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

Diff of /code/trunk/ChangeLog

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

revision 221 by ph10, Fri Aug 17 09:25:08 2007 UTC revision 1614 by ph10, Sun Nov 29 17:38:25 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 17-Aug-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 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    
25    Version 8.38 23-November-2015
26    -----------------------------
27    
28    1.  If a group that contained a recursive back reference also contained a
29        forward reference subroutine call followed by a non-forward-reference
30        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
31        compile correct code, leading to undefined behaviour or an internally
32        detected error. This bug was discovered by the LLVM fuzzer.
33    
34    2.  Quantification of certain items (e.g. atomic back references) could cause
35        incorrect code to be compiled when recursive forward references were
36        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
37        This bug was discovered by the LLVM fuzzer.
38    
39    3.  A repeated conditional group whose condition was a reference by name caused
40        a buffer overflow if there was more than one group with the given name.
41        This bug was discovered by the LLVM fuzzer.
42    
43    4.  A recursive back reference by name within a group that had the same name as
44        another group caused a buffer overflow. For example:
45        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
46    
47    5.  A forward reference by name to a group whose number is the same as the
48        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
49        a buffer overflow at compile time. This bug was discovered by the LLVM
50        fuzzer.
51    
52    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
53        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
54    
55    7.  Another buffer overflow bug involved duplicate named groups with a
56        reference between their definition, with a group that reset capture
57        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
58        fixed by always allowing for more memory, even if not needed. (A proper fix
59        is implemented in PCRE2, but it involves more refactoring.)
60    
61    8.  There was no check for integer overflow in subroutine calls such as (?123).
62    
63    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
64        being treated as a literal 'l' instead of causing an error.
65    
66    10. There was a buffer overflow if pcre_exec() was called with an ovector of
67        size 1. This bug was found by american fuzzy lop.
68    
69    11. If a non-capturing group containing a conditional group that could match
70        an empty string was repeated, it was not identified as matching an empty
71        string itself. For example: /^(?:(?(1)x|)+)+$()/.
72    
73    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
74        \a and \e in test subject lines.
75    
76    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
77        instead of the EBCDIC value.
78    
79    14. The handling of \c in an EBCDIC environment has been revised so that it is
80        now compatible with the specification in Perl's perlebcdic page.
81    
82    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
83        ASCII/Unicode. This has now been added to the list of characters that are
84        recognized as white space in EBCDIC.
85    
86    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
87        error (correctly) when used outside a class, but did not give an error
88        within a class.
89    
90    17. \h within a class was incorrectly compiled in EBCDIC environments.
91    
92    18. A pattern with an unmatched closing parenthesis that contained a backward
93        assertion which itself contained a forward reference caused buffer
94        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
95    
96    19. JIT should return with error when the compiled pattern requires more stack
97        space than the maximum.
98    
99    20. A possessively repeated conditional group that could match an empty string,
100        for example, /(?(R))*+/, was incorrectly compiled.
101    
102    21. Fix infinite recursion in the JIT compiler when certain patterns such as
103        /(?:|a|){100}x/ are analysed.
104    
105    22. Some patterns with character classes involving [: and \\ were incorrectly
106        compiled and could cause reading from uninitialized memory or an incorrect
107        error diagnosis.
108    
109    23. Pathological patterns containing many nested occurrences of [: caused
110        pcre_compile() to run for a very long time.
111    
112    24. A conditional group with only one branch has an implicit empty alternative
113        branch and must therefore be treated as potentially matching an empty
114        string.
115    
116    25. If (?R was followed by - or + incorrect behaviour happened instead of a
117        diagnostic.
118    
119    26. Arrange to give up on finding the minimum matching length for overly
120        complex patterns.
121    
122    27. Similar to (4) above: in a pattern with duplicated named groups and an
123        occurrence of (?| it is possible for an apparently non-recursive back
124        reference to become recursive if a later named group with the relevant
125        number is encountered. This could lead to a buffer overflow. Wen Guanxing
126        from Venustech ADLAB discovered this bug.
127    
128    28. If pcregrep was given the -q option with -c or -l, or when handling a
129        binary file, it incorrectly wrote output to stdout.
130    
131    29. The JIT compiler did not restore the control verb head in case of *THEN
132        control verbs. This issue was found by Karl Skomski with a custom LLVM
133        fuzzer.
134    
135    30. Error messages for syntax errors following \g and \k were giving inaccurate
136        offsets in the pattern.
137    
138    31. Added a check for integer overflow in conditions (?(<digits>) and
139        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
140        fuzzer.
141    
142    32. Handling recursive references such as (?2) when the reference is to a group
143        later in the pattern uses code that is very hacked about and error-prone.
144        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
145        give an internal error if it is obvious that compiling has gone wrong.
146    
147    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
148        This issue was found by Karl Skomski with a custom LLVM fuzzer.
149    
150    34. The JIT compiler should restore the control chain for empty possessive
151        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
152    
153    35. Match limit check added to JIT recursion. This issue was found by Karl
154        Skomski with a custom LLVM fuzzer.
155    
156    36. Yet another case similar to 27 above has been circumvented by an
157        unconditional allocation of extra memory. This issue is fixed "properly" in
158        PCRE2 by refactoring the way references are handled. Wen Guanxing
159        from Venustech ADLAB discovered this bug.
160    
161    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
162        with a custom LLVM fuzzer.
163    
164    38. Fixed a corner case of range optimization in JIT.
165    
166    39. An incorrect error "overran compiling workspace" was given if there were
167        exactly enough group forward references such that the last one extended
168        into the workspace safety margin. The next one would have expanded the
169        workspace. The test for overflow was not including the safety margin.
170    
171    40. A match limit issue is fixed in JIT which was found by Karl Skomski
172        with a custom LLVM fuzzer.
173    
174    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
175        work under Windows. (Why has it taken so long for anyone to notice?)
176    
177    42. In a character class such as [\W\p{Any}] where both a negative-type escape
178        ("not a word character") and a property escape were present, the property
179        escape was being ignored.
180    
181    43. Fix crash caused by very long (*MARK) or (*THEN) names.
182    
183    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
184        by a single ASCII character in a class item, was incorrectly compiled in
185        UCP mode. The POSIX class got lost, but only if the single character
186        followed it.
187    
188    45. [:punct:] in UCP mode was matching some characters in the range 128-255
189        that should not have been matched.
190    
191    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
192        class, all characters with code points greater than 255 are in the class.
193        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
194        such as \w are turned into Unicode properties), wide characters were not
195        correctly handled, and could fail to match.
196    
197    
198    Version 8.37 28-April-2015
199    --------------------------
200    
201    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
202        for those parentheses to be closed with whatever has been captured so far.
203        However, it was failing to mark any other groups between the hightest
204        capture so far and the currrent group as "unset". Thus, the ovector for
205        those groups contained whatever was previously there. An example is the
206        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
207    
208    2.  If an assertion condition was quantified with a minimum of zero (an odd
209        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
210    
211    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
212        unrecognized modifier, a crash could occur.
213    
214    4.  An attempt to do global matching in pcretest with a zero-length ovector
215        caused a crash.
216    
217    5.  Fixed a memory leak during matching that could occur for a subpattern
218        subroutine call (recursive or otherwise) if the number of captured groups
219        that had to be saved was greater than ten.
220    
221    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
222        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
223        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
224    
225    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
226        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
227    
228    8. When a pattern is compiled, it remembers the highest back reference so that
229       when matching, if the ovector is too small, extra memory can be obtained to
230       use instead. A conditional subpattern whose condition is a check on a
231       capture having happened, such as, for example in the pattern
232       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
233       setting the highest backreference number. This mattered only if pcre_exec()
234       was called with an ovector that was too small to hold the capture, and there
235       was no other kind of back reference (a situation which is probably quite
236       rare). The effect of the bug was that the condition was always treated as
237       FALSE when the capture could not be consulted, leading to a incorrect
238       behaviour by pcre_exec(). This bug has been fixed.
239    
240    9. A reference to a duplicated named group (either a back reference or a test
241       for being set in a conditional) that occurred in a part of the pattern where
242       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
243       to be incorrectly calculated, leading to overwriting.
244    
245    10. A mutually recursive set of back references such as (\2)(\1) caused a
246        segfault at study time (while trying to find the minimum matching length).
247        The infinite loop is now broken (with the minimum length unset, that is,
248        zero).
249    
250    11. If an assertion that was used as a condition was quantified with a minimum
251        of zero, matching went wrong. In particular, if the whole group had
252        unlimited repetition and could match an empty string, a segfault was
253        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
254        assertions to be quantified, but not if they are being used as conditions,
255        so the above pattern is faulted by Perl. PCRE has now been changed so that
256        it also rejects such patterns.
257    
258    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
259        failed to allow the zero-repeat case if pcre2_exec() was called with an
260        ovector too small to capture the group.
261    
262    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
263        Red Hat Product Security:
264    
265        (a) A crash if /K and /F were both set with the option to save the compiled
266        pattern.
267    
268        (b) Another crash if the option to print captured substrings in a callout
269        was combined with setting a null ovector, for example \O\C+ as a subject
270        string.
271    
272    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
273        forward reference repeated a large (but limited) number of times within a
274        repeated outer group that has a zero minimum quantifier, caused incorrect
275        code to be compiled, leading to the error "internal error:
276        previously-checked referenced subpattern not found" when an incorrect
277        memory address was read. This bug was reported as "heap overflow",
278        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
279        CVE-2015-2325.
280    
281    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
282        call within a group that also contained a recursive back reference caused
283        incorrect code to be compiled. This bug was reported as "heap overflow",
284        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
285        number CVE-2015-2326.
286    
287    24. Computing the size of the JIT read-only data in advance has been a source
288        of various issues, and new ones are still appear unfortunately. To fix
289        existing and future issues, size computation is eliminated from the code,
290        and replaced by on-demand memory allocation.
291    
292    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
293        adjacent to the end of the range, and the range contained characters with
294        more than one other case, caused incorrect behaviour when compiled in UTF
295        mode. In that example, the range a-j was left out of the class.
296    
297    26. Fix JIT compilation of conditional blocks, which assertion
298        is converted to (*FAIL). E.g: /(?(?!))/.
299    
300    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
301        discovered by the LLVM fuzzer.
302    
303    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
304        when this assertion was used as a condition, for example (?(?!)a|b). In
305        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
306        error about an unsupported item.
307    
308    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
309        possessification code could take exponential time to complete. A recursion
310        depth limit of 1000 has been imposed to limit the resources used by this
311        optimization.
312    
313    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
314        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
315        because \S ensures they are all in the class. The code for doing this was
316        interacting badly with the code for computing the amount of space needed to
317        compile the pattern, leading to a buffer overflow. This bug was discovered
318        by the LLVM fuzzer.
319    
320    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
321        other kinds of group caused stack overflow at compile time. This bug was
322        discovered by the LLVM fuzzer.
323    
324    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
325        between a subroutine call and its quantifier was incorrectly compiled,
326        leading to buffer overflow or other errors. This bug was discovered by the
327        LLVM fuzzer.
328    
329    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
330        assertion after (?(. The code was failing to check the character after
331        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
332        was discovered by the LLVM fuzzer.
333    
334    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
335        a fixed maximum following a group that contains a subroutine reference was
336        incorrectly compiled and could trigger buffer overflow. This bug was
337        discovered by the LLVM fuzzer.
338    
339    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
340        caused a stack overflow instead of the diagnosis of a non-fixed length
341        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
342    
343    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
344        (e.g. /(?<=\Ka)/) could make pcregrep loop.
345    
346    37. There was a similar problem to 36 in pcretest for global matches.
347    
348    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
349        and a subsequent item in the pattern caused a non-match, backtracking over
350        the repeated \X did not stop, but carried on past the start of the subject,
351        causing reference to random memory and/or a segfault. There were also some
352        other cases where backtracking after \C could crash. This set of bugs was
353        discovered by the LLVM fuzzer.
354    
355    39. The function for finding the minimum length of a matching string could take
356        a very long time if mutual recursion was present many times in a pattern,
357        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
358        method has been implemented. This infelicity was discovered by the LLVM
359        fuzzer.
360    
361    40. Static linking against the PCRE library using the pkg-config module was
362        failing on missing pthread symbols.
363    
364    
365    Version 8.36 26-September-2014
366    ------------------------------
367    
368    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
369        -Wmissing-field-initializers and -Wunused-parameter.
370    
371    2.  The tests for quantifiers being too big (greater than 65535) were being
372        applied after reading the number, and stupidly assuming that integer
373        overflow would give a negative number. The tests are now applied as the
374        numbers are read.
375    
376    3.  Tidy code in pcre_exec.c where two branches that used to be different are
377        now the same.
378    
379    4.  The JIT compiler did not generate match limit checks for certain
380        bracketed expressions with quantifiers. This may lead to exponential
381        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
382        issue should be resolved now.
383    
384    5.  Fixed an issue, which occures when nested alternatives are optimized
385        with table jumps.
386    
387    6.  Inserted two casts and changed some ints to size_t in the light of some
388        reported 64-bit compiler warnings (Bugzilla 1477).
389    
390    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
391        an empty string, which sometimes were behaving incorrectly in the
392        interpreter (though correctly in the JIT matcher). This pcretest input is
393        an example:
394    
395          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
396          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
397    
398        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
399        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
400        for an empty string was breaking the inner loop and carrying on at a lower
401        level, when possessive repeated groups should always return to a higher
402        level as they have no backtrack points in them. The empty string test now
403        occurs at the outer level.
404    
405    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
406        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
407    
408    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
409        doesn't).
410    
411    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
412        to hold a pattern to be incorrectly computed (too small) when there were
413        named back references to duplicated names. This could cause "internal
414        error: code overflow" or "double free or corruption" or other memory
415        handling errors.
416    
417    11. When named subpatterns had the same prefixes, back references could be
418        confused. For example, in this pattern:
419    
420          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
421    
422        the reference to 'Name' was incorrectly treated as a reference to a
423        duplicate name.
424    
425    12. A pattern such as /^s?c/mi8 where the optional character has more than
426        one "other case" was incorrectly compiled such that it would only try to
427        match starting at "c".
428    
429    13. When a pattern starting with \s was studied, VT was not included in the
430        list of possible starting characters; this should have been part of the
431        8.34/18 patch.
432    
433    14. If a character class started [\Qx]... where x is any character, the class
434        was incorrectly terminated at the ].
435    
436    15. If a pattern that started with a caseless match for a character with more
437        than one "other case" was studied, PCRE did not set up the starting code
438        unit bit map for the list of possible characters. Now it does. This is an
439        optimization improvement, not a bug fix.
440    
441    16. The Unicode data tables have been updated to Unicode 7.0.0.
442    
443    17. Fixed a number of memory leaks in pcregrep.
444    
445    18. Avoid a compiler warning (from some compilers) for a function call with
446        a cast that removes "const" from an lvalue by using an intermediate
447        variable (to which the compiler does not object).
448    
449    19. Incorrect code was compiled if a group that contained an internal recursive
450        back reference was optional (had quantifier with a minimum of zero). This
451        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
452        caused segmentation faults because of stack overflows at compile time.
453    
454    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
455        group that is quantified with an indefinite repeat, caused a compile-time
456        loop which used up all the system stack and provoked a segmentation fault.
457        This was not the same bug as 19 above.
458    
459    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
460        Patch by Mike Frysinger.
461    
462    
463    Version 8.35 04-April-2014
464    --------------------------
465    
466    1.  A new flag is set, when property checks are present in an XCLASS.
467        When this flag is not set, PCRE can perform certain optimizations
468        such as studying these XCLASS-es.
469    
470    2.  The auto-possessification of character sets were improved: a normal
471        and an extended character set can be compared now. Furthermore
472        the JIT compiler optimizes more character set checks.
473    
474    3.  Got rid of some compiler warnings for potentially uninitialized variables
475        that show up only when compiled with -O2.
476    
477    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
478        of a match later then the end of the match. The pcretest program was not
479        handling the case sensibly - it was outputting from the start to the next
480        binary zero. It now reports this situation in a message, and outputs the
481        text from the end to the start.
482    
483    5.  Fast forward search is improved in JIT. Instead of the first three
484        characters, any three characters with fixed position can be searched.
485        Search order: first, last, middle.
486    
487    6.  Improve character range checks in JIT. Characters are read by an inprecise
488        function now, which returns with an unknown value if the character code is
489        above a certain threshold (e.g: 256). The only limitation is that the value
490        must be bigger than the threshold as well. This function is useful when
491        the characters above the threshold are handled in the same way.
492    
493    7.  The macros whose names start with RAWUCHAR are placeholders for a future
494        mode in which only the bottom 21 bits of 32-bit data items are used. To
495        make this more memorable for those maintaining the code, the names have
496        been changed to start with UCHAR21, and an extensive comment has been added
497        to their definition.
498    
499    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
500        to the export list in Makefile.am (they were accidentally omitted from the
501        8.34 tarball).
502    
503    9.  The informational output from pcretest used the phrase "starting byte set"
504        which is inappropriate for the 16-bit and 32-bit libraries. As the output
505        for "first char" and "need char" really means "non-UTF-char", I've changed
506        "byte" to "char", and slightly reworded the output. The documentation about
507        these values has also been (I hope) clarified.
508    
509    10. Another JIT related optimization: use table jumps for selecting the correct
510        backtracking path, when more than four alternatives are present inside a
511        bracket.
512    
513    11. Empty match is not possible, when the minimum length is greater than zero,
514        and there is no \K in the pattern. JIT should avoid empty match checks in
515        such cases.
516    
517    12. In a caseless character class with UCP support, when a character with more
518        than one alternative case was not the first character of a range, not all
519        the alternative cases were added to the class. For example, s and \x{17f}
520        are both alternative cases for S: the class [RST] was handled correctly,
521        but [R-T] was not.
522    
523    13. The configure.ac file always checked for pthread support when JIT was
524        enabled. This is not used in Windows, so I have put this test inside a
525        check for the presence of windows.h (which was already tested for).
526    
527    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
528        The algorithm provides a way to skip certain starting offsets, and usually
529        faster than linear prefix searches.
530    
531    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
532        as for 'fr_FR' and 'french'. For some reason, however, it then used the
533        Windows-specific input and output files, which have 'french' screwed in.
534        So this could never have worked. One of the problems with locales is that
535        they aren't always the same. I have now updated RunTest so that it checks
536        the output of the locale test (test 3) against three different output
537        files, and it allows the test to pass if any one of them matches. With luck
538        this should make the test pass on some versions of Solaris where it was
539        failing. Because of the uncertainty, the script did not used to stop if
540        test 3 failed; it now does. If further versions of a French locale ever
541        come to light, they can now easily be added.
542    
543    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
544        there was a message during ./configure, but it did not stop. This now
545        provokes an error. The invalid example in README has been corrected.
546        If a value less than the minimum is given, the minimum value has always
547        been used, but now a warning is given.
548    
549    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
550        was a bug in the test system, which is now fixed. Also, the list of various
551        configurations that are tested for each release did not have one with both
552        16/32 bits and --enable-bar-anycrlf. It now does.
553    
554    18. pcretest was missing "-C bsr" for displaying the \R default setting.
555    
556    19. Little endian PowerPC systems are supported now by the JIT compiler.
557    
558    20. The fast forward newline mechanism could enter to an infinite loop on
559        certain invalid UTF-8 input. Although we don't support these cases
560        this issue can be fixed by a performance optimization.
561    
562    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
563        not take account if existing stack usage. There is now a new global
564        variable called pcre_stack_guard that can be set to point to an external
565        function to check stack availability. It is called at the start of
566        processing every parenthesized group.
567    
568    22. A typo in the code meant that in ungreedy mode the max/min qualifier
569        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
570        not match "ab".
571    
572    23. When UTF was disabled, the JIT program reported some incorrect compile
573        errors. These messages are silenced now.
574    
575    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
576        compiler.
577    
578    25. Change all the temporary files used in RunGrepTest to be different to those
579        used by RunTest so that the tests can be run simultaneously, for example by
580        "make -j check".
581    
582    
583    Version 8.34 15-December-2013
584    -----------------------------
585    
586    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
587        executable memory. Patch inspired by Carsten Klein.
588    
589    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
590        this macro is never tested and has no effect, because the work to support
591        coverage involves only compiling and linking options and special targets in
592        the Makefile. The comment in config.h implied that defining the macro would
593        enable coverage support, which is totally false. There was also support for
594        setting this macro in the CMake files (my fault, I just copied it from
595        configure). SUPPORT_GCOV has now been removed.
596    
597    3.  Make a small performance improvement in strlen16() and strlen32() in
598        pcretest.
599    
600    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
601        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
602        these cases). There was also one in pcretest.c.
603    
604    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
605    
606    6.  In UTF mode, the code for checking whether a group could match an empty
607        string (which is used for indefinitely repeated groups to allow for
608        breaking an infinite loop) was broken when the group contained a repeated
609        negated single-character class with a character that occupied more than one
610        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
611        UTF-8 mode). The effect was undefined: the group might or might not be
612        deemed as matching an empty string, or the program might have crashed.
613    
614    7.  The code for checking whether a group could match an empty string was not
615        recognizing that \h, \H, \v, \V, and \R must match a character.
616    
617    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
618        an empty string. If it can, pcretest shows this in its information output.
619    
620    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
621        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
622        matched by pcre_exec() without using JIT:
623    
624        (a) If the rest of the pattern did not match after a maximal run of
625            grapheme clusters, the code for backing up to try with fewer of them
626            did not always back up over a full grapheme when characters that do not
627            have the modifier quality were involved, e.g. Hangul syllables.
628    
629        (b) If the match point in a subject started with modifier character, and
630            there was no match, the code could incorrectly back up beyond the match
631            point, and potentially beyond the first character in the subject,
632            leading to a segfault or an incorrect match result.
633    
634    10. A conditional group with an assertion condition could lead to PCRE
635        recording an incorrect first data item for a match if no other first data
636        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
637        first data item, and therefore matched "ca" after "c" instead of at the
638        start.
639    
640    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
641        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
642    
643    12. The source of pcregrep now includes z/OS-specific code so that it can be
644        compiled for z/OS as part of the special z/OS distribution.
645    
646    13. Added the -T and -TM options to pcretest.
647    
648    14. The code in pcre_compile.c for creating the table of named capturing groups
649        has been refactored. Instead of creating the table dynamically during the
650        actual compiling pass, the information is remembered during the pre-compile
651        pass (on the stack unless there are more than 20 named groups, in which
652        case malloc() is used) and the whole table is created before the actual
653        compile happens. This has simplified the code (it is now nearly 150 lines
654        shorter) and prepared the way for better handling of references to groups
655        with duplicate names.
656    
657    15. A back reference to a named subpattern when there is more than one of the
658        same name now checks them in the order in which they appear in the pattern.
659        The first one that is set is used for the reference. Previously only the
660        first one was inspected. This change makes PCRE more compatible with Perl.
661    
662    16. Unicode character properties were updated from Unicode 6.3.0.
663    
664    17. The compile-time code for auto-possessification has been refactored, based
665        on a patch by Zoltan Herczeg. It now happens after instead of during
666        compilation. The code is cleaner, and more cases are handled. The option
667        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
668        options in pcretest are provided to set it. It can also be set by
669        (*NO_AUTO_POSSESS) at the start of a pattern.
670    
671    18. The character VT has been added to the default ("C" locale) set of
672        characters that match \s and are generally treated as white space,
673        following this same change in Perl 5.18. There is now no difference between
674        "Perl space" and "POSIX space". Whether VT is treated as white space in
675        other locales depends on the locale.
676    
677    19. The code for checking named groups as conditions, either for being set or
678        for being recursed, has been refactored (this is related to 14 and 15
679        above). Processing unduplicated named groups should now be as fast at
680        numerical groups, and processing duplicated groups should be faster than
681        before.
682    
683    20. Two patches to the CMake build system, by Alexander Barkov:
684    
685          (1) Replace the "source" command by "." in CMakeLists.txt because
686              "source" is a bash-ism.
687    
688          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
689              without these the CMake build does not work on Solaris.
690    
691    21. Perl has changed its handling of \8 and \9. If there is no previously
692        encountered capturing group of those numbers, they are treated as the
693        literal characters 8 and 9 instead of a binary zero followed by the
694        literals. PCRE now does the same.
695    
696    22. Following Perl, added \o{} to specify codepoints in octal, making it
697        possible to specify values greater than 0777 and also making them
698        unambiguous.
699    
700    23. Perl now gives an error for missing closing braces after \x{... instead of
701        treating the string as literal. PCRE now does the same.
702    
703    24. RunTest used to grumble if an inappropriate test was selected explicitly,
704        but just skip it when running all tests. This make it awkward to run ranges
705        of tests when one of them was inappropriate. Now it just skips any
706        inappropriate tests, as it always did when running all tests.
707    
708    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
709        character types such as \d or \w, too many callouts were inserted, and the
710        data that they returned was rubbish.
711    
712    26. In UCP mode, \s was not matching two of the characters that Perl matches,
713        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
714        were matched by \h. The code has now been refactored so that the lists of
715        the horizontal and vertical whitespace characters used for \h and \v (which
716        are defined only in one place) are now also used for \s.
717    
718    27. Add JIT support for the 64 bit TileGX architecture.
719        Patch by Jiong Wang (Tilera Corporation).
720    
721    28. Possessive quantifiers for classes (both explicit and automatically
722        generated) now use special opcodes instead of wrapping in ONCE brackets.
723    
724    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
725        (because it's meaningless), this was not happening when PCRE_CASELESS was
726        set. Not wrong, but inefficient.
727    
728    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
729        properties for \w, \d, etc) is present in a test regex. Otherwise if the
730        test contains no characters greater than 255, Perl doesn't realise it
731        should be using Unicode semantics.
732    
733    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
734        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
735        does in Unicode mode.
736    
737    32. Added the "forbid" facility to pcretest so that putting tests into the
738        wrong test files can sometimes be quickly detected.
739    
740    33. There is now a limit (default 250) on the depth of nesting of parentheses.
741        This limit is imposed to control the amount of system stack used at compile
742        time. It can be changed at build time by --with-parens-nest-limit=xxx or
743        the equivalent in CMake.
744    
745    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
746        errors. Perl warns for these when in warning mode, but PCRE has no facility
747        for giving warnings.
748    
749    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
750        However, this was not working for (?!) because it is optimized to (*FAIL),
751        for which PCRE does not allow quantifiers. The optimization is now disabled
752        when a quantifier follows (?!). I can't see any use for this, but it makes
753        things uniform.
754    
755    36. Perl no longer allows group names to start with digits, so I have made this
756        change also in PCRE. It simplifies the code a bit.
757    
758    37. In extended mode, Perl ignores spaces before a + that indicates a
759        possessive quantifier. PCRE allowed a space before the quantifier, but not
760        before the possessive +. It now does.
761    
762    38. The use of \K (reset reported match start) within a repeated possessive
763        group such as (a\Kb)*+ was not working.
764    
765    40. Document that the same character tables must be used at compile time and
766        run time, and that the facility to pass tables to pcre_exec() and
767        pcre_dfa_exec() is for use only with saved/restored patterns.
768    
769    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
770        features for Builds with MSVC:
771    
772        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
773           builds). The .rc files can be used to set FileDescription and many other
774           attributes.
775    
776        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
777           This allows higher-level build scripts which want .pdb files to avoid
778           hard-coding the exact files needed."
779    
780    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
781        mean "start of word" and "end of word", respectively, as a transition aid.
782    
783    43. A minimizing repeat of a class containing codepoints greater than 255 in
784        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
785        compiled to use the heap for recursion.
786    
787    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
788        is configured.
789    
790    
791    Version 8.33 28-May-2013
792    ------------------------
793    
794    1.  Added 'U' to some constants that are compared to unsigned integers, to
795        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
796        variables that are added to signed variables, to ensure the result is
797        signed and can be negated.
798    
799    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
800        CMake config files.
801    
802    3.  Revise the creation of config.h.generic so that all boolean macros are
803        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
804        overriding via -D on the command line possible.
805    
806    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
807        to unsigned int is reported to make a quite noticeable speed difference in
808        a specific Windows environment. Testing on Linux did also appear to show
809        some benefit (and it is clearly not harmful). Also fixed the definition of
810        Xop which should be unsigned.
811    
812    5.  Related to (4), changing the definition of the intermediate variable cc
813        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
814        improvements.
815    
816    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
817        unnecessary spaces.
818    
819    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
820        and later.
821    
822    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
823    
824    9.  Optimizing fast_forward_start_bits in JIT.
825    
826    10. Adding support for callouts in JIT, and fixing some issues revealed
827        during this work. Namely:
828    
829        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
830    
831        (b) Minimum length was not checked before the matching is started.
832    
833    11. The value of capture_last that is passed to callouts was incorrect in some
834        cases when there was a capture on one path that was subsequently abandoned
835        after a backtrack. Also, the capture_last value is now reset after a
836        recursion, since all captures are also reset in this case.
837    
838    12. The interpreter no longer returns the "too many substrings" error in the
839        case when an overflowing capture is in a branch that is subsequently
840        abandoned after a backtrack.
841    
842    13. In the pathological case when an offset vector of size 2 is used, pcretest
843        now prints out the matched string after a yield of 0 or 1.
844    
845    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
846        Only supported by the JIT compiler at the moment.
847    
848    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
849    
850    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
851        offset of the starting point of the matching process, provided the offsets
852        vector is large enough.
853    
854    17. The \A escape now records a lookbehind value of 1, though its execution
855        does not actually inspect the previous character. This is to ensure that,
856        in partial multi-segment matching, at least one character from the old
857        segment is retained when a new segment is processed. Otherwise, if there
858        are no lookbehinds in the pattern, \A might match incorrectly at the start
859        of a new segment.
860    
861    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
862    
863    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
864        gives some modest performance improvement in 8-bit mode.
865    
866    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
867        be expressed in certain programming languages using Universal Character
868        Names.
869    
870    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
871        which points out that "non characters" are not "characters that may not
872        appear in Unicode strings" but rather "characters that are reserved for
873        internal use and have only local meaning".
874    
875    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
876        there was a conditional group that depended on an assertion, if the
877        assertion was false, the callout that immediately followed the alternation
878        in the condition was skipped when pcre_exec() was used for matching.
879    
880    23. Allow an explicit callout to be inserted before an assertion that is the
881        condition for a conditional group, for compatibility with automatic
882        callouts, which always insert a callout at this point.
883    
884    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
885        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
886    
887    25. (*PRUNE) is now supported by the JIT compiler.
888    
889    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
890    
891    27. Fix the case where there are two or more SKIPs with arguments that may be
892        ignored.
893    
894    28. (*SKIP) is now supported by the JIT compiler.
895    
896    29. (*THEN) is now supported by the JIT compiler.
897    
898    30. Update RunTest with additional test selector options.
899    
900    31. The way PCRE handles backtracking verbs has been changed in two ways.
901    
902        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
903        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
904        backtracking. In some cases this makes it more Perl-compatible, but Perl's
905        rather obscure rules do not always do the same thing.
906    
907        (2) Previously, backtracking verbs were confined within assertions. This is
908        no longer the case for positive assertions, except for (*ACCEPT). Again,
909        this sometimes improves Perl compatibility, and sometimes does not.
910    
911    32. A number of tests that were in test 2 because Perl did things differently
912        have been moved to test 1, because either Perl or PCRE has changed, and
913        these tests are now compatible.
914    
915    32. Backtracking control verbs are now handled in the same way in JIT and
916        interpreter.
917    
918    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
919        contained a forward subroutine reference caused a compile error.
920    
921    34. Auto-detect and optimize limited repetitions in JIT.
922    
923    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
924        blocking (*UTF) etc.
925    
926    36. In the interpreter, maximizing pattern repetitions for characters and
927        character types now use tail recursion, which reduces stack usage.
928    
929    37. The value of the max lookbehind was not correctly preserved if a compiled
930        and saved regex was reloaded on a host of different endianness.
931    
932    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
933        of the compiled pattern block, expand the flags field from 16 to 32 bits
934        because it was almost full.
935    
936    39. Try madvise first before posix_madvise.
937    
938    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
939        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
940        original change locked out all matches of empty strings. This has been
941        changed so that one match of an empty string per line is recognized.
942        Subsequent searches on the same line (for colouring or for --only-matching,
943        for example) do not recognize empty strings.
944    
945    41. Applied a user patch to fix a number of spelling mistakes in comments.
946    
947    42. Data lines longer than 65536 caused pcretest to crash.
948    
949    43. Clarified the data type for length and startoffset arguments for pcre_exec
950        and pcre_dfa_exec in the function-specific man pages, where they were
951        explicitly stated to be in bytes, never having been updated. I also added
952        some clarification to the pcreapi man page.
953    
954    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
955        a segmentation fault.
956    
957    
958    Version 8.32 30-November-2012
959    -----------------------------
960    
961    1.  Improved JIT compiler optimizations for first character search and single
962        character iterators.
963    
964    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
965        Patch by Daniel Richard G.
966    
967    3.  Single character iterator optimizations in the JIT compiler.
968    
969    4.  Improved JIT compiler optimizations for character ranges.
970    
971    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
972        Reported by Giuseppe D'Angelo.
973    
974    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
975        of a line, was being set incorrectly in cases where .* appeared inside
976        atomic brackets at the start of a pattern, or where there was a subsequent
977        *PRUNE or *SKIP.
978    
979    7.  Improved instruction cache flush for POWER/PowerPC.
980        Patch by Daniel Richard G.
981    
982    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
983        grep:
984    
985        (a) There is now no limit to the number of patterns to be matched.
986    
987        (b) An error is given if a pattern is too long.
988    
989        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
990            are now supported.
991    
992        (d) --exclude-from and --include-from (multiple use) have been added.
993    
994        (e) Exclusions and inclusions now apply to all files and directories, not
995            just to those obtained from scanning a directory recursively.
996    
997        (f) Multiple uses of -f and --file-list are now supported.
998    
999        (g) In a Windows environment, the default for -d has been changed from
1000            "read" (the GNU grep default) to "skip", because otherwise the presence
1001            of a directory in the file list provokes an error.
1002    
1003        (h) The documentation has been revised and clarified in places.
1004    
1005    9.  Improve the matching speed of capturing brackets.
1006    
1007    10. Changed the meaning of \X so that it now matches a Unicode extended
1008        grapheme cluster.
1009    
1010    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1011        out POSIX threads when JIT support is configured.
1012    
1013    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1014    
1015    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1016        structure could go wrong in environments where size_t is not the same size
1017        as int.
1018    
1019    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1020        set.
1021    
1022    15. The EBCDIC support had decayed; later updates to the code had included
1023        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1024        general tidy up of EBCDIC-related issues, and the documentation was also
1025        not quite right. There is now a test that can be run on ASCII systems to
1026        check some of the EBCDIC-related things (but is it not a full test).
1027    
1028    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1029        in a small tidy to the code.
1030    
1031    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1032    
1033    18. If the --only-matching (-o) option in pcregrep is specified multiple
1034        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1035        substrings matched by the 1st and 2nd capturing parentheses. A separating
1036        string can be specified by --om-separator (default empty).
1037    
1038    19. Improving the first n character searches.
1039    
1040    20. Turn case lists for horizontal and vertical white space into macros so that
1041        they are defined only once.
1042    
1043    21. This set of changes together give more compatible Unicode case-folding
1044        behaviour for characters that have more than one other case when UCP
1045        support is available.
1046    
1047        (a) The Unicode property table now has offsets into a new table of sets of
1048            three or more characters that are case-equivalent. The MultiStage2.py
1049            script that generates these tables (the pcre_ucd.c file) now scans
1050            CaseFolding.txt instead of UnicodeData.txt for character case
1051            information.
1052    
1053        (b) The code for adding characters or ranges of characters to a character
1054            class has been abstracted into a generalized function that also handles
1055            case-independence. In UTF-mode with UCP support, this uses the new data
1056            to handle characters with more than one other case.
1057    
1058        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1059            whose other case is greater than 256 are now correctly matched
1060            caselessly. Previously, the high codepoint matched the low one, but not
1061            vice versa.
1062    
1063        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1064            of the new class addition function, using character lists defined as
1065            macros alongside the case definitions of 20 above.
1066    
1067        (e) Caseless back references now work with characters that have more than
1068            one other case.
1069    
1070        (f) General caseless matching of characters with more than one other case
1071            is supported.
1072    
1073    22. Unicode character properties were updated from Unicode 6.2.0
1074    
1075    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1076    
1077    24. Add support for 32-bit character strings, and UTF-32
1078    
1079    25. Major JIT compiler update (code refactoring and bugfixing).
1080        Experimental Sparc 32 support is added.
1081    
1082    26. Applied a modified version of Daniel Richard G's patch to create
1083        pcre.h.generic and config.h.generic by "make" instead of in the
1084        PrepareRelease script.
1085    
1086    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1087        pcre_compile.c when checking for a zero character.
1088    
1089    28. Introducing a native interface for JIT. Through this interface, the compiled
1090        machine code can be directly executed. The purpose of this interface is to
1091        provide fast pattern matching, so several sanity checks are not performed.
1092        However, feature tests are still performed. The new interface provides
1093        1.4x speedup compared to the old one.
1094    
1095    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1096        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1097        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1098    
1099    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1100        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1101        the "old" RFC 2279). Instead, it ended up passing a negative length to
1102        pcre_exec().
1103    
1104    31. Add support for GCC's visibility feature to hide internal functions.
1105    
1106    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1107        "unknown -C option" after outputting 0 or 1.
1108    
1109    33. There is now support for generating a code coverage report for the test
1110        suite in environments where gcc is the compiler and lcov is installed. This
1111        is mainly for the benefit of the developers.
1112    
1113    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1114        unaddressable using valgrind annotations, allowing valgrind to detect
1115        invalid memory accesses. This is mainly for the benefit of the developers.
1116    
1117    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1118    
1119    26. Give configure error if --enable-cpp but no C++ compiler found.
1120    
1121    
1122    Version 8.31 06-July-2012
1123    -------------------------
1124    
1125    1.  Fixing a wrong JIT test case and some compiler warnings.
1126    
1127    2.  Removed a bashism from the RunTest script.
1128    
1129    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1130        to unsigned type, result still unsigned" that was given by an MS compiler
1131        on encountering the code "-sizeof(xxx)".
1132    
1133    4.  Partial matching support is added to the JIT compiler.
1134    
1135    5.  Fixed several bugs concerned with partial matching of items that consist
1136        of more than one character:
1137    
1138        (a) /^(..)\1/ did not partially match "aba" because checking references was
1139            done on an "all or nothing" basis. This also applied to repeated
1140            references.
1141    
1142        (b) \R did not give a hard partial match if \r was found at the end of the
1143            subject.
1144    
1145        (c) \X did not give a hard partial match after matching one or more
1146            characters at the end of the subject.
1147    
1148        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1149            a partial match for the string "\r".
1150    
1151        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1152            a partial match for a CR character at the end of the subject string.
1153    
1154    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1155        running pcretest, the text "(JIT)" added to the output whenever JIT is
1156        actually used to run the match.
1157    
1158    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1159        or /S+[+] with a digit between 1 and 7.
1160    
1161    8.  OP_NOT now supports any UTF character not just single-byte ones.
1162    
1163    9.  (*MARK) control verb is now supported by the JIT compiler.
1164    
1165    10. The command "./RunTest list" lists the available tests without actually
1166        running any of them. (Because I keep forgetting what they all are.)
1167    
1168    11. Add PCRE_INFO_MAXLOOKBEHIND.
1169    
1170    12. Applied a (slightly modified) user-supplied patch that improves performance
1171        when the heap is used for recursion (compiled with --disable-stack-for-
1172        recursion). Instead of malloc and free for each heap frame each time a
1173        logical recursion happens, frames are retained on a chain and re-used where
1174        possible. This sometimes gives as much as 30% improvement.
1175    
1176    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1177        call.
1178    
1179    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1180    
1181    15. It is now possible to link pcretest with libedit as an alternative to
1182        libreadline.
1183    
1184    16. (*COMMIT) control verb is now supported by the JIT compiler.
1185    
1186    17. The Unicode data tables have been updated to Unicode 6.1.0.
1187    
1188    18. Added --file-list option to pcregrep.
1189    
1190    19. Added binary file support to pcregrep, including the -a, --binary-files,
1191        -I, and --text options.
1192    
1193    20. The madvise function is renamed for posix_madvise for QNX compatibility
1194        reasons. Fixed by Giuseppe D'Angelo.
1195    
1196    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1197    
1198    22. Changed the option for creating symbolic links for 16-bit man pages from
1199        -s to -sf so that re-installing does not cause issues.
1200    
1201    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1202    
1203    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1204        matches in certain environments (the workspace was not being correctly
1205        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1206        some of the workspace data at the beginning of a restart.
1207    
1208    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1209        was not doing so when it should - probably a typo introduced by SVN 528
1210        (change 8.10/14).
1211    
1212    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1213        \w+ when the character tables indicated that \x{c4} was a word character.
1214        There were several related cases, all because the tests for doing a table
1215        lookup were testing for characters less than 127 instead of 255.
1216    
1217    27. If a pattern contains capturing parentheses that are not used in a match,
1218        their slots in the ovector are set to -1. For those that are higher than
1219        any matched groups, this happens at the end of processing. In the case when
1220        there were back references that the ovector was too small to contain
1221        (causing temporary malloc'd memory to be used during matching), and the
1222        highest capturing number was not used, memory off the end of the ovector
1223        was incorrectly being set to -1. (It was using the size of the temporary
1224        memory instead of the true size.)
1225    
1226    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1227        ovector size, it uses memory at the end of the block that it has got.
1228    
1229    29. Check for an overlong MARK name and give an error at compile time. The
1230        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1231    
1232    30. JIT compiler update.
1233    
1234    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1235        Rill for the patch.
1236    
1237    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1238    
1239    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1240    
1241    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1242        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1243        was enabled.
1244    
1245    35. Improve JIT code generation for greedy plus quantifier.
1246    
1247    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1248        1 to "aa" instead of to an empty string. The bug affected repeated groups
1249        that could potentially match an empty string.
1250    
1251    37. Optimizing single character iterators in JIT.
1252    
1253    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1254        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1255        codepoints that are too big for the mode are faulted, and in a UTF mode,
1256        disallowed codepoints are also faulted.
1257    
1258    39. If PCRE was compiled with UTF support, in three places in the DFA
1259        matcher there was code that should only have been obeyed in UTF mode, but
1260        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1261        processing when bytes with values greater than 127 were present. In 16-bit
1262        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1263        both cases the values are those that cannot be the first data item in a UTF
1264        character. The three items that might have provoked this were recursions,
1265        possessively repeated groups, and atomic groups.
1266    
1267    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1268        and pcregrep, because some OS require shared objects to be explicitly
1269        passed to ld, causing the link step to fail if they are not.
1270    
1271    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1272        mode, patterns that started with \h* or \R* might be incorrectly matched.
1273    
1274    
1275    Version 8.30 04-February-2012
1276    -----------------------------
1277    
1278    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1279        name is defined in ctype.h.
1280    
1281    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1282        only in quite long subpatterns.
1283    
1284    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1285        since it was replaced by pcre_fullinfo() in February 2000.
1286    
1287    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1288        match a (*MARK), and the match failed at the start of the subject, a
1289        reference to memory before the start of the subject could occur. This bug
1290        was introduced by fix 17 of release 8.21.
1291    
1292    5.  A reference to an unset group with zero minimum repetition was giving
1293        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1294        /(another)?(\1?)test/ matched against "hello world test". This bug was
1295        introduced in release 8.13.
1296    
1297    6.  Add support for 16-bit character strings (a large amount of work involving
1298        many changes and refactorings).
1299    
1300    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1301        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1302        from a file.
1303    
1304    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1305        rounding is not applied in this particular case).
1306    
1307    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1308        if they appear, or are escaped, in patterns.
1309    
1310    10. Get rid of a number of -Wunused-but-set-variable warnings.
1311    
1312    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1313        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1314        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1315        also returns the mark "x". This bug applied to capturing parentheses,
1316        non-capturing parentheses, and atomic parentheses. It also applied to some
1317        assertions.
1318    
1319    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1320        information out of configure.ac instead of relying on pcre.h.generic, which
1321        is not stored in the repository.
1322    
1323    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1324        -lreadline.
1325    
1326    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1327    
1328    15. Applied Graycode's patch to put the top-level frame on the stack rather
1329        than the heap when not using the stack for recursion. This gives a
1330        performance improvement in many cases when recursion is not deep.
1331    
1332    16. Experimental code added to "pcretest -C" to output the stack frame size.
1333    
1334    
1335    Version 8.21 12-Dec-2011
1336    ------------------------
1337    
1338    1.  Updating the JIT compiler.
1339    
1340    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1341        are added as well.
1342    
1343    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1344        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1345        calling _pcre_jit_exec. Some extra comments are added.
1346    
1347    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1348        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1349        was introduced by change 18 for 8.20.
1350    
1351    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1352        ECMA-262 standard.
1353    
1354    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1355        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1356        This bug was probably introduced by change 9 of 8.13.
1357    
1358    7.  While fixing 6 above, I noticed that a number of other items were being
1359        incorrectly rejected as "not fixed length". This arose partly because newer
1360        opcodes had not been added to the fixed-length checking code. I have (a)
1361        corrected the bug and added tests for these items, and (b) arranged for an
1362        error to occur if an unknown opcode is encountered while checking for fixed
1363        length instead of just assuming "not fixed length". The items that were
1364        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1365        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1366        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1367    
1368    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1369        being incorrectly compiled and would have given unpredicatble results.
1370    
1371    9.  A possessively repeated subpattern with minimum repeat count greater than
1372        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1373        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1374        the first (A) could occur when it should not.
1375    
1376    10. Add a cast and remove a redundant test from the code.
1377    
1378    11. JIT should use pcre_malloc/pcre_free for allocation.
1379    
1380    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1381        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1382        is anything other than /usr, -L is still shown).
1383    
1384    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1385    
1386    14. Perl does not support \N without a following name in a [] class; PCRE now
1387        also gives an error.
1388    
1389    15. If a forward reference was repeated with an upper limit of around 2000,
1390        it caused the error "internal error: overran compiling workspace". The
1391        maximum number of forward references (including repeats) was limited by the
1392        internal workspace, and dependent on the LINK_SIZE. The code has been
1393        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1394        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1395        of around 200,000 forward references. While doing this, I also speeded up
1396        the filling in of repeated forward references.
1397    
1398    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1399        incorrectly expecting the subject to contain another "a" after the start.
1400    
1401    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1402        in the match, the SKIP should be ignored. This was not happening; instead
1403        the SKIP was being treated as NOMATCH. For patterns such as
1404        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1405        tested.
1406    
1407    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1408        now much more compatible with Perl, in particular in cases where the result
1409        is a non-match for a non-anchored pattern. For example, if
1410        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1411        "m", where previously it did not return a name. A side effect of this
1412        change is that for partial matches, the last encountered mark name is
1413        returned, as for non matches. A number of tests that were previously not
1414        Perl-compatible have been moved into the Perl-compatible test files. The
1415        refactoring has had the pleasing side effect of removing one argument from
1416        the match() function, thus reducing its stack requirements.
1417    
1418    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1419        subsequent uses of /S (without +) incorrectly behaved like /S+.
1420    
1421    21. Retrieve executable code size support for the JIT compiler and fixing
1422        some warnings.
1423    
1424    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1425        not work when the shorter character appeared right at the end of the
1426        subject string.
1427    
1428    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1429        systems.
1430    
1431    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1432        output it when the /M option is used in pcretest.
1433    
1434    25. The CheckMan script was not being included in the distribution. Also, added
1435        an explicit "perl" to run Perl scripts from the PrepareRelease script
1436        because this is reportedly needed in Windows.
1437    
1438    26. If study data was being save in a file and studying had not found a set of
1439        "starts with" bytes for the pattern, the data written to the file (though
1440        never used) was taken from uninitialized memory and so caused valgrind to
1441        complain.
1442    
1443    27. Updated RunTest.bat as provided by Sheri Pierce.
1444    
1445    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1446    
1447    29. Computation of memory usage for the table of capturing group names was
1448        giving an unnecessarily large value.
1449    
1450    
1451    Version 8.20 21-Oct-2011
1452    ------------------------
1453    
1454    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1455        a POSIX class. After further experiments with Perl, which convinced me that
1456        Perl has bugs and confusions, a closing square bracket is no longer allowed
1457        in a POSIX name. This bug also affected patterns with classes that started
1458        with full stops.
1459    
1460    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1461        captured substring, but while checking the failing first alternative,
1462        substring 1 is temporarily captured. If the output vector supplied to
1463        pcre_exec() was not big enough for this capture, the yield of the function
1464        was still zero ("insufficient space for captured substrings"). This cannot
1465        be totally fixed without adding another stack variable, which seems a lot
1466        of expense for a edge case. However, I have improved the situation in cases
1467        such as /(a)(b)x|abc/ matched against "abc", where the return code
1468        indicates that fewer than the maximum number of slots in the ovector have
1469        been set.
1470    
1471    3.  Related to (2) above: when there are more back references in a pattern than
1472        slots in the output vector, pcre_exec() uses temporary memory during
1473        matching, and copies in the captures as far as possible afterwards. It was
1474        using the entire output vector, but this conflicts with the specification
1475        that only 2/3 is used for passing back captured substrings. Now it uses
1476        only the first 2/3, for compatibility. This is, of course, another edge
1477        case.
1478    
1479    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1480        main code base, and can be used by building with --enable-jit. When this is
1481        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1482        runtime --no-jit option is given.
1483    
1484    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1485        ovector, the return from the function was zero, implying that there were
1486        other matches that did not fit. The correct "exactly full" value is now
1487        returned.
1488    
1489    6.  If a subpattern that was called recursively or as a subroutine contained
1490        (*PRUNE) or any other control that caused it to give a non-standard return,
1491        invalid errors such as "Error -26 (nested recursion at the same subject
1492        position)" or even infinite loops could occur.
1493    
1494    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1495        computing the minimum length on reaching *ACCEPT, and so ended up with the
1496        wrong value of 1 rather than 0. Further investigation indicates that
1497        computing a minimum subject length in the presence of *ACCEPT is difficult
1498        (think back references, subroutine calls), and so I have changed the code
1499        so that no minimum is registered for a pattern that contains *ACCEPT.
1500    
1501    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1502        it was not handled as intended. [But see 16 below.]
1503    
1504    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1505        Sheri Pierce.
1506    
1507    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1508        the first byte in a match must be "a".
1509    
1510    11. Change 17 for 8.13 increased the recursion depth for patterns like
1511        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1512        pattern contains any instances of (*THEN). If it does not, the old
1513        optimizations are restored. It would be nice to do this on a per-group
1514        basis, but at the moment that is not feasible.
1515    
1516    12. In some environments, the output of pcretest -C is CRLF terminated. This
1517        broke RunTest's code that checks for the link size. A single white space
1518        character after the value is now allowed for.
1519    
1520    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1521        For "fr", it uses the Windows-specific input and output files.
1522    
1523    14. If (*THEN) appeared in a group that was called recursively or as a
1524        subroutine, it did not work as intended. [But see next item.]
1525    
1526    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1527        pattern fragments (but not containing any | characters). If A and B are
1528        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1529        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1530        D. In other words, Perl considers parentheses that do not contain any |
1531        characters to be part of a surrounding alternative, whereas PCRE was
1532        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1533        differently. PCRE now behaves in the same way as Perl, except in the case
1534        of subroutine/recursion calls such as (?1) which have in any case always
1535        been different (but PCRE had them first :-).
1536    
1537    16. Related to 15 above: Perl does not treat the | in a conditional group as
1538        creating alternatives. Such a group is treated in the same way as an
1539        ordinary group without any | characters when processing (*THEN). PCRE has
1540        been changed to match Perl's behaviour.
1541    
1542    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1543        RunGrepTest script failed.
1544    
1545    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1546        inevitable for groups that contain captures, but it can lead to a lot of
1547        stack use in large patterns. The old behaviour has been restored for atomic
1548        groups that do not contain any capturing parentheses.
1549    
1550    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1551        suppress the check for a minimum subject length at run time. (If it was
1552        given to pcre_exec() or pcre_dfa_exec() it did work.)
1553    
1554    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1555        fail to work when decoding hex characters in data strings in EBCDIC
1556        environments.
1557    
1558    21. It appears that in at least one Mac OS environment, the isxdigit() function
1559        is implemented as a macro that evaluates to its argument more than once,
1560        contravening the C 90 Standard (I haven't checked a later standard). There
1561        was an instance in pcretest which caused it to go wrong when processing
1562        \x{...} escapes in subject strings. The has been rewritten to avoid using
1563        things like p++ in the argument of isxdigit().
1564    
1565    
1566    Version 8.13 16-Aug-2011
1567    ------------------------
1568    
1569    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1570    
1571    2.  Two minor typos in pcre_internal.h have been fixed.
1572    
1573    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1574        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1575        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1576    
1577    4.  There were a number of related bugs in the code for matching backrefences
1578        caselessly in UTF-8 mode when codes for the characters concerned were
1579        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1580        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1581        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1582        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1583        2-byte code at the end of the subject (it thought there wasn't enough data
1584        left).
1585    
1586    5.  Comprehensive information about what went wrong is now returned by
1587        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1588        as the output vector has at least 2 elements. The offset of the start of
1589        the failing character and a reason code are placed in the vector.
1590    
1591    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1592        now returned is for the first byte of the failing character, instead of the
1593        last byte inspected. This is an incompatible change, but I hope it is small
1594        enough not to be a problem. It makes the returned offset consistent with
1595        pcre_exec() and pcre_dfa_exec().
1596    
1597    7.  pcretest now gives a text phrase as well as the error number when
1598        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1599        failure, the offset and reason code are output.
1600    
1601    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1602        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1603        back over a single character (\n). This seems wrong (because it treated the
1604        two characters as a single entity when going forwards), conflicts with the
1605        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1606        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1607        has been changed.
1608    
1609    9.  Some internal refactoring has changed the processing so that the handling
1610        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1611        time (the PCRE_DOTALL option was changed this way some time ago: version
1612        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1613        which was always a bit of a fudge. It also means that there is one less
1614        argument for the match() function, which reduces its stack requirements
1615        slightly. This change also fixes an incompatibility with Perl: the pattern
1616        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1617    
1618    10. More internal refactoring has drastically reduced the number of recursive
1619        calls to match() for possessively repeated groups such as (abc)++ when
1620        using pcre_exec().
1621    
1622    11. While implementing 10, a number of bugs in the handling of groups were
1623        discovered and fixed:
1624    
1625        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1626        (a|)*(?1) gave a compile-time internal error.
1627        ((a|)+)+  did not notice that the outer group could match an empty string.
1628        (^a|^)+   was not marked as anchored.
1629        (.*a|.*)+ was not marked as matching at start or after a newline.
1630    
1631    12. Yet more internal refactoring has removed another argument from the match()
1632        function. Special calls to this function are now indicated by setting a
1633        value in a variable in the "match data" data block.
1634    
1635    13. Be more explicit in pcre_study() instead of relying on "default" for
1636        opcodes that mean there is no starting character; this means that when new
1637        ones are added and accidentally left out of pcre_study(), testing should
1638        pick them up.
1639    
1640    14. The -s option of pcretest has been documented for ages as being an old
1641        synonym of -m (show memory usage). I have changed it to mean "force study
1642        for every regex", that is, assume /S for every regex. This is similar to -i
1643        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1644        using it. It makes it easier to run collections of tests with and without
1645        study enabled, and thereby test pcre_study() more easily. All the standard
1646        tests are now run with and without -s (but some patterns can be marked as
1647        "never study" - see 20 below).
1648    
1649    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1650        restoration of the capturing data to the outer values was not happening
1651        correctly.
1652    
1653    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1654        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1655        pattern had matched an empty string, and so incorrectly returned a no
1656        match.
1657    
1658    17. There was optimizing code for the last branch of non-capturing parentheses,
1659        and also for the obeyed branch of a conditional subexpression, which used
1660        tail recursion to cut down on stack usage. Unfortunately, now that there is
1661        the possibility of (*THEN) occurring in these branches, tail recursion is
1662        no longer possible because the return has to be checked for (*THEN). These
1663        two optimizations have therefore been removed. [But see 8.20/11 above.]
1664    
1665    18. If a pattern containing \R was studied, it was assumed that \R always
1666        matched two bytes, thus causing the minimum subject length to be
1667        incorrectly computed because \R can also match just one byte.
1668    
1669    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1670        was incorrectly computed.
1671    
1672    20. If /S is present twice on a test pattern in pcretest input, it now
1673        *disables* studying, thereby overriding the use of -s on the command line
1674        (see 14 above). This is necessary for one or two tests to keep the output
1675        identical in both cases.
1676    
1677    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1678        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1679    
1680    22. When an atomic group that contained a capturing parenthesis was
1681        successfully matched, but the branch in which it appeared failed, the
1682        capturing was not being forgotten if a higher numbered group was later
1683        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1684        group 1 to "a", when in fact it should be unset. This applied to multi-
1685        branched capturing and non-capturing groups, repeated or not, and also to
1686        positive assertions (capturing in negative assertions does not happen
1687        in PCRE) and also to nested atomic groups.
1688    
1689    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1690        subject after a captured substring, to make it easier to tell which of a
1691        number of identical substrings has been captured.
1692    
1693    24. The way atomic groups are processed by pcre_exec() has been changed so that
1694        if they are repeated, backtracking one repetition now resets captured
1695        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1696        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1697        "aaa". Previously, it would have been "a". As part of this code
1698        refactoring, the way recursive calls are handled has also been changed.
1699    
1700    25. If an assertion condition captured any substrings, they were not passed
1701        back unless some other capturing happened later. For example, if
1702        (?(?=(a))a) was matched against "a", no capturing was returned.
1703    
1704    26. When studying a pattern that contained subroutine calls or assertions,
1705        the code for finding the minimum length of a possible match was handling
1706        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1707        group 1 called group 2 while simultaneously a separate group 2 called group
1708        1). A stack overflow occurred in this case. I have fixed this by limiting
1709        the recursion depth to 10.
1710    
1711    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1712        Fortmann. This supports explicit test numbers on the command line, and has
1713        argument validation and error reporting.
1714    
1715    28. An instance of \X with an unlimited repeat could fail if at any point the
1716        first character it looked at was a mark character.
1717    
1718    29. Some minor code refactoring concerning Unicode properties and scripts
1719        should reduce the stack requirement of match() slightly.
1720    
1721    30. Added the '=' option to pcretest to check the setting of unused capturing
1722        slots at the end of the pattern, which are documented as being -1, but are
1723        not included in the return count.
1724    
1725    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1726        compiled something random. Now it gives a compile-time error (as does
1727        Perl).
1728    
1729    32. A *MARK encountered during the processing of a positive assertion is now
1730        recorded and passed back (compatible with Perl).
1731    
1732    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1733        had alternative anchored branches, the search for a second match in a line
1734        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1735        matched the line "0102" twice. The same bug affected patterns that started
1736        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1737        twice.
1738    
1739    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1740        does, and because of capturing effects, quantifying parenthesized
1741        assertions may at times be useful. Quantifiers are now allowed for
1742        parenthesized assertions.
1743    
1744    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1745    
1746    36. \g was being checked for fancy things in a character class, when it should
1747        just be a literal "g".
1748    
1749    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1750        appearance of a nested POSIX class supersedes an apparent external class.
1751        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1752        unescaped square brackets may also appear as part of class names. For
1753        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1754        more like Perl. (But see 8.20/1 above.)
1755    
1756    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1757        was because it thought it was \N{name}, which is not supported).
1758    
1759    39. Add minix to OS list not supporting the -S option in pcretest.
1760    
1761    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1762        cannot analyze patterns in sufficient detail to catch mutual recursions
1763        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1764        subgroup is called recursively as a subpattern for a second time at the
1765        same position in the subject string. In previous releases this might have
1766        been caught by the recursion limit, or it might have run out of stack.
1767    
1768    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1769        happen only once. PCRE was, however incorrectly giving a compile time error
1770        "recursive call could loop indefinitely" because it cannot analyze the
1771        pattern in sufficient detail. The compile time test no longer happens when
1772        PCRE is compiling a conditional subpattern, but actual runaway loops are
1773        now caught at runtime (see 40 above).
1774    
1775    42. It seems that Perl allows any characters other than a closing parenthesis
1776        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1777        has been changed to be the same.
1778    
1779    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1780        as not to get warnings when autogen.sh is called. Also changed
1781        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1782    
1783    44. To help people who use pcregrep to scan files containing exceedingly long
1784        lines, the following changes have been made:
1785    
1786        (a) The default value of the buffer size parameter has been increased from
1787            8K to 20K. (The actual buffer used is three times this size.)
1788    
1789        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1790            PCRE is built.
1791    
1792        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1793            to be set at run time.
1794    
1795        (d) Numerical values in pcregrep options can be followed by K or M, for
1796            example --buffer-size=50K.
1797    
1798        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1799            given and the return code is set to 2.
1800    
1801    45. Add a pointer to the latest mark to the callout data block.
1802    
1803    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1804        partial match of an empty string instead of no match. This was specific to
1805        the use of ".".
1806    
1807    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1808        complete match instead of a partial match. This bug was dependent on both
1809        the PCRE_UTF8 and PCRE_DOTALL options being set.
1810    
1811    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1812        starting byte set, because \b was not being ignored.
1813    
1814    
1815    Version 8.12 15-Jan-2011
1816    ------------------------
1817    
1818    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1819        checks for such things as part of the documentation building process.
1820    
1821    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1822        --match-limit and --recursion-limit options (added for 8.11). In
1823        particular, this made one of the standard tests fail. (The integer value
1824        went into the wrong half of a long int.)
1825    
1826    3.  If the --colour option was given to pcregrep with -v (invert match), it
1827        did strange things, either producing crazy output, or crashing. It should,
1828        of course, ignore a request for colour when reporting lines that do not
1829        match.
1830    
1831    4.  Another pcregrep bug caused similar problems if --colour was specified with
1832        -M (multiline) and the pattern match finished with a line ending.
1833    
1834    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1835        matched in multiline mode, the following line was shown as part of the
1836        match. This seems wrong, so I have changed it.
1837    
1838    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1839        the check for further matches in the same line (so they could be coloured)
1840        to overrun the end of the current line. If another match was found, it was
1841        incorrectly shown (and then shown again when found in the next line).
1842    
1843    7.  If pcregrep was compiled under Windows, there was a reference to the
1844        function pcregrep_exit() before it was defined. I am assuming this was
1845        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1846        reported by a user. I've moved the definition above the reference.
1847    
1848    
1849    Version 8.11 10-Dec-2010
1850    ------------------------
1851    
1852    1.  (*THEN) was not working properly if there were untried alternatives prior
1853        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1854        backtracked to try for "b" instead of moving to the next alternative branch
1855        at the same level (in this case, to look for "c"). The Perl documentation
1856        is clear that when (*THEN) is backtracked onto, it goes to the "next
1857        alternative in the innermost enclosing group".
1858    
1859    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1860        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1861        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1862        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1863        (*THEN).
1864    
1865    3.  If \s appeared in a character class, it removed the VT character from
1866        the class, even if it had been included by some previous item, for example
1867        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1868        of \s, but is part of the POSIX "space" class.)
1869    
1870    4.  A partial match never returns an empty string (because you can always
1871        match an empty string at the end of the subject); however the checking for
1872        an empty string was starting at the "start of match" point. This has been
1873        changed to the "earliest inspected character" point, because the returned
1874        data for a partial match starts at this character. This means that, for
1875        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1876        (previously it gave "no match").
1877    
1878    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1879        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1880        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1881        has an implication that the given string is incomplete (because a partial
1882        match is preferred over a full match). For this reason, these items now
1883        give a partial match in this situation. [Aside: previously, the one case
1884        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1885        match rather than a full match, which was wrong by the old rules, but is
1886        now correct.]
1887    
1888    6.  There was a bug in the handling of #-introduced comments, recognized when
1889        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1890        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1891        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1892        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1893        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1894        places in pcre_compile().
1895    
1896    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1897        comments when looking ahead for named forward references to subpatterns,
1898        the only newline sequence it recognized was NL. It now handles newlines
1899        according to the set newline convention.
1900    
1901    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1902        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1903        cater for a lack of strerror(). These oversights have been fixed.
1904    
1905    9.  Added --match-limit and --recursion-limit to pcregrep.
1906    
1907    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1908    
1909    11. When the -o option was used, pcregrep was setting a return code of 1, even
1910        when matches were found, and --line-buffered was not being honoured.
1911    
1912    12. Added an optional parentheses number to the -o and --only-matching options
1913        of pcregrep.
1914    
1915    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1916        can match an empty string. The code to do it in pcretest and pcredemo
1917        needed fixing:
1918    
1919        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1920            only one byte after an empty string match just before CRLF (this case
1921            just got forgotten; "any" and "anycrlf" were OK).
1922    
1923        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1924            mode when an empty string match preceded an ASCII character followed by
1925            a non-ASCII character. (The code for advancing by one character rather
1926            than one byte was nonsense.)
1927    
1928        (c) The pcredemo.c sample program did not have any code at all to handle
1929            the cases when CRLF is a valid newline sequence.
1930    
1931    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1932        as a starting offset was within the subject string. There is now a new
1933        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1934        negative or greater than the length of the string. In order to test this,
1935        pcretest is extended to allow the setting of negative starting offsets.
1936    
1937    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1938        starting offset points to the beginning of a UTF-8 character was
1939        unnecessarily clumsy. I tidied it up.
1940    
1941    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1942        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1943    
1944    17. Nobody had reported that the --include_dir option, which was added in
1945        release 7.7 should have been called --include-dir (hyphen, not underscore)
1946        for compatibility with GNU grep. I have changed it to --include-dir, but
1947        left --include_dir as an undocumented synonym, and the same for
1948        --exclude-dir, though that is not available in GNU grep, at least as of
1949        release 2.5.4.
1950    
1951    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1952        characters from a string of bytes) have been redefined so as not to use
1953        loops, in order to improve performance in some environments. At the same
1954        time, I abstracted some of the common code into auxiliary macros to save
1955        repetition (this should not affect the compiled code).
1956    
1957    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1958        compile-time error is now given if \c is not followed by an ASCII
1959        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1960        different, and any byte value is allowed.)
1961    
1962    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1963        START_OPTIMIZE option, which is now allowed at compile time - but just
1964        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1965        to pcregrep and other applications that have no direct access to PCRE
1966        options. The new /Y option in pcretest sets this option when calling
1967        pcre_compile().
1968    
1969    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1970        back references. Groups containing recursive back references were forced to
1971        be atomic by that change, but in the case of named groups, the amount of
1972        memory required was incorrectly computed, leading to "Failed: internal
1973        error: code overflow". This has been fixed.
1974    
1975    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1976        pcretest.c, to avoid build problems in some Borland environments.
1977    
1978    
1979    Version 8.10 25-Jun-2010
1980    ------------------------
1981    
1982    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1983        THEN.
1984    
1985    2.  (*ACCEPT) was not working when inside an atomic group.
1986    
1987    3.  Inside a character class, \B is treated as a literal by default, but
1988        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1989        causes the error). The code is unchanged, but I tidied the documentation.
1990    
1991    4.  Inside a character class, PCRE always treated \R and \X as literals,
1992        whereas Perl faults them if its -w option is set. I have changed PCRE so
1993        that it faults them when PCRE_EXTRA is set.
1994    
1995    5.  Added support for \N, which always matches any character other than
1996        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1997    
1998    6.  When compiling pcregrep with newer versions of gcc which may have
1999        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2000        declared with attribute warn_unused_result" were given. Just casting the
2001        result to (void) does not stop the warnings; a more elaborate fudge is
2002        needed. I've used a macro to implement this.
2003    
2004    7.  Minor change to pcretest.c to avoid a compiler warning.
2005    
2006    8.  Added four artifical Unicode properties to help with an option to make
2007        \s etc use properties (see next item). The new properties are: Xan
2008        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2009    
2010    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2011        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2012        this option. Modified pcretest to add /W to test this facility. Added
2013        REG_UCP to make it available via the POSIX interface.
2014    
2015    10. Added --line-buffered to pcregrep.
2016    
2017    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2018        studied, and the match started with a letter with a code point greater than
2019        127 whose first byte was different to the first byte of the other case of
2020        the letter, the other case of this starting letter was not recognized
2021        (#976).
2022    
2023    12. If a pattern that was studied started with a repeated Unicode property
2024        test, for example, \p{Nd}+, there was the theoretical possibility of
2025        setting up an incorrect bitmap of starting bytes, but fortunately it could
2026        not have actually happened in practice until change 8 above was made (it
2027        added property types that matched character-matching opcodes).
2028    
2029    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2030        possible starting bytes for non-anchored patterns.
2031    
2032    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2033        \R, and also a number of cases that involve Unicode properties, both
2034        explicit and implicit when PCRE_UCP is set.
2035    
2036    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2037        input, it could crash or give wrong results if characters with values
2038        greater than 0xc0 were present in the subject string. (Detail: it assumed
2039        UTF-8 input when processing these items.)
2040    
2041    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2042        size_t is 64-bit (#991).
2043    
2044    17. Added a check for running out of memory when PCRE is compiled with
2045        --disable-stack-for-recursion (#990).
2046    
2047    18. If the last data line in a file for pcretest does not have a newline on
2048        the end, a newline was missing in the output.
2049    
2050    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2051        less than 128) in its various bitmaps. However, there is a facility for
2052        generating tables according to the current locale when PCRE is compiled. It
2053        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2054        characters, are recognized by isspace() and therefore were getting set in
2055        these tables, and indeed these tables seem to approximate to ISO 8859. This
2056        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2057        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2058        which of course cannot start UTF-8 characters. I have changed the code so
2059        that only real ASCII characters (less than 128) and the correct starting
2060        bytes for UTF-8 encodings are set for characters greater than 127 when in
2061        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2062        altogether.)
2063    
2064    20. Added the /T option to pcretest so as to be able to run tests with non-
2065        standard character tables, thus making it possible to include the tests
2066        used for 19 above in the standard set of tests.
2067    
2068    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2069        reference to a subpattern the other side of a comment that contains an
2070        opening parenthesis caused either an internal compiling error, or a
2071        reference to the wrong subpattern.
2072    
2073    
2074    Version 8.02 19-Mar-2010
2075    ------------------------
2076    
2077    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2078    
2079    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2080        configured.
2081    
2082    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2083        original author of that file, following a query about its status.
2084    
2085    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2086        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2087    
2088    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2089        quantifier applied to a forward-referencing subroutine call, could compile
2090        incorrect code or give the error "internal error: previously-checked
2091        referenced subpattern not found".
2092    
2093    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2094        variables to point to external functions. For these systems, therefore,
2095        pcre_malloc etc. are now initialized to local functions that call the
2096        relevant global functions.
2097    
2098    7.  There were two entries missing in the vectors called coptable and poptable
2099        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2100        I've fixed the data, and added a kludgy way of testing at compile time that
2101        the lengths are correct (equal to the number of opcodes).
2102    
2103    8.  Following on from 7, I added a similar kludge to check the length of the
2104        eint vector in pcreposix.c.
2105    
2106    9.  Error texts for pcre_compile() are held as one long string to avoid too
2107        much relocation at load time. To find a text, the string is searched,
2108        counting zeros. There was no check for running off the end of the string,
2109        which could happen if a new error number was added without updating the
2110        string.
2111    
2112    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2113    
2114    11. \K was not working if it appeared in an atomic group or in a group that
2115        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2116        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2117        the assertion is positive, but not if it is negative.
2118    
2119    12. Change 11 fortuitously reduced the size of the stack frame used in the
2120        "match()" function of pcre_exec.c by one pointer. Forthcoming
2121        implementation of support for (*MARK) will need an extra pointer on the
2122        stack; I have reserved it now, so that the stack frame size does not
2123        decrease.
2124    
2125    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2126        item in branch that calls a recursion is a subroutine call - as in the
2127        second branch in the above example - was incorrectly given the compile-
2128        time error "recursive call could loop indefinitely" because pcre_compile()
2129        was not correctly checking the subroutine for matching a non-empty string.
2130    
2131    14. The checks for overrunning compiling workspace could trigger after an
2132        overrun had occurred. This is a "should never occur" error, but it can be
2133        triggered by pathological patterns such as hundreds of nested parentheses.
2134        The checks now trigger 100 bytes before the end of the workspace.
2135    
2136    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2137    
2138    
2139    Version 8.01 19-Jan-2010
2140    ------------------------
2141    
2142    1.  If a pattern contained a conditional subpattern with only one branch (in
2143        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2144        computed the wrong minimum data length (which is of course zero for such
2145        subpatterns). This could cause incorrect "no match" results.
2146    
2147    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2148        the pattern is reset in the first branch, pcre_compile() failed with
2149        "internal error: code overflow at offset...". This happened only when
2150        the reset was to the original external option setting. (An optimization
2151        abstracts leading options settings into an external setting, which was the
2152        cause of this.)
2153    
2154    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2155        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2156        assertion pattern did not match (meaning that the assertion was true), it
2157        was incorrectly treated as false if the SKIP had been reached during the
2158        matching. This also applied to assertions used as conditions.
2159    
2160    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2161        assertion subpattern, including such a pattern used as a condition,
2162        unpredictable results occurred, instead of the error return
2163        PCRE_ERROR_DFA_UITEM.
2164    
2165    5.  The C++ GlobalReplace function was not working like Perl for the special
2166        situation when an empty string is matched. It now does the fancy magic
2167        stuff that is necessary.
2168    
2169    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2170        removed. (These were left over from very, very early versions of PCRE.)
2171    
2172    7.  Some cosmetic changes to the code to make life easier when compiling it
2173        as part of something else:
2174    
2175        (a) Change DEBUG to PCRE_DEBUG.
2176    
2177        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2178            called "current" as "current_branch", to prevent a collision with the
2179            Linux macro when compiled as a kernel module.
2180    
2181        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2182            prevent a collision with the Linux macro when compiled as a kernel
2183            module.
2184    
2185    8.  In pcre_compile() there are some checks for integer overflows that used to
2186        cast potentially large values to (double). This has been changed to that
2187        when building, a check for int64_t is made, and if it is found, it is used
2188        instead, thus avoiding the use of floating point arithmetic. (There is no
2189        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2190        double.
2191    
2192    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2193        2005 (difference between two addresses compared to an unsigned value).
2194    
2195    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2196        custom one, because of the following reported problem in Windows:
2197    
2198          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2199              under Win32.
2200          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2201              therefore missing the function definition.
2202          - The compiler thus generates a "C" signature for the test function.
2203          - The linker fails to find the "C" function.
2204          - PCRE fails to configure if asked to do so against libbz2.
2205    
2206    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2207        messages were output:
2208    
2209          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2210          rerunning libtoolize, to keep the correct libtool macros in-tree.
2211          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2212    
2213        I have done both of these things.
2214    
2215    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2216        most of the time, it *can* run out if it is given a pattern that contains a
2217        runaway infinite recursion. I updated the discussion in the pcrestack man
2218        page.
2219    
2220    13. Now that we have gone to the x.xx style of version numbers, the minor
2221        version may start with zero. Using 08 or 09 is a bad idea because users
2222        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2223        interpreted as invalid octal numbers. I've updated the previous comment in
2224        configure.ac, and also added a check that gives an error if 08 or 09 are
2225        used.
2226    
2227    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2228        causing partial matching to fail when the end of the subject matched \W
2229        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2230    
2231    15. There were some discrepancies between the declarations in pcre_internal.h
2232        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2233        their definitions. The declarations used "const uschar *" and the
2234        definitions used USPTR. Even though USPTR is normally defined as "const
2235        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2236        reported that: "This difference in casting confuses some C++ compilers, for
2237        example, SunCC recognizes above declarations as different functions and
2238        generates broken code for hbpcre." I have changed the declarations to use
2239        USPTR.
2240    
2241    16. GNU libtool is named differently on some systems. The autogen.sh script now
2242        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2243        (FreeBSD).
2244    
2245    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2246        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2247        comment: "Figure out how to create a longlong from a string: strtoll and
2248        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2249        instance, but it only takes 2 args instead of 3!"
2250    
2251    18. A subtle bug concerned with back references has been fixed by a change of
2252        specification, with a corresponding code fix. A pattern such as
2253        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2254        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2255        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2256        same bug. Such groups have to be quantified to be useful, or contained
2257        inside another quantified group. (If there's no repetition, the reference
2258        can never match.) The problem arises because, having left the group and
2259        moved on to the rest of the pattern, a later failure that backtracks into
2260        the group uses the captured value from the final iteration of the group
2261        rather than the correct earlier one. I have fixed this in PCRE by forcing
2262        any group that contains a reference to itself to be an atomic group; that
2263        is, there cannot be any backtracking into it once it has completed. This is
2264        similar to recursive and subroutine calls.
2265    
2266    
2267    Version 8.00 19-Oct-09
2268    ----------------------
2269    
2270    1.  The table for translating pcre_compile() error codes into POSIX error codes
2271        was out-of-date, and there was no check on the pcre_compile() error code
2272        being within the table. This could lead to an OK return being given in
2273        error.
2274    
2275    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2276        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2277        in a Windows environment.
2278    
2279    3.  The pcregrep --count option prints the count for each file even when it is
2280        zero, as does GNU grep. However, pcregrep was also printing all files when
2281        --files-with-matches was added. Now, when both options are given, it prints
2282        counts only for those files that have at least one match. (GNU grep just
2283        prints the file name in this circumstance, but including the count seems
2284        more useful - otherwise, why use --count?) Also ensured that the
2285        combination -clh just lists non-zero counts, with no names.
2286    
2287    4.  The long form of the pcregrep -F option was incorrectly implemented as
2288        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2289        but it seems right to fix it, and I didn't think it was worth preserving
2290        the old behaviour.
2291    
2292    5.  The command line items --regex=pattern and --regexp=pattern were not
2293        recognized by pcregrep, which required --regex pattern or --regexp pattern
2294        (with a space rather than an '='). The man page documented the '=' forms,
2295        which are compatible with GNU grep; these now work.
2296    
2297    6.  No libpcreposix.pc file was created for pkg-config; there was just
2298        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2299    
2300    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2301        when UCP support is not needed, by modifying the Python script that
2302        generates it from Unicode data files. This should not matter if the module
2303        is correctly used as a library, but I received one complaint about 50K of
2304        unwanted data. My guess is that the person linked everything into his
2305        program rather than using a library. Anyway, it does no harm.
2306    
2307    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2308        was a minimum greater than 1 for a wide character in a possessive
2309        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2310        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2311        character. Chaos in the form of incorrect output or a compiling loop could
2312        result.
2313    
2314    9.  The restrictions on what a pattern can contain when partial matching is
2315        requested for pcre_exec() have been removed. All patterns can now be
2316        partially matched by this function. In addition, if there are at least two
2317        slots in the offset vector, the offset of the earliest inspected character
2318        for the match and the offset of the end of the subject are set in them when
2319        PCRE_ERROR_PARTIAL is returned.
2320    
2321    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2322        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2323        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2324        and may be more useful for multi-segment matching.
2325    
2326    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2327        used to be given if ever the end of the subject was reached; now it is
2328        given only if matching could not proceed because another character was
2329        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2330        string "Z", which now yields "no match" instead of "partial match". In the
2331        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2332        final character ended with (*FAIL).
2333    
2334    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2335        if the pattern had a "must contain" character that was already found in the
2336        earlier partial match, unless partial matching was again requested. For
2337        example, with the pattern /dog.(body)?/, the "must contain" character is
2338        "g". If the first part-match was for the string "dog", restarting with
2339        "sbody" failed. This bug has been fixed.
2340    
2341    13. The string returned by pcre_dfa_exec() after a partial match has been
2342        changed so that it starts at the first inspected character rather than the
2343        first character of the match. This makes a difference only if the pattern
2344        starts with a lookbehind assertion or \b or \B (\K is not supported by
2345        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2346        matching functions compatible, and I think it's the right thing to do.
2347    
2348    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2349        so that the demonstration program is easily available in environments where
2350        PCRE has not been installed from source.
2351    
2352    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2353        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2354        library.
2355    
2356    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2357        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2358        is not the first non-POSIX option to be added. Clearly some people find
2359        these options useful.
2360    
2361    17. If a caller to the POSIX matching function regexec() passes a non-zero
2362        value for nmatch with a NULL value for pmatch, the value of
2363        nmatch is forced to zero.
2364    
2365    18. RunGrepTest did not have a test for the availability of the -u option of
2366        the diff command, as RunTest does. It now checks in the same way as
2367        RunTest, and also checks for the -b option.
2368    
2369    19. If an odd number of negated classes containing just a single character
2370        interposed, within parentheses, between a forward reference to a named
2371        subpattern and the definition of the subpattern, compilation crashed with
2372        an internal error, complaining that it could not find the referenced
2373        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2374        [The bug was that it was starting one character too far in when skipping
2375        over the character class, thus treating the ] as data rather than
2376        terminating the class. This meant it could skip too much.]
2377    
2378    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2379        /g option in pcretest when the pattern contains \K, which makes it possible
2380        to have an empty string match not at the start, even when the pattern is
2381        anchored. Updated pcretest and pcredemo to use this option.
2382    
2383    21. If the maximum number of capturing subpatterns in a recursion was greater
2384        than the maximum at the outer level, the higher number was returned, but
2385        with unset values at the outer level. The correct (outer level) value is
2386        now given.
2387    
2388    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2389        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2390        make it do so. The string so far is captured, making this feature
2391        compatible with Perl.
2392    
2393    23. The tests have been re-organized, adding tests 11 and 12, to make it
2394        possible to check the Perl 5.10 features against Perl 5.10.
2395    
2396    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2397        pattern matches a fixed length string. PCRE did not allow this; now it
2398        does. Neither allows recursion.
2399    
2400    25. I finally figured out how to implement a request to provide the minimum
2401        length of subject string that was needed in order to match a given pattern.
2402        (It was back references and recursion that I had previously got hung up
2403        on.) This code has now been added to pcre_study(); it finds a lower bound
2404        to the length of subject needed. It is not necessarily the greatest lower
2405        bound, but using it to avoid searching strings that are too short does give
2406        some useful speed-ups. The value is available to calling programs via
2407        pcre_fullinfo().
2408    
2409    26. While implementing 25, I discovered to my embarrassment that pcretest had
2410        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2411        study optimizations had never been tested with that matching function.
2412        Oops. What is worse, even when it was passed study data, there was a bug in
2413        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2414        were also very few tests of studied patterns with pcre_dfa_exec().
2415    
2416    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2417        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2418        on the other side of the coin, they are no longer allowed to have different
2419        names, because these cannot be distinguished in PCRE, and this has caused
2420        confusion. (This is a difference from Perl.)
2421    
2422    28. When duplicate subpattern names are present (necessarily with different
2423        numbers, as required by 27 above), and a test is made by name in a
2424        conditional pattern, either for a subpattern having been matched, or for
2425        recursion in such a pattern, all the associated numbered subpatterns are
2426        tested, and the overall condition is true if the condition is true for any
2427        one of them. This is the way Perl works, and is also more like the way
2428        testing by number works.
2429    
2430    
2431    Version 7.9 11-Apr-09
2432    ---------------------
2433    
2434    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2435        (pcretest), all targets were linked against these libraries. This included
2436        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2437        libraries. This caused unwanted dependencies to be created. This problem
2438        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2439        pcretest is linked with readline.
2440    
2441    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2442        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2443        moved outside it again, because FALSE and TRUE are already defined in AIX,
2444        but BOOL is not.
2445    
2446    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2447        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2448    
2449    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2450        hyphens) following file names and line numbers when outputting matching
2451        lines. This is not true; no spaces are inserted. I have also clarified the
2452        wording for the --colour (or --color) option.
2453    
2454    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2455        was not coloured; this is different to GNU grep, so I have changed it to be
2456        the same.
2457    
2458    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2459        each matching line was coloured. Now it goes on to look for further matches
2460        of any of the test patterns, which is the same behaviour as GNU grep.
2461    
2462    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2463        doesn't make sense to accept an empty string match in pcregrep, so I have
2464        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2465        seems to be how GNU grep behaves. [But see later change 40 for release
2466        8.33.]
2467    
2468    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2469        start or after a newline", because the conditional assertion was not being
2470        correctly handled. The rule now is that both the assertion and what follows
2471        in the first alternative must satisfy the test.
2472    
2473    9.  If auto-callout was enabled in a pattern with a conditional group whose
2474        condition was an assertion, PCRE could crash during matching, both with
2475        pcre_exec() and pcre_dfa_exec().
2476    
2477    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2478        used for matching.
2479    
2480    11. Unicode property support in character classes was not working for
2481        characters (bytes) greater than 127 when not in UTF-8 mode.
2482    
2483    12. Added the -M command line option to pcretest.
2484    
2485    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2486    
2487    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2488    
2489    16. Added comments and documentation about mis-use of no_arg in the C++
2490        wrapper.
2491    
2492    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2493        from Martin Jerabek that uses macro names for all relevant character and
2494        string constants.
2495    
2496    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2497        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2498        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2499        these, but not everybody uses configure.
2500    
2501    19. A conditional group that had only one branch was not being correctly
2502        recognized as an item that could match an empty string. This meant that an
2503        enclosing group might also not be so recognized, causing infinite looping
2504        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2505        with the subject "ab", where knowledge that the repeated group can match
2506        nothing is needed in order to break the loop.
2507    
2508    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2509        exec(), but without supplying a callout function, matching went wrong.
2510    
2511    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2512        leak if the size of the offset vector was greater than 30. When the vector
2513        is smaller, the saved offsets during recursion go onto a local stack
2514        vector, but for larger vectors malloc() is used. It was failing to free
2515        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2516        error, in fact).
2517    
2518    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2519        heapframe that is used only when UTF-8 support is enabled. This caused no
2520        problem, but was untidy.
2521    
2522    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2523        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2524        included within another project.
2525    
2526    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2527        slightly modified by me:
2528    
2529          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2530              not building pcregrep.
2531    
2532          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2533              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2534    
2535    25. Forward references, both numeric and by name, in patterns that made use of
2536        duplicate group numbers, could behave incorrectly or give incorrect errors,
2537        because when scanning forward to find the reference group, PCRE was not
2538        taking into account the duplicate group numbers. A pattern such as
2539        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2540    
2541    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2542        the feature of a custom pointer more persuasive (as requested by a user).
2543    
2544    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2545        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2546        pre-defined.
2547    
2548    28. Added support for (*UTF8) at the start of a pattern.
2549    
2550    29. Arrange for flags added by the "release type" setting in CMake to be shown
2551        in the configuration summary.
2552    
2553    
2554    Version 7.8 05-Sep-08
2555    ---------------------
2556    
2557    1.  Replaced UCP searching code with optimized version as implemented for Ad
2558        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2559        stage table and inline lookup instead of a function, giving speed ups of 2
2560        to 5 times on some simple patterns that I tested. Permission was given to
2561        distribute the MultiStage2.py script that generates the tables (it's not in
2562        the tarball, but is in the Subversion repository).
2563    
2564    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2565        scripts.
2566    
2567    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2568        a group with a zero qualifier. The result of the study could be incorrect,
2569        or the function might crash, depending on the pattern.
2570    
2571    4.  Caseless matching was not working for non-ASCII characters in back
2572        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2573        It now works when Unicode Property Support is available.
2574    
2575    5.  In pcretest, an escape such as \x{de} in the data was always generating
2576        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2577        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2578        truncation.
2579    
2580    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2581    
2582    7.  Added two (int) casts to pcregrep when printing the difference of two
2583        pointers, in case they are 64-bit values.
2584    
2585    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2586        test 2 if it fails.
2587    
2588    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2589        and a #define of that name to empty if it is not externally set. This is to
2590        allow users of MSVC to set it if necessary.
2591    
2592    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2593        the convenience functions in the pcre_get.c source file.
2594    
2595    11. An option change at the start of a pattern that had top-level alternatives
2596        could cause overwriting and/or a crash. This command provoked a crash in
2597        some environments:
2598    
2599          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2600    
2601        This potential security problem was recorded as CVE-2008-2371.
2602    
2603    12. For a pattern where the match had to start at the beginning or immediately
2604        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2605        pcre_dfa_exec() could read past the end of the passed subject if there was
2606        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2607        pcretest so that it places the subject at the end of its malloc-ed buffer.
2608    
2609    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2610        exec() might read past the end of the data buffer in UTF-8 mode.
2611    
2612    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2613        the data contained the byte 0x85 as part of a UTF-8 character within its
2614        first line. This applied both to normal and DFA matching.
2615    
2616    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2617        /^[^d]*?$/8 failed to match "abc".
2618    
2619    16. Added a missing copyright notice to pcrecpp_internal.h.
2620    
2621    17. Make it more clear in the documentation that values returned from
2622        pcre_exec() in ovector are byte offsets, not character counts.
2623    
2624    18. Tidied a few places to stop certain compilers from issuing warnings.
2625    
2626    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2627        supplied by Stefan Weber. I made a further small update for 7.8 because
2628        there is a change of source arrangements: the pcre_searchfuncs.c module is
2629        replaced by pcre_ucd.c.
2630    
2631    
2632    Version 7.7 07-May-08
2633    ---------------------
2634    
2635    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2636        a string to a long long, pretend we don't even have a long long." This is
2637        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2638    
2639    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2640        pre-7.6 versions, which defined a global no_arg variable instead of putting
2641        it in the RE class. (See also #8 below.)
2642    
2643    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2644        Lopes.
2645    
2646    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2647    
2648        (1) The include/exclude patterns were being applied to the whole pathnames
2649            of files, instead of just to the final components.
2650    
2651        (2) If there was more than one level of directory, the subdirectories were
2652            skipped unless they satisfied the include/exclude conditions. This is
2653            inconsistent with GNU grep (and could even be seen as contrary to the
2654            pcregrep specification - which I improved to make it absolutely clear).
2655            The action now is always to scan all levels of directory, and just
2656            apply the include/exclude patterns to regular files.
2657    
2658    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2659        --exclude_dir in the tests to avoid scanning .svn directories.
2660    
2661    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2662        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2663        doesn't support NULs in patterns.
2664    
2665    7.  Added some missing "const"s to declarations of static tables in
2666        pcre_compile.c and pcre_dfa_exec.c.
2667    
2668    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2669        caused by fix #2  above. (Subsequently also a second patch to fix the
2670        first patch. And a third patch - this was a messy problem.)
2671    
2672    9.  Applied Craig's patch to remove the use of push_back().
2673    
2674    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2675        matching function regexec().
2676    
2677    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2678        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2679        references. PCRE supports relative numbers with this syntax (I don't think
2680        Oniguruma does).
2681    
2682    12. Previously, a group with a zero repeat such as (...){0} was completely
2683        omitted from the compiled regex. However, this means that if the group
2684        was called as a subroutine from elsewhere in the pattern, things went wrong
2685        (an internal error was given). Such groups are now left in the compiled
2686        pattern, with a new opcode that causes them to be skipped at execution
2687        time.
2688    
2689    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2690        to the way PCRE behaves:
2691    
2692        (a) A lone ] character is dis-allowed (Perl treats it as data).
2693    
2694        (b) A back reference to an unmatched subpattern matches an empty string
2695            (Perl fails the current match path).
2696    
2697        (c) A data ] in a character class must be notated as \] because if the
2698            first data character in a class is ], it defines an empty class. (In
2699            Perl it is not possible to have an empty class.) The empty class []
2700            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2701            The negative empty class [^] matches any one character, independently
2702            of the DOTALL setting.
2703    
2704    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2705        non-existent subpattern following a character class starting with ']' and
2706        containing () gave an internal compiling error instead of "reference to
2707        non-existent subpattern". Fortunately, when the pattern did exist, the
2708        compiled code was correct. (When scanning forwards to check for the
2709        existence of the subpattern, it was treating the data ']' as terminating
2710        the class, so got the count wrong. When actually compiling, the reference
2711        was subsequently set up correctly.)
2712    
2713    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2714        it was being rejected as not supported by pcre_dfa_exec(), even though
2715        other assertions are supported. I have made pcre_dfa_exec() support
2716        (*FAIL).
2717    
2718    16. The implementation of 13c above involved the invention of a new opcode,
2719        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2720        cannot be changed at match time, I realized I could make a small
2721        improvement to matching performance by compiling OP_ALLANY instead of
2722        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2723        on the OP_ANY path.
2724    
2725    17. Compiling pcretest on Windows with readline support failed without the
2726        following two fixes: (1) Make the unistd.h include conditional on
2727        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2728    
2729    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2730        ncurses library to be included for pcretest when ReadLine support is
2731        requested, but also to allow for it to be overridden. This patch came from
2732        Daniel Bergström.
2733    
2734    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2735        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2736        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2737        spotting this.
2738    
2739    
2740    Version 7.6 28-Jan-08
2741    ---------------------
2742    
2743    1.  A character class containing a very large number of characters with
2744        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2745        overflow.
2746    
2747    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2748        HAVE_LONG_LONG is not defined.
2749    
2750    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2751        bring them up to date and include new features. This patch includes:
2752    
2753        - Fixed PH's badly added libz and libbz2 support.
2754        - Fixed a problem with static linking.
2755        - Added pcredemo. [But later removed - see 7 below.]
2756        - Fixed dftables problem and added an option.
2757        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2758            HAVE_LONG_LONG.
2759        - Added readline support for pcretest.
2760        - Added an listing of the option settings after cmake has run.
2761    
2762    4.  A user submitted a patch to Makefile that makes it easy to create
2763        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2764        Makefile.am that cause it to include this special target, without
2765        affecting anything else. Note that the same mingw target plus all
2766        the other distribution libraries and programs are now supported
2767        when configuring with CMake (see 6 below) instead of with
2768        Configure/Make.
2769    
2770    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2771        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2772        exported in the Windows port". It has not yet been confirmed that the patch
2773        solves the problem, but it does no harm.
2774    
2775    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2776        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2777        with CMake, and also correct the comment about stack recursion.
2778    
2779    7.  Remove the automatic building of pcredemo from the ./configure system and
2780        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2781        of a program that users should build themselves after PCRE is installed, so
2782        building it automatically is not really right. What is more, it gave
2783        trouble in some build environments.
2784    
2785    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2786    
2787    
2788    Version 7.5 10-Jan-08
2789    ---------------------
2790    
2791    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2792        values in parens when parsing an RE using the C++ wrapper."
2793    
2794    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2795        Characters greater than 255 were excluded from the class instead of being
2796        included.
2797    
2798    3.  The same bug as (2) above applied to negated POSIX classes such as
2799        [:^space:].
2800    
2801    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2802        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2803        I have changed it.
2804    
2805    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2806        first named subpattern) and a construct such as (?&a) would reference the
2807        first named subpattern whose name started with "a" (in other words, the
2808        length check was missing). Both these problems are fixed. "Subpattern name
2809        expected" is now given for (?&) (a zero-length name), and this patch also
2810        makes it give the same error for \k'' (previously it complained that that
2811        was a reference to a non-existent subpattern).
2812    
2813    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2814        this is right because (?- can be followed by option settings as well as by
2815        digits. I have, however, made the messages clearer.
2816    
2817    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2818        than the number used in the conditional) now cause a compile-time error.
2819        This is actually not compatible with Perl, which accepts such patterns, but
2820        treats the conditional as always being FALSE (as PCRE used to), but it
2821        seems to me that giving a diagnostic is better.
2822    
2823    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2824        and messages.
2825    
2826    9.  Fix two occurrences of "backslash" in comments that should have been
2827        "backspace".
2828    
2829    10. Remove two redundant lines of code that can never be obeyed (their function
2830        was moved elsewhere).
2831    
2832    11. The program that makes PCRE's Unicode character property table had a bug
2833        which caused it to generate incorrect table entries for sequences of
2834        characters that have the same character type, but are in different scripts.
2835        It amalgamated them into a single range, with the script of the first of
2836        them. In other words, some characters were in the wrong script. There were
2837        thirteen such cases, affecting characters in the following ranges:
2838    
2839          U+002b0 - U+002c1
2840          U+0060c - U+0060d
2841          U+0061e - U+00612
2842          U+0064b - U+0065e
2843          U+0074d - U+0076d
2844          U+01800 - U+01805
2845          U+01d00 - U+01d77
2846          U+01d9b - U+01dbf
2847          U+0200b - U+0200f
2848          U+030fc - U+030fe
2849          U+03260 - U+0327f
2850          U+0fb46 - U+0fbb1
2851          U+10450 - U+1049d
2852    
2853    12. The -o option (show only the matching part of a line) for pcregrep was not
2854        compatible with GNU grep in that, if there was more than one match in a
2855        line, it showed only the first of them. It now behaves in the same way as
2856        GNU grep.
2857    
2858    13. If the -o and -v options were combined for pcregrep, it printed a blank
2859        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2860        does the same. The return code can be used to tell if there were any
2861        non-matching lines.
2862    
2863    14. Added --file-offsets and --line-offsets to pcregrep.
2864    
2865    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2866        infinitely looping recursion. The bug was that positive lookaheads were not
2867        being skipped when checking for a possible empty match (negative lookaheads
2868        and both kinds of lookbehind were skipped).
2869    
2870    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2871        inclusion of <windows.h> to before rather than after the definition of
2872        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2873    
2874    17. Specifying a possessive quantifier with a specific limit for a Unicode
2875        character property caused pcre_compile() to compile bad code, which led at
2876        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2877        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2878        caused the error; without that there was no problem.
2879    
2880    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2881    
2882    19. Added --enable-pcretest-libreadline.
2883    
2884    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2885        RE::GlobalReplace(). As a result, the number of replacements returned was
2886        double what it should be. I removed one of the increments, but Craig sent a
2887        later patch that removed the other one (the right fix) and added unit tests
2888        that check the return values (which was not done before).
2889    
2890    21. Several CMake things:
2891    
2892        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2893            the names libpcre and libpcreposix, not just pcre and pcreposix.
2894    
2895        (2) The above change means that pcretest and pcregrep are now correctly
2896            linked with the newly-built libraries, not previously installed ones.
2897    
2898        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2899    
2900    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2901        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2902        UTF-8 newline character). The key issue is that the pattern starts .*;
2903        this means that the match must be either at the beginning, or after a
2904        newline. The bug was in the code for advancing after a failed match and
2905        checking that the new position followed a newline. It was not taking
2906        account of UTF-8 characters correctly.
2907    
2908    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2909        character classes. PCRE was not treating the sequence [:...:] as a
2910        character class unless the ... were all letters. Perl, however, seems to
2911        allow any characters between [: and :], though of course it rejects as
2912        unknown any "names" that contain non-letters, because all the known class
2913        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2914        for example, whereas PCRE did not - it did not recognize a POSIX character
2915        class. This seemed a bit dangerous, so the code has been changed to be
2916        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2917        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2918        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2919        Perl does, and where it didn't before.
2920    
2921    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2922        Windows environments %n is disabled by default.
2923    
2924    
2925    Version 7.4 21-Sep-07
2926    ---------------------
2927    
2928    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2929        means that a class such as [\s] counted as "explicit reference to CR or
2930        LF". That isn't really right - the whole point of the change was to try to
2931        help when there was an actual mention of one of the two characters. So now
2932        the change happens only if \r or \n (or a literal CR or LF) character is
2933        encountered.
2934    
2935    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2936        of both had grown to the point where there were only 3 bits left.
2937        Fortunately, there was spare space in the data structure, and so I have
2938        moved the internal flags into a new 16-bit field to free up more option
2939        bits.
2940    
2941    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2942        but did not set the internal JCHANGED flag - either of these is enough to
2943        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2944        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2945        start sets both bits.
2946    
2947    4.  Added options (at build time, compile time, exec time) to change \R from
2948        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2949    
2950    5.  doc/pcresyntax.html was missing from the distribution.
2951    
2952    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2953        compatibility, even though it is no longer used.
2954    
2955    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2956        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2957        windows.h file is present (where different names are used). [This was
2958        reversed later after testing - see 16 below.]
2959    
2960    8.  Changed all #include <config.h> to #include "config.h". There were also
2961        some further <pcre.h> cases that I changed to "pcre.h".
2962    
2963    9.  When pcregrep was used with the --colour option, it missed the line ending
2964        sequence off the lines that it output.
2965    
2966    10. It was pointed out to me that arrays of string pointers cause lots of
2967        relocations when a shared library is dynamically loaded. A technique of
2968        using a single long string with a table of offsets can drastically reduce
2969        these. I have refactored PCRE in four places to do this. The result is
2970        dramatic:
2971    
2972          Originally:                          290
2973          After changing UCP table:            187
2974          After changing error message table:   43
2975          After changing table of "verbs"       36
2976          After changing table of Posix names   22
2977    
2978        Thanks to the folks working on Gregex for glib for this insight.
2979    
2980    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2981        unicode-properties was also set.
2982    
2983    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2984    
2985    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2986        checked only for CRLF.
2987    
2988    14. Added casts to pcretest.c to avoid compiler warnings.
2989    
2990    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2991    
2992    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2993        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2994        entirely. This removes changes made in 7 above.
2995    
2996    17. The CMake files have been updated, and there is now more information about
2997        building with CMake in the NON-UNIX-USE document.
2998    
2999    
3000    Version 7.3 28-Aug-07
3001  ---------------------  ---------------------
3002    
3003   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 143  Version 7.3 17-Aug-07 Line 3139  Version 7.3 17-Aug-07
3139  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte  24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3140      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").      characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3141    
3142    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3143    
3144    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3145        character were causing crashes (broken optimization).
3146    
3147    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3148        \p or \P) caused a compile-time loop.
3149    
3150    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3151        break. For example, the unstudied pattern [\r\n]A does not match the string
3152        "\r\nA" because change 7.0/46 below moves the current point on by two
3153        characters after failing to match at the start. However, the pattern \nA
3154        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3155        the same is true. There doesn't seem any very clean way out of this, but
3156        what I have chosen to do makes the common cases work: PCRE now takes note
3157        of whether there can be an explicit match for \r or \n anywhere in the
3158        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3159        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3160        pattern has explicit CR or LF references.
3161    
3162    29. Added (*CR) etc for changing newline setting at start of pattern.
3163    
3164    
3165  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3166  ---------------------  ---------------------

Legend:
Removed from v.221  
changed lines
  Added in v.1614

  ViewVC Help
Powered by ViewVC 1.1.5