
1259    
1260    39. If PCRE was compiled with UTF support, in three places in the DFA
1261        matcher there was code that should only have been obeyed in UTF mode, but
1262        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1263        processing when bytes with values greater than 127 were present. In 16-bit
1264        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1265        both cases the values are those that cannot be the first data item in a UTF
1266        character. The three items that might have provoked this were recursions,
1267        possessively repeated groups, and atomic groups.
1268    
1269    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1270        and pcregrep, because some OS require shared objects to be explicitly
1271        passed to ld, causing the link step to fail if they are not.
1272    
1273    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1274        mode, patterns that started with \h* or \R* might be incorrectly matched.
1275    
1276    
1277    Version 8.30 04-February-2012
1278    -----------------------------
1279    
1280    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1281        name is defined in ctype.h.
1282    
1283    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1284        only in quite long subpatterns.
1285    
1286    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1287        since it was replaced by pcre_fullinfo() in February 2000.
1288    
1289    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1290        match a (*MARK), and the match failed at the start of the subject, a
1291        reference to memory before the start of the subject could occur. This bug
1292        was introduced by fix 17 of release 8.21.
1293    
1294    5.  A reference to an unset group with zero minimum repetition was giving
1295        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1296        /(another)?(\1?)test/ matched against "hello world test". This bug was
1297        introduced in release 8.13.
1298    
1299    6.  Add support for 16-bit character strings (a large amount of work involving
1300        many changes and refactorings).
1301    
1302    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1303        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1304        from a file.
1305    
1306    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1307        rounding is not applied in this particular case).
1308    
1309    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1310        if they appear, or are escaped, in patterns.
1311    
1312    10. Get rid of a number of -Wunused-but-set-variable warnings.
1313    
1314    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1315        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1316        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1317        also returns the mark "x". This bug applied to capturing parentheses,
1318        non-capturing parentheses, and atomic parentheses. It also applied to some
1319        assertions.
1320    
1321    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1322        information out of configure.ac instead of relying on pcre.h.generic, which
1323        is not stored in the repository.
1324    
1325    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1326        -lreadline.
1327    
1328    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1329    
1330    15. Applied Graycode's patch to put the top-level frame on the stack rather
1331        than the heap when not using the stack for recursion. This gives a
1332        performance improvement in many cases when recursion is not deep.
1333    
1334    16. Experimental code added to "pcretest -C" to output the stack frame size.
1335    
1336    
1337    Version 8.21 12-Dec-2011
1338    ------------------------
1339    
1340    1.  Updating the JIT compiler.
1341    
1342    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1343        are added as well.
1344    
1345    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1346        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1347        calling _pcre_jit_exec. Some extra comments are added.
1348    
1349    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1350        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1351        was introduced by change 18 for 8.20.
1352    
1353    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1354        ECMA-262 standard.
1355    
1356    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1357        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1358        This bug was probably introduced by change 9 of 8.13.
1359    
1360    7.  While fixing 6 above, I noticed that a number of other items were being
1361        incorrectly rejected as "not fixed length". This arose partly because newer
1362        opcodes had not been added to the fixed-length checking code. I have (a)
1363        corrected the bug and added tests for these items, and (b) arranged for an
1364        error to occur if an unknown opcode is encountered while checking for fixed
1365        length instead of just assuming "not fixed length". The items that were
1366        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1367        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1368        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1369    
1370    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1371        being incorrectly compiled and would have given unpredicatble results.
1372    
1373    9.  A possessively repeated subpattern with minimum repeat count greater than
1374        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1375        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1376        the first (A) could occur when it should not.
1377    
1378    10. Add a cast and remove a redundant test from the code.
1379    
1380    11. JIT should use pcre_malloc/pcre_free for allocation.
1381    
1382    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1383        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1384        is anything other than /usr, -L is still shown).
1385    
1386    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1387    
1388    14. Perl does not support \N without a following name in a [] class; PCRE now
1389        also gives an error.
1390    
1391    15. If a forward reference was repeated with an upper limit of around 2000,
1392        it caused the error "internal error: overran compiling workspace". The
1393        maximum number of forward references (including repeats) was limited by the
1394        internal workspace, and dependent on the LINK_SIZE. The code has been
1395        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1396        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1397        of around 200,000 forward references. While doing this, I also speeded up
1398        the filling in of repeated forward references.
1399    
1400    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1401        incorrectly expecting the subject to contain another "a" after the start.
1402    
1403    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1404        in the match, the SKIP should be ignored. This was not happening; instead
1405        the SKIP was being treated as NOMATCH. For patterns such as
1406        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1407        tested.
1408    
1409    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1410        now much more compatible with Perl, in particular in cases where the result
1411        is a non-match for a non-anchored pattern. For example, if
1412        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1413        "m", where previously it did not return a name. A side effect of this
1414        change is that for partial matches, the last encountered mark name is
1415        returned, as for non matches. A number of tests that were previously not
1416        Perl-compatible have been moved into the Perl-compatible test files. The
1417        refactoring has had the pleasing side effect of removing one argument from
1418        the match() function, thus reducing its stack requirements.
1419    
1420    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1421        subsequent uses of /S (without +) incorrectly behaved like /S+.
1422    
1423    21. Retrieve executable code size support for the JIT compiler and fixing
1424        some warnings.
1425    
1426    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1427        not work when the shorter character appeared right at the end of the
1428        subject string.
1429    
1430    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1431        systems.
1432    
1433    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1434        output it when the /M option is used in pcretest.
1435    
1436    25. The CheckMan script was not being included in the distribution. Also, added
1437        an explicit "perl" to run Perl scripts from the PrepareRelease script
1438        because this is reportedly needed in Windows.
1439    
1440    26. If study data was being save in a file and studying had not found a set of
1441        "starts with" bytes for the pattern, the data written to the file (though
1442        never used) was taken from uninitialized memory and so caused valgrind to
1443        complain.
1444    
1445    27. Updated RunTest.bat as provided by Sheri Pierce.
1446    
1447    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1448    
1449    29. Computation of memory usage for the table of capturing group names was
1450        giving an unnecessarily large value.
1451    
1452    
1453    Version 8.20 21-Oct-2011
1454    ------------------------
1455    
1456    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1457        a POSIX class. After further experiments with Perl, which convinced me that
1458        Perl has bugs and confusions, a closing square bracket is no longer allowed
1459        in a POSIX name. This bug also affected patterns with classes that started
1460        with full stops.
1461    
1462    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1463        captured substring, but while checking the failing first alternative,
1464        substring 1 is temporarily captured. If the output vector supplied to
1465        pcre_exec() was not big enough for this capture, the yield of the function
1466        was still zero ("insufficient space for captured substrings"). This cannot
1467        be totally fixed without adding another stack variable, which seems a lot
1468        of expense for a edge case. However, I have improved the situation in cases
1469        such as /(a)(b)x|abc/ matched against "abc", where the return code
1470        indicates that fewer than the maximum number of slots in the ovector have
1471        been set.
1472    
1473    3.  Related to (2) above: when there are more back references in a pattern than
1474        slots in the output vector, pcre_exec() uses temporary memory during
1475        matching, and copies in the captures as far as possible afterwards. It was
1476        using the entire output vector, but this conflicts with the specification
1477        that only 2/3 is used for passing back captured substrings. Now it uses
1478        only the first 2/3, for compatibility. This is, of course, another edge
1479        case.
1480    
1481    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1482        main code base, and can be used by building with --enable-jit. When this is
1483        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1484        runtime --no-jit option is given.
1485    
1486    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1487        ovector, the return from the function was zero, implying that there were
1488        other matches that did not fit. The correct "exactly full" value is now
1489        returned.
1490    
1491    6.  If a subpattern that was called recursively or as a subroutine contained
1492        (*PRUNE) or any other control that caused it to give a non-standard return,
1493        invalid errors such as "Error -26 (nested recursion at the same subject
1494        position)" or even infinite loops could occur.
1495    
1496    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1497        computing the minimum length on reaching *ACCEPT, and so ended up with the
1498        wrong value of 1 rather than 0. Further investigation indicates that
1499        computing a minimum subject length in the presence of *ACCEPT is difficult
1500        (think back references, subroutine calls), and so I have changed the code
1501        so that no minimum is registered for a pattern that contains *ACCEPT.
1502    
1503    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1504        it was not handled as intended. [But see 16 below.]
1505    
1506    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1507        Sheri Pierce.
1508    
1509    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1510        the first byte in a match must be "a".
1511    
1512    11. Change 17 for 8.13 increased the recursion depth for patterns like
1513        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1514        pattern contains any instances of (*THEN). If it does not, the old
1515        optimizations are restored. It would be nice to do this on a per-group
1516        basis, but at the moment that is not feasible.
1517    
1518    12. In some environments, the output of pcretest -C is CRLF terminated. This
1519        broke RunTest's code that checks for the link size. A single white space
1520        character after the value is now allowed for.
1521    
1522    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1523        For "fr", it uses the Windows-specific input and output files.
1524    
1525    14. If (*THEN) appeared in a group that was called recursively or as a
1526        subroutine, it did not work as intended. [But see next item.]
1527    
1528    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1529        pattern fragments (but not containing any | characters). If A and B are
1530        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1531        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1532        D. In other words, Perl considers parentheses that do not contain any |
1533        characters to be part of a surrounding alternative, whereas PCRE was
1534        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1535        differently. PCRE now behaves in the same way as Perl, except in the case
1536        of subroutine/recursion calls such as (?1) which have in any case always
1537        been different (but PCRE had them first :-).
1538    
1539    16. Related to 15 above: Perl does not treat the | in a conditional group as
1540        creating alternatives. Such a group is treated in the same way as an
1541        ordinary group without any | characters when processing (*THEN). PCRE has
1542        been changed to match Perl's behaviour.
1543    
1544    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1545        RunGrepTest script failed.
1546    
1547    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1548        inevitable for groups that contain captures, but it can lead to a lot of
1549        stack use in large patterns. The old behaviour has been restored for atomic
1550        groups that do not contain any capturing parentheses.
1551    
1552    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1553        suppress the check for a minimum subject length at run time. (If it was
1554        given to pcre_exec() or pcre_dfa_exec() it did work.)
1555    
1556    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1557        fail to work when decoding hex characters in data strings in EBCDIC
1558        environments.
1559    
1560    21. It appears that in at least one Mac OS environment, the isxdigit() function
1561        is implemented as a macro that evaluates to its argument more than once,
1562        contravening the C 90 Standard (I haven't checked a later standard). There
1563        was an instance in pcretest which caused it to go wrong when processing
1564        \x{...} escapes in subject strings. The has been rewritten to avoid using
1565        things like p++ in the argument of isxdigit().
1566    
1567    
1568    Version 8.13 16-Aug-2011
1569    ------------------------
1570    
1571    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1572    
1573    2.  Two minor typos in pcre_internal.h have been fixed.
1574    
1575    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1576        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1577        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1578    
1579    4.  There were a number of related bugs in the code for matching backrefences
1580        caselessly in UTF-8 mode when codes for the characters concerned were
1581        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1582        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1583        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1584        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1585        2-byte code at the end of the subject (it thought there wasn't enough data
1586        left).
1587    
1588    5.  Comprehensive information about what went wrong is now returned by
1589        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1590        as the output vector has at least 2 elements. The offset of the start of
1591        the failing character and a reason code are placed in the vector.
1592    
1593    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1594        now returned is for the first byte of the failing character, instead of the
1595        last byte inspected. This is an incompatible change, but I hope it is small
1596        enough not to be a problem. It makes the returned offset consistent with
1597        pcre_exec() and pcre_dfa_exec().
1598    
1599    7.  pcretest now gives a text phrase as well as the error number when
1600        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1601        failure, the offset and reason code are output.
1602    
1603    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1604        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1605        back over a single character (\n). This seems wrong (because it treated the
1606        two characters as a single entity when going forwards), conflicts with the
1607        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1608        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1609        has been changed.
1610    
1611    9.  Some internal refactoring has changed the processing so that the handling
1612        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1613        time (the PCRE_DOTALL option was changed this way some time ago: version
1614        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1615        which was always a bit of a fudge. It also means that there is one less
1616        argument for the match() function, which reduces its stack requirements
1617        slightly. This change also fixes an incompatibility with Perl: the pattern
1618        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1619    
1620    10. More internal refactoring has drastically reduced the number of recursive
1621        calls to match() for possessively repeated groups such as (abc)++ when
1622        using pcre_exec().
1623    
1624    11. While implementing 10, a number of bugs in the handling of groups were
1625        discovered and fixed:
1626    
1627        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1628        (a|)*(?1) gave a compile-time internal error.
1629        ((a|)+)+  did not notice that the outer group could match an empty string.
1630        (^a|^)+   was not marked as anchored.
1631        (.*a|.*)+ was not marked as matching at start or after a newline.
1632    
1633    12. Yet more internal refactoring has removed another argument from the match()
1634        function. Special calls to this function are now indicated by setting a
1635        value in a variable in the "match data" data block.
1636    
1637    13. Be more explicit in pcre_study() instead of relying on "default" for
1638        opcodes that mean there is no starting character; this means that when new
1639        ones are added and accidentally left out of pcre_study(), testing should
1640        pick them up.
1641    
1642    14. The -s option of pcretest has been documented for ages as being an old
1643        synonym of -m (show memory usage). I have changed it to mean "force study
1644        for every regex", that is, assume /S for every regex. This is similar to -i
1645        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1646        using it. It makes it easier to run collections of tests with and without
1647        study enabled, and thereby test pcre_study() more easily. All the standard
1648        tests are now run with and without -s (but some patterns can be marked as
1649        "never study" - see 20 below).
1650    
1651    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1652        restoration of the capturing data to the outer values was not happening
1653        correctly.
1654    
1655    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1656        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1657        pattern had matched an empty string, and so incorrectly returned a no
1658        match.
1659    
1660    17. There was optimizing code for the last branch of non-capturing parentheses,
1661        and also for the obeyed branch of a conditional subexpression, which used
1662        tail recursion to cut down on stack usage. Unfortunately, now that there is
1663        the possibility of (*THEN) occurring in these branches, tail recursion is
1664        no longer possible because the return has to be checked for (*THEN). These
1665        two optimizations have therefore been removed. [But see 8.20/11 above.]
1666    
1667    18. If a pattern containing \R was studied, it was assumed that \R always
1668        matched two bytes, thus causing the minimum subject length to be
1669        incorrectly computed because \R can also match just one byte.
1670    
1671    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1672        was incorrectly computed.
1673    
1674    20. If /S is present twice on a test pattern in pcretest input, it now
1675        *disables* studying, thereby overriding the use of -s on the command line
1676        (see 14 above). This is necessary for one or two tests to keep the output
1677        identical in both cases.
1678    
1679    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1680        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1681    
1682    22. When an atomic group that contained a capturing parenthesis was
1683        successfully matched, but the branch in which it appeared failed, the
1684        capturing was not being forgotten if a higher numbered group was later
1685        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1686        group 1 to "a", when in fact it should be unset. This applied to multi-
1687        branched capturing and non-capturing groups, repeated or not, and also to
1688        positive assertions (capturing in negative assertions does not happen
1689        in PCRE) and also to nested atomic groups.
1690    
1691    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1692        subject after a captured substring, to make it easier to tell which of a
1693        number of identical substrings has been captured.
1694    
1695    24. The way atomic groups are processed by pcre_exec() has been changed so that
1696        if they are repeated, backtracking one repetition now resets captured
1697        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1698        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1699        "aaa". Previously, it would have been "a". As part of this code
1700        refactoring, the way recursive calls are handled has also been changed.
1701    
1702    25. If an assertion condition captured any substrings, they were not passed
1703        back unless some other capturing happened later. For example, if
1704        (?(?=(a))a) was matched against "a", no capturing was returned.
1705    
1706    26. When studying a pattern that contained subroutine calls or assertions,
1707        the code for finding the minimum length of a possible match was handling
1708        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1709        group 1 called group 2 while simultaneously a separate group 2 called group
1710        1). A stack overflow occurred in this case. I have fixed this by limiting
1711        the recursion depth to 10.
1712    
1713    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1714        Fortmann. This supports explicit test numbers on the command line, and has
1715        argument validation and error reporting.
1716    
1717    28. An instance of \X with an unlimited repeat could fail if at any point the
1718        first character it looked at was a mark character.
1719    
1720    29. Some minor code refactoring concerning Unicode properties and scripts
1721        should reduce the stack requirement of match() slightly.
1722    
1723    30. Added the '=' option to pcretest to check the setting of unused capturing
1724        slots at the end of the pattern, which are documented as being -1, but are
1725        not included in the return count.
1726    
1727    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1728        compiled something random. Now it gives a compile-time error (as does
1729        Perl).
1730    
1731    32. A *MARK encountered during the processing of a positive assertion is now
1732        recorded and passed back (compatible with Perl).
1733    
1734    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1735        had alternative anchored branches, the search for a second match in a line
1736        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1737        matched the line "0102" twice. The same bug affected patterns that started
1738        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1739        twice.
1740    
1741    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1742        does, and because of capturing effects, quantifying parenthesized
1743        assertions may at times be useful. Quantifiers are now allowed for
1744        parenthesized assertions.
1745    
1746    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1747    
1748    36. \g was being checked for fancy things in a character class, when it should
1749        just be a literal "g".
1750    
1751    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1752        appearance of a nested POSIX class supersedes an apparent external class.
1753        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1754        unescaped square brackets may also appear as part of class names. For
1755        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1756        more like Perl. (But see 8.20/1 above.)
1757    
1758    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1759        was because it thought it was \N{name}, which is not supported).
1760    
1761    39. Add minix to OS list not supporting the -S option in pcretest.
1762    
1763    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1764        cannot analyze patterns in sufficient detail to catch mutual recursions
1765        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1766        subgroup is called recursively as a subpattern for a second time at the
1767        same position in the subject string. In previous releases this might have
1768        been caught by the recursion limit, or it might have run out of stack.
1769    
1770    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1771        happen only once. PCRE was, however incorrectly giving a compile time error
1772        "recursive call could loop indefinitely" because it cannot analyze the
1773        pattern in sufficient detail. The compile time test no longer happens when
1774        PCRE is compiling a conditional subpattern, but actual runaway loops are
1775        now caught at runtime (see 40 above).
1776    
1777    42. It seems that Perl allows any characters other than a closing parenthesis
1778        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1779        has been changed to be the same.
1780    
1781    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1782        as not to get warnings when autogen.sh is called. Also changed
1783        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1784    
1785    44. To help people who use pcregrep to scan files containing exceedingly long
1786        lines, the following changes have been made:
1787    
1788        (a) The default value of the buffer size parameter has been increased from
1789            8K to 20K. (The actual buffer used is three times this size.)
1790    
1791        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1792            PCRE is built.
1793    
1794        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1795            to be set at run time.
1796    
1797        (d) Numerical values in pcregrep options can be followed by K or M, for
1798            example --buffer-size=50K.
1799    
1800        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1801            given and the return code is set to 2.
1802    
1803    45. Add a pointer to the latest mark to the callout data block.
1804    
1805    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1806        partial match of an empty string instead of no match. This was specific to
1807        the use of ".".
1808    
1809    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1810        complete match instead of a partial match. This bug was dependent on both
1811        the PCRE_UTF8 and PCRE_DOTALL options being set.
1812    
1813    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1814        starting byte set, because \b was not being ignored.
1815    
1816    
1817    Version 8.12 15-Jan-2011
1818    ------------------------
1819    
1820    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1821        checks for such things as part of the documentation building process.
1822    
1823    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1824        --match-limit and --recursion-limit options (added for 8.11). In
1825        particular, this made one of the standard tests fail. (The integer value
1826        went into the wrong half of a long int.)
1827    
1828    3.  If the --colour option was given to pcregrep with -v (invert match), it
1829        did strange things, either producing crazy output, or crashing. It should,
1830        of course, ignore a request for colour when reporting lines that do not
1831        match.
1832    
1833    4.  Another pcregrep bug caused similar problems if --colour was specified with
1834        -M (multiline) and the pattern match finished with a line ending.
1835    
1836    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1837        matched in multiline mode, the following line was shown as part of the
1838        match. This seems wrong, so I have changed it.
1839    
1840    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1841        the check for further matches in the same line (so they could be coloured)
1842        to overrun the end of the current line. If another match was found, it was
1843        incorrectly shown (and then shown again when found in the next line).
1844    
1845    7.  If pcregrep was compiled under Windows, there was a reference to the
1846        function pcregrep_exit() before it was defined. I am assuming this was
1847        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1848        reported by a user. I've moved the definition above the reference.
1849    
1850    
1851    Version 8.11 10-Dec-2010
1852    ------------------------
1853    
1854    1.  (*THEN) was not working properly if there were untried alternatives prior
1855        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1856        backtracked to try for "b" instead of moving to the next alternative branch
1857        at the same level (in this case, to look for "c"). The Perl documentation
1858        is clear that when (*THEN) is backtracked onto, it goes to the "next
1859        alternative in the innermost enclosing group".
1860    
1861    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1862        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1863        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1864        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1865        (*THEN).
1866    
1867    3.  If \s appeared in a character class, it removed the VT character from
1868        the class, even if it had been included by some previous item, for example
1869        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1870        of \s, but is part of the POSIX "space" class.)
1871    
1872    4.  A partial match never returns an empty string (because you can always
1873        match an empty string at the end of the subject); however the checking for
1874        an empty string was starting at the "start of match" point. This has been
1875        changed to the "earliest inspected character" point, because the returned
1876        data for a partial match starts at this character. This means that, for
1877        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1878        (previously it gave "no match").
1879    
1880    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1881        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1882        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1883        has an implication that the given string is incomplete (because a partial
1884        match is preferred over a full match). For this reason, these items now
1885        give a partial match in this situation. [Aside: previously, the one case
1886        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1887        match rather than a full match, which was wrong by the old rules, but is
1888        now correct.]
1889    
1890    6.  There was a bug in the handling of #-introduced comments, recognized when
1891        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1892        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1893        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1894        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1895        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1896        places in pcre_compile().
1897    
1898    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1899        comments when looking ahead for named forward references to subpatterns,
1900        the only newline sequence it recognized was NL. It now handles newlines
1901        according to the set newline convention.
1902    
1903    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1904        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1905        cater for a lack of strerror(). These oversights have been fixed.
1906    
1907    9.  Added --match-limit and --recursion-limit to pcregrep.
1908    
1909    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1910    
1911    11. When the -o option was used, pcregrep was setting a return code of 1, even
1912        when matches were found, and --line-buffered was not being honoured.
1913    
1914    12. Added an optional parentheses number to the -o and --only-matching options
1915        of pcregrep.
1916    
1917    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1918        can match an empty string. The code to do it in pcretest and pcredemo
1919        needed fixing:
1920    
1921        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1922            only one byte after an empty string match just before CRLF (this case
1923            just got forgotten; "any" and "anycrlf" were OK).
1924    
1925        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1926            mode when an empty string match preceded an ASCII character followed by
1927            a non-ASCII character. (The code for advancing by one character rather
1928            than one byte was nonsense.)
1929    
1930        (c) The pcredemo.c sample program did not have any code at all to handle
1931            the cases when CRLF is a valid newline sequence.
1932    
1933    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1934        as a starting offset was within the subject string. There is now a new
1935        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1936        negative or greater than the length of the string. In order to test this,
1937        pcretest is extended to allow the setting of negative starting offsets.
1938    
1939    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1940        starting offset points to the beginning of a UTF-8 character was
1941        unnecessarily clumsy. I tidied it up.
1942    
1943    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1944        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1945    
1946    17. Nobody had reported that the --include_dir option, which was added in
1947        release 7.7 should have been called --include-dir (hyphen, not underscore)
1948        for compatibility with GNU grep. I have changed it to --include-dir, but
1949        left --include_dir as an undocumented synonym, and the same for
1950        --exclude-dir, though that is not available in GNU grep, at least as of
1951        release 2.5.4.
1952    
1953    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1954        characters from a string of bytes) have been redefined so as not to use
1955        loops, in order to improve performance in some environments. At the same
1956        time, I abstracted some of the common code into auxiliary macros to save
1957        repetition (this should not affect the compiled code).
1958    
1959    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1960        compile-time error is now given if \c is not followed by an ASCII
1961        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1962        different, and any byte value is allowed.)
1963    
1964    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1965        START_OPTIMIZE option, which is now allowed at compile time - but just
1966        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1967        to pcregrep and other applications that have no direct access to PCRE
1968        options. The new /Y option in pcretest sets this option when calling
1969        pcre_compile().
1970    
1971    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1972        back references. Groups containing recursive back references were forced to
1973        be atomic by that change, but in the case of named groups, the amount of
1974        memory required was incorrectly computed, leading to "Failed: internal
1975        error: code overflow". This has been fixed.
1976    
1977    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1978        pcretest.c, to avoid build problems in some Borland environments.
1979    
1980    
1981    Version 8.10 25-Jun-2010
1982    ------------------------
1983    
1984    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1985        THEN.
1986    
1987    2.  (*ACCEPT) was not working when inside an atomic group.
1988    
1989    3.  Inside a character class, \B is treated as a literal by default, but
1990        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1991        causes the error). The code is unchanged, but I tidied the documentation.
1992    
1993    4.  Inside a character class, PCRE always treated \R and \X as literals,
1994        whereas Perl faults them if its -w option is set. I have changed PCRE so
1995        that it faults them when PCRE_EXTRA is set.
1996    
1997    5.  Added support for \N, which always matches any character other than
1998        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1999    
2000    6.  When compiling pcregrep with newer versions of gcc which may have
2001        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2002        declared with attribute warn_unused_result" were given. Just casting the
2003        result to (void) does not stop the warnings; a more elaborate fudge is
2004        needed. I've used a macro to implement this.
2005    
2006    7.  Minor change to pcretest.c to avoid a compiler warning.
2007    
2008    8.  Added four artifical Unicode properties to help with an option to make
2009        \s etc use properties (see next item). The new properties are: Xan
2010        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2011    
2012    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2013        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2014        this option. Modified pcretest to add /W to test this facility. Added
2015        REG_UCP to make it available via the POSIX interface.
2016    
2017    10. Added --line-buffered to pcregrep.
2018    
2019    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2020        studied, and the match started with a letter with a code point greater than
2021        127 whose first byte was different to the first byte of the other case of
2022        the letter, the other case of this starting letter was not recognized
2023        (#976).
2024    
2025    12. If a pattern that was studied started with a repeated Unicode property
2026        test, for example, \p{Nd}+, there was the theoretical possibility of
2027        setting up an incorrect bitmap of starting bytes, but fortunately it could
2028        not have actually happened in practice until change 8 above was made (it
2029        added property types that matched character-matching opcodes).
2030    
2031    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2032        possible starting bytes for non-anchored patterns.
2033    
2034    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2035        \R, and also a number of cases that involve Unicode properties, both
2036        explicit and implicit when PCRE_UCP is set.
2037    
2038    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2039        input, it could crash or give wrong results if characters with values
2040        greater than 0xc0 were present in the subject string. (Detail: it assumed
2041        UTF-8 input when processing these items.)
2042    
2043    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2044        size_t is 64-bit (#991).
2045    
2046    17. Added a check for running out of memory when PCRE is compiled with
2047        --disable-stack-for-recursion (#990).
2048    
2049    18. If the last data line in a file for pcretest does not have a newline on
2050        the end, a newline was missing in the output.
2051    
2052    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2053        less than 128) in its various bitmaps. However, there is a facility for
2054        generating tables according to the current locale when PCRE is compiled. It
2055        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2056        characters, are recognized by isspace() and therefore were getting set in
2057        these tables, and indeed these tables seem to approximate to ISO 8859. This
2058        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2059        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2060        which of course cannot start UTF-8 characters. I have changed the code so
2061        that only real ASCII characters (less than 128) and the correct starting
2062        bytes for UTF-8 encodings are set for characters greater than 127 when in
2063        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2064        altogether.)
2065    
2066    20. Added the /T option to pcretest so as to be able to run tests with non-
2067        standard character tables, thus making it possible to include the tests
2068        used for 19 above in the standard set of tests.
2069    
2070    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2071        reference to a subpattern the other side of a comment that contains an
2072        opening parenthesis caused either an internal compiling error, or a
2073        reference to the wrong subpattern.
2074    
2075    
2076    Version 8.02 19-Mar-2010
2077  ------------------------  ------------------------
2078    
2079  1.  The Unicode data tables have been updated to Unicode 5.2.0.  1.  The Unicode data tables have been updated to Unicode 5.2.0.
2080    
2081    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2082        configured.
2083    
2084    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2085        original author of that file, following a query about its status.
2086    
2087    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2088        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2089    
2090    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2091        quantifier applied to a forward-referencing subroutine call, could compile
2092        incorrect code or give the error "internal error: previously-checked
2093        referenced subpattern not found".
2094    
2095    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2096        variables to point to external functions. For these systems, therefore,
2097        pcre_malloc etc. are now initialized to local functions that call the
2098        relevant global functions.
2099    
2100    7.  There were two entries missing in the vectors called coptable and poptable
2101        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2102        I've fixed the data, and added a kludgy way of testing at compile time that
2103        the lengths are correct (equal to the number of opcodes).
2104    
2105    8.  Following on from 7, I added a similar kludge to check the length of the
2106        eint vector in pcreposix.c.
2107    
2108    9.  Error texts for pcre_compile() are held as one long string to avoid too
2109        much relocation at load time. To find a text, the string is searched,
2110        counting zeros. There was no check for running off the end of the string,
2111        which could happen if a new error number was added without updating the
2112        string.
2113    
2114    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2115    
2116    11. \K was not working if it appeared in an atomic group or in a group that
2117        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2118        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2119        the assertion is positive, but not if it is negative.
2120    
2121    12. Change 11 fortuitously reduced the size of the stack frame used in the
2122        "match()" function of pcre_exec.c by one pointer. Forthcoming
2123        implementation of support for (*MARK) will need an extra pointer on the
2124        stack; I have reserved it now, so that the stack frame size does not
2125        decrease.
2126    
2127    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2128        item in branch that calls a recursion is a subroutine call - as in the
2129        second branch in the above example - was incorrectly given the compile-
2130        time error "recursive call could loop indefinitely" because pcre_compile()
2131        was not correctly checking the subroutine for matching a non-empty string.
2132    
2133    14. The checks for overrunning compiling workspace could trigger after an
2134        overrun had occurred. This is a "should never occur" error, but it can be
2135        triggered by pathological patterns such as hundreds of nested parentheses.
2136        The checks now trigger 100 bytes before the end of the workspace.
2137    
2138    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2139    
2140    
2141  Version 8.01 19-Jan-2010  Version 8.01 19-Jan-2010
2142  ------------------------  ------------------------
# Line 333  Version 7.9 11-Apr-09 Line 2464  Version 7.9 11-Apr-09
2464  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
2465      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
2466      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2467      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2468        8.33.]
2469    
2470  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
2471      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 576  Version 7.7 07-May-08 Line 2708  Version 7.7 07-May-08
2708      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2709      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2710      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2711      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2712      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2713      was subsequently set up correctly.)      was subsequently set up correctly.)
2714    

Legend:
Removed from v.491  
changed lines
  Added in v.1615

  ViewVC Help
Powered by ViewVC 1.1.5