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

Diff of /code/trunk/ChangeLog

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

revision 200 by ph10, Wed Aug 1 09:10:40 2007 UTC revision 1611 by ph10, Thu Nov 26 20:29:13 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 05-Jul-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 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    
15    Version 8.38 23-November-2015
16    -----------------------------
17    
18    1.  If a group that contained a recursive back reference also contained a
19        forward reference subroutine call followed by a non-forward-reference
20        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
21        compile correct code, leading to undefined behaviour or an internally
22        detected error. This bug was discovered by the LLVM fuzzer.
23    
24    2.  Quantification of certain items (e.g. atomic back references) could cause
25        incorrect code to be compiled when recursive forward references were
26        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
27        This bug was discovered by the LLVM fuzzer.
28    
29    3.  A repeated conditional group whose condition was a reference by name caused
30        a buffer overflow if there was more than one group with the given name.
31        This bug was discovered by the LLVM fuzzer.
32    
33    4.  A recursive back reference by name within a group that had the same name as
34        another group caused a buffer overflow. For example:
35        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
36    
37    5.  A forward reference by name to a group whose number is the same as the
38        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
39        a buffer overflow at compile time. This bug was discovered by the LLVM
40        fuzzer.
41    
42    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
43        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
44    
45    7.  Another buffer overflow bug involved duplicate named groups with a
46        reference between their definition, with a group that reset capture
47        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
48        fixed by always allowing for more memory, even if not needed. (A proper fix
49        is implemented in PCRE2, but it involves more refactoring.)
50    
51    8.  There was no check for integer overflow in subroutine calls such as (?123).
52    
53    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
54        being treated as a literal 'l' instead of causing an error.
55    
56    10. There was a buffer overflow if pcre_exec() was called with an ovector of
57        size 1. This bug was found by american fuzzy lop.
58    
59    11. If a non-capturing group containing a conditional group that could match
60        an empty string was repeated, it was not identified as matching an empty
61        string itself. For example: /^(?:(?(1)x|)+)+$()/.
62    
63    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
64        \a and \e in test subject lines.
65    
66    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
67        instead of the EBCDIC value.
68    
69    14. The handling of \c in an EBCDIC environment has been revised so that it is
70        now compatible with the specification in Perl's perlebcdic page.
71    
72    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
73        ASCII/Unicode. This has now been added to the list of characters that are
74        recognized as white space in EBCDIC.
75    
76    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
77        error (correctly) when used outside a class, but did not give an error
78        within a class.
79    
80    17. \h within a class was incorrectly compiled in EBCDIC environments.
81    
82    18. A pattern with an unmatched closing parenthesis that contained a backward
83        assertion which itself contained a forward reference caused buffer
84        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
85    
86    19. JIT should return with error when the compiled pattern requires more stack
87        space than the maximum.
88    
89    20. A possessively repeated conditional group that could match an empty string,
90        for example, /(?(R))*+/, was incorrectly compiled.
91    
92    21. Fix infinite recursion in the JIT compiler when certain patterns such as
93        /(?:|a|){100}x/ are analysed.
94    
95    22. Some patterns with character classes involving [: and \\ were incorrectly
96        compiled and could cause reading from uninitialized memory or an incorrect
97        error diagnosis.
98    
99    23. Pathological patterns containing many nested occurrences of [: caused
100        pcre_compile() to run for a very long time.
101    
102    24. A conditional group with only one branch has an implicit empty alternative
103        branch and must therefore be treated as potentially matching an empty
104        string.
105    
106    25. If (?R was followed by - or + incorrect behaviour happened instead of a
107        diagnostic.
108    
109    26. Arrange to give up on finding the minimum matching length for overly
110        complex patterns.
111    
112    27. Similar to (4) above: in a pattern with duplicated named groups and an
113        occurrence of (?| it is possible for an apparently non-recursive back
114        reference to become recursive if a later named group with the relevant
115        number is encountered. This could lead to a buffer overflow. Wen Guanxing
116        from Venustech ADLAB discovered this bug.
117    
118    28. If pcregrep was given the -q option with -c or -l, or when handling a
119        binary file, it incorrectly wrote output to stdout.
120    
121    29. The JIT compiler did not restore the control verb head in case of *THEN
122        control verbs. This issue was found by Karl Skomski with a custom LLVM
123        fuzzer.
124    
125    30. Error messages for syntax errors following \g and \k were giving inaccurate
126        offsets in the pattern.
127    
128    31. Added a check for integer overflow in conditions (?(<digits>) and
129        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
130        fuzzer.
131    
132    32. Handling recursive references such as (?2) when the reference is to a group
133        later in the pattern uses code that is very hacked about and error-prone.
134        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
135        give an internal error if it is obvious that compiling has gone wrong.
136    
137    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
138        This issue was found by Karl Skomski with a custom LLVM fuzzer.
139    
140    34. The JIT compiler should restore the control chain for empty possessive
141        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
142    
143    35. Match limit check added to JIT recursion. This issue was found by Karl
144        Skomski with a custom LLVM fuzzer.
145    
146    36. Yet another case similar to 27 above has been circumvented by an
147        unconditional allocation of extra memory. This issue is fixed "properly" in
148        PCRE2 by refactoring the way references are handled. Wen Guanxing
149        from Venustech ADLAB discovered this bug.
150    
151    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
152        with a custom LLVM fuzzer.
153    
154    38. Fixed a corner case of range optimization in JIT.
155    
156    39. An incorrect error "overran compiling workspace" was given if there were
157        exactly enough group forward references such that the last one extended
158        into the workspace safety margin. The next one would have expanded the
159        workspace. The test for overflow was not including the safety margin.
160    
161    40. A match limit issue is fixed in JIT which was found by Karl Skomski
162        with a custom LLVM fuzzer.
163    
164    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
165        work under Windows. (Why has it taken so long for anyone to notice?)
166    
167    42. In a character class such as [\W\p{Any}] where both a negative-type escape
168        ("not a word character") and a property escape were present, the property
169        escape was being ignored.
170    
171    43. Fix crash caused by very long (*MARK) or (*THEN) names.
172    
173    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
174        by a single ASCII character in a class item, was incorrectly compiled in
175        UCP mode. The POSIX class got lost, but only if the single character
176        followed it.
177    
178    45. [:punct:] in UCP mode was matching some characters in the range 128-255
179        that should not have been matched.
180    
181    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
182        class, all characters with code points greater than 255 are in the class.
183        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
184        such as \w are turned into Unicode properties), wide characters were not
185        correctly handled, and could fail to match.
186    
187    
188    Version 8.37 28-April-2015
189    --------------------------
190    
191    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
192        for those parentheses to be closed with whatever has been captured so far.
193        However, it was failing to mark any other groups between the hightest
194        capture so far and the currrent group as "unset". Thus, the ovector for
195        those groups contained whatever was previously there. An example is the
196        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
197    
198    2.  If an assertion condition was quantified with a minimum of zero (an odd
199        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
200    
201    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
202        unrecognized modifier, a crash could occur.
203    
204    4.  An attempt to do global matching in pcretest with a zero-length ovector
205        caused a crash.
206    
207    5.  Fixed a memory leak during matching that could occur for a subpattern
208        subroutine call (recursive or otherwise) if the number of captured groups
209        that had to be saved was greater than ten.
210    
211    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
212        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
213        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
214    
215    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
216        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
217    
218    8. When a pattern is compiled, it remembers the highest back reference so that
219       when matching, if the ovector is too small, extra memory can be obtained to
220       use instead. A conditional subpattern whose condition is a check on a
221       capture having happened, such as, for example in the pattern
222       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
223       setting the highest backreference number. This mattered only if pcre_exec()
224       was called with an ovector that was too small to hold the capture, and there
225       was no other kind of back reference (a situation which is probably quite
226       rare). The effect of the bug was that the condition was always treated as
227       FALSE when the capture could not be consulted, leading to a incorrect
228       behaviour by pcre_exec(). This bug has been fixed.
229    
230    9. A reference to a duplicated named group (either a back reference or a test
231       for being set in a conditional) that occurred in a part of the pattern where
232       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
233       to be incorrectly calculated, leading to overwriting.
234    
235    10. A mutually recursive set of back references such as (\2)(\1) caused a
236        segfault at study time (while trying to find the minimum matching length).
237        The infinite loop is now broken (with the minimum length unset, that is,
238        zero).
239    
240    11. If an assertion that was used as a condition was quantified with a minimum
241        of zero, matching went wrong. In particular, if the whole group had
242        unlimited repetition and could match an empty string, a segfault was
243        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
244        assertions to be quantified, but not if they are being used as conditions,
245        so the above pattern is faulted by Perl. PCRE has now been changed so that
246        it also rejects such patterns.
247    
248    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
249        failed to allow the zero-repeat case if pcre2_exec() was called with an
250        ovector too small to capture the group.
251    
252    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
253        Red Hat Product Security:
254    
255        (a) A crash if /K and /F were both set with the option to save the compiled
256        pattern.
257    
258        (b) Another crash if the option to print captured substrings in a callout
259        was combined with setting a null ovector, for example \O\C+ as a subject
260        string.
261    
262    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
263        forward reference repeated a large (but limited) number of times within a
264        repeated outer group that has a zero minimum quantifier, caused incorrect
265        code to be compiled, leading to the error "internal error:
266        previously-checked referenced subpattern not found" when an incorrect
267        memory address was read. This bug was reported as "heap overflow",
268        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
269        CVE-2015-2325.
270    
271    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
272        call within a group that also contained a recursive back reference caused
273        incorrect code to be compiled. This bug was reported as "heap overflow",
274        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
275        number CVE-2015-2326.
276    
277    24. Computing the size of the JIT read-only data in advance has been a source
278        of various issues, and new ones are still appear unfortunately. To fix
279        existing and future issues, size computation is eliminated from the code,
280        and replaced by on-demand memory allocation.
281    
282    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
283        adjacent to the end of the range, and the range contained characters with
284        more than one other case, caused incorrect behaviour when compiled in UTF
285        mode. In that example, the range a-j was left out of the class.
286    
287    26. Fix JIT compilation of conditional blocks, which assertion
288        is converted to (*FAIL). E.g: /(?(?!))/.
289    
290    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
291        discovered by the LLVM fuzzer.
292    
293    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
294        when this assertion was used as a condition, for example (?(?!)a|b). In
295        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
296        error about an unsupported item.
297    
298    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
299        possessification code could take exponential time to complete. A recursion
300        depth limit of 1000 has been imposed to limit the resources used by this
301        optimization.
302    
303    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
304        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
305        because \S ensures they are all in the class. The code for doing this was
306        interacting badly with the code for computing the amount of space needed to
307        compile the pattern, leading to a buffer overflow. This bug was discovered
308        by the LLVM fuzzer.
309    
310    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
311        other kinds of group caused stack overflow at compile time. This bug was
312        discovered by the LLVM fuzzer.
313    
314    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
315        between a subroutine call and its quantifier was incorrectly compiled,
316        leading to buffer overflow or other errors. This bug was discovered by the
317        LLVM fuzzer.
318    
319    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
320        assertion after (?(. The code was failing to check the character after
321        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
322        was discovered by the LLVM fuzzer.
323    
324    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
325        a fixed maximum following a group that contains a subroutine reference was
326        incorrectly compiled and could trigger buffer overflow. This bug was
327        discovered by the LLVM fuzzer.
328    
329    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
330        caused a stack overflow instead of the diagnosis of a non-fixed length
331        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
332    
333    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
334        (e.g. /(?<=\Ka)/) could make pcregrep loop.
335    
336    37. There was a similar problem to 36 in pcretest for global matches.
337    
338    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
339        and a subsequent item in the pattern caused a non-match, backtracking over
340        the repeated \X did not stop, but carried on past the start of the subject,
341        causing reference to random memory and/or a segfault. There were also some
342        other cases where backtracking after \C could crash. This set of bugs was
343        discovered by the LLVM fuzzer.
344    
345    39. The function for finding the minimum length of a matching string could take
346        a very long time if mutual recursion was present many times in a pattern,
347        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
348        method has been implemented. This infelicity was discovered by the LLVM
349        fuzzer.
350    
351    40. Static linking against the PCRE library using the pkg-config module was
352        failing on missing pthread symbols.
353    
354    
355    Version 8.36 26-September-2014
356    ------------------------------
357    
358    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
359        -Wmissing-field-initializers and -Wunused-parameter.
360    
361    2.  The tests for quantifiers being too big (greater than 65535) were being
362        applied after reading the number, and stupidly assuming that integer
363        overflow would give a negative number. The tests are now applied as the
364        numbers are read.
365    
366    3.  Tidy code in pcre_exec.c where two branches that used to be different are
367        now the same.
368    
369    4.  The JIT compiler did not generate match limit checks for certain
370        bracketed expressions with quantifiers. This may lead to exponential
371        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
372        issue should be resolved now.
373    
374    5.  Fixed an issue, which occures when nested alternatives are optimized
375        with table jumps.
376    
377    6.  Inserted two casts and changed some ints to size_t in the light of some
378        reported 64-bit compiler warnings (Bugzilla 1477).
379    
380    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
381        an empty string, which sometimes were behaving incorrectly in the
382        interpreter (though correctly in the JIT matcher). This pcretest input is
383        an example:
384    
385          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
386          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
387    
388        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
389        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
390        for an empty string was breaking the inner loop and carrying on at a lower
391        level, when possessive repeated groups should always return to a higher
392        level as they have no backtrack points in them. The empty string test now
393        occurs at the outer level.
394    
395    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
396        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
397    
398    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
399        doesn't).
400    
401    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
402        to hold a pattern to be incorrectly computed (too small) when there were
403        named back references to duplicated names. This could cause "internal
404        error: code overflow" or "double free or corruption" or other memory
405        handling errors.
406    
407    11. When named subpatterns had the same prefixes, back references could be
408        confused. For example, in this pattern:
409    
410          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
411    
412        the reference to 'Name' was incorrectly treated as a reference to a
413        duplicate name.
414    
415    12. A pattern such as /^s?c/mi8 where the optional character has more than
416        one "other case" was incorrectly compiled such that it would only try to
417        match starting at "c".
418    
419    13. When a pattern starting with \s was studied, VT was not included in the
420        list of possible starting characters; this should have been part of the
421        8.34/18 patch.
422    
423    14. If a character class started [\Qx]... where x is any character, the class
424        was incorrectly terminated at the ].
425    
426    15. If a pattern that started with a caseless match for a character with more
427        than one "other case" was studied, PCRE did not set up the starting code
428        unit bit map for the list of possible characters. Now it does. This is an
429        optimization improvement, not a bug fix.
430    
431    16. The Unicode data tables have been updated to Unicode 7.0.0.
432    
433    17. Fixed a number of memory leaks in pcregrep.
434    
435    18. Avoid a compiler warning (from some compilers) for a function call with
436        a cast that removes "const" from an lvalue by using an intermediate
437        variable (to which the compiler does not object).
438    
439    19. Incorrect code was compiled if a group that contained an internal recursive
440        back reference was optional (had quantifier with a minimum of zero). This
441        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
442        caused segmentation faults because of stack overflows at compile time.
443    
444    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
445        group that is quantified with an indefinite repeat, caused a compile-time
446        loop which used up all the system stack and provoked a segmentation fault.
447        This was not the same bug as 19 above.
448    
449    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
450        Patch by Mike Frysinger.
451    
452    
453    Version 8.35 04-April-2014
454    --------------------------
455    
456    1.  A new flag is set, when property checks are present in an XCLASS.
457        When this flag is not set, PCRE can perform certain optimizations
458        such as studying these XCLASS-es.
459    
460    2.  The auto-possessification of character sets were improved: a normal
461        and an extended character set can be compared now. Furthermore
462        the JIT compiler optimizes more character set checks.
463    
464    3.  Got rid of some compiler warnings for potentially uninitialized variables
465        that show up only when compiled with -O2.
466    
467    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
468        of a match later then the end of the match. The pcretest program was not
469        handling the case sensibly - it was outputting from the start to the next
470        binary zero. It now reports this situation in a message, and outputs the
471        text from the end to the start.
472    
473    5.  Fast forward search is improved in JIT. Instead of the first three
474        characters, any three characters with fixed position can be searched.
475        Search order: first, last, middle.
476    
477    6.  Improve character range checks in JIT. Characters are read by an inprecise
478        function now, which returns with an unknown value if the character code is
479        above a certain threshold (e.g: 256). The only limitation is that the value
480        must be bigger than the threshold as well. This function is useful when
481        the characters above the threshold are handled in the same way.
482    
483    7.  The macros whose names start with RAWUCHAR are placeholders for a future
484        mode in which only the bottom 21 bits of 32-bit data items are used. To
485        make this more memorable for those maintaining the code, the names have
486        been changed to start with UCHAR21, and an extensive comment has been added
487        to their definition.
488    
489    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
490        to the export list in Makefile.am (they were accidentally omitted from the
491        8.34 tarball).
492    
493    9.  The informational output from pcretest used the phrase "starting byte set"
494        which is inappropriate for the 16-bit and 32-bit libraries. As the output
495        for "first char" and "need char" really means "non-UTF-char", I've changed
496        "byte" to "char", and slightly reworded the output. The documentation about
497        these values has also been (I hope) clarified.
498    
499    10. Another JIT related optimization: use table jumps for selecting the correct
500        backtracking path, when more than four alternatives are present inside a
501        bracket.
502    
503    11. Empty match is not possible, when the minimum length is greater than zero,
504        and there is no \K in the pattern. JIT should avoid empty match checks in
505        such cases.
506    
507    12. In a caseless character class with UCP support, when a character with more
508        than one alternative case was not the first character of a range, not all
509        the alternative cases were added to the class. For example, s and \x{17f}
510        are both alternative cases for S: the class [RST] was handled correctly,
511        but [R-T] was not.
512    
513    13. The configure.ac file always checked for pthread support when JIT was
514        enabled. This is not used in Windows, so I have put this test inside a
515        check for the presence of windows.h (which was already tested for).
516    
517    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
518        The algorithm provides a way to skip certain starting offsets, and usually
519        faster than linear prefix searches.
520    
521    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
522        as for 'fr_FR' and 'french'. For some reason, however, it then used the
523        Windows-specific input and output files, which have 'french' screwed in.
524        So this could never have worked. One of the problems with locales is that
525        they aren't always the same. I have now updated RunTest so that it checks
526        the output of the locale test (test 3) against three different output
527        files, and it allows the test to pass if any one of them matches. With luck
528        this should make the test pass on some versions of Solaris where it was
529        failing. Because of the uncertainty, the script did not used to stop if
530        test 3 failed; it now does. If further versions of a French locale ever
531        come to light, they can now easily be added.
532    
533    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
534        there was a message during ./configure, but it did not stop. This now
535        provokes an error. The invalid example in README has been corrected.
536        If a value less than the minimum is given, the minimum value has always
537        been used, but now a warning is given.
538    
539    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
540        was a bug in the test system, which is now fixed. Also, the list of various
541        configurations that are tested for each release did not have one with both
542        16/32 bits and --enable-bar-anycrlf. It now does.
543    
544    18. pcretest was missing "-C bsr" for displaying the \R default setting.
545    
546    19. Little endian PowerPC systems are supported now by the JIT compiler.
547    
548    20. The fast forward newline mechanism could enter to an infinite loop on
549        certain invalid UTF-8 input. Although we don't support these cases
550        this issue can be fixed by a performance optimization.
551    
552    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
553        not take account if existing stack usage. There is now a new global
554        variable called pcre_stack_guard that can be set to point to an external
555        function to check stack availability. It is called at the start of
556        processing every parenthesized group.
557    
558    22. A typo in the code meant that in ungreedy mode the max/min qualifier
559        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
560        not match "ab".
561    
562    23. When UTF was disabled, the JIT program reported some incorrect compile
563        errors. These messages are silenced now.
564    
565    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
566        compiler.
567    
568    25. Change all the temporary files used in RunGrepTest to be different to those
569        used by RunTest so that the tests can be run simultaneously, for example by
570        "make -j check".
571    
572    
573    Version 8.34 15-December-2013
574    -----------------------------
575    
576    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
577        executable memory. Patch inspired by Carsten Klein.
578    
579    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
580        this macro is never tested and has no effect, because the work to support
581        coverage involves only compiling and linking options and special targets in
582        the Makefile. The comment in config.h implied that defining the macro would
583        enable coverage support, which is totally false. There was also support for
584        setting this macro in the CMake files (my fault, I just copied it from
585        configure). SUPPORT_GCOV has now been removed.
586    
587    3.  Make a small performance improvement in strlen16() and strlen32() in
588        pcretest.
589    
590    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
591        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
592        these cases). There was also one in pcretest.c.
593    
594    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
595    
596    6.  In UTF mode, the code for checking whether a group could match an empty
597        string (which is used for indefinitely repeated groups to allow for
598        breaking an infinite loop) was broken when the group contained a repeated
599        negated single-character class with a character that occupied more than one
600        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
601        UTF-8 mode). The effect was undefined: the group might or might not be
602        deemed as matching an empty string, or the program might have crashed.
603    
604    7.  The code for checking whether a group could match an empty string was not
605        recognizing that \h, \H, \v, \V, and \R must match a character.
606    
607    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
608        an empty string. If it can, pcretest shows this in its information output.
609    
610    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
611        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
612        matched by pcre_exec() without using JIT:
613    
614        (a) If the rest of the pattern did not match after a maximal run of
615            grapheme clusters, the code for backing up to try with fewer of them
616            did not always back up over a full grapheme when characters that do not
617            have the modifier quality were involved, e.g. Hangul syllables.
618    
619        (b) If the match point in a subject started with modifier character, and
620            there was no match, the code could incorrectly back up beyond the match
621            point, and potentially beyond the first character in the subject,
622            leading to a segfault or an incorrect match result.
623    
624    10. A conditional group with an assertion condition could lead to PCRE
625        recording an incorrect first data item for a match if no other first data
626        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
627        first data item, and therefore matched "ca" after "c" instead of at the
628        start.
629    
630    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
631        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
632    
633    12. The source of pcregrep now includes z/OS-specific code so that it can be
634        compiled for z/OS as part of the special z/OS distribution.
635    
636    13. Added the -T and -TM options to pcretest.
637    
638    14. The code in pcre_compile.c for creating the table of named capturing groups
639        has been refactored. Instead of creating the table dynamically during the
640        actual compiling pass, the information is remembered during the pre-compile
641        pass (on the stack unless there are more than 20 named groups, in which
642        case malloc() is used) and the whole table is created before the actual
643        compile happens. This has simplified the code (it is now nearly 150 lines
644        shorter) and prepared the way for better handling of references to groups
645        with duplicate names.
646    
647    15. A back reference to a named subpattern when there is more than one of the
648        same name now checks them in the order in which they appear in the pattern.
649        The first one that is set is used for the reference. Previously only the
650        first one was inspected. This change makes PCRE more compatible with Perl.
651    
652    16. Unicode character properties were updated from Unicode 6.3.0.
653    
654    17. The compile-time code for auto-possessification has been refactored, based
655        on a patch by Zoltan Herczeg. It now happens after instead of during
656        compilation. The code is cleaner, and more cases are handled. The option
657        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
658        options in pcretest are provided to set it. It can also be set by
659        (*NO_AUTO_POSSESS) at the start of a pattern.
660    
661    18. The character VT has been added to the default ("C" locale) set of
662        characters that match \s and are generally treated as white space,
663        following this same change in Perl 5.18. There is now no difference between
664        "Perl space" and "POSIX space". Whether VT is treated as white space in
665        other locales depends on the locale.
666    
667    19. The code for checking named groups as conditions, either for being set or
668        for being recursed, has been refactored (this is related to 14 and 15
669        above). Processing unduplicated named groups should now be as fast at
670        numerical groups, and processing duplicated groups should be faster than
671        before.
672    
673    20. Two patches to the CMake build system, by Alexander Barkov:
674    
675          (1) Replace the "source" command by "." in CMakeLists.txt because
676              "source" is a bash-ism.
677    
678          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
679              without these the CMake build does not work on Solaris.
680    
681    21. Perl has changed its handling of \8 and \9. If there is no previously
682        encountered capturing group of those numbers, they are treated as the
683        literal characters 8 and 9 instead of a binary zero followed by the
684        literals. PCRE now does the same.
685    
686    22. Following Perl, added \o{} to specify codepoints in octal, making it
687        possible to specify values greater than 0777 and also making them
688        unambiguous.
689    
690    23. Perl now gives an error for missing closing braces after \x{... instead of
691        treating the string as literal. PCRE now does the same.
692    
693    24. RunTest used to grumble if an inappropriate test was selected explicitly,
694        but just skip it when running all tests. This make it awkward to run ranges
695        of tests when one of them was inappropriate. Now it just skips any
696        inappropriate tests, as it always did when running all tests.
697    
698    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
699        character types such as \d or \w, too many callouts were inserted, and the
700        data that they returned was rubbish.
701    
702    26. In UCP mode, \s was not matching two of the characters that Perl matches,
703        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
704        were matched by \h. The code has now been refactored so that the lists of
705        the horizontal and vertical whitespace characters used for \h and \v (which
706        are defined only in one place) are now also used for \s.
707    
708    27. Add JIT support for the 64 bit TileGX architecture.
709        Patch by Jiong Wang (Tilera Corporation).
710    
711    28. Possessive quantifiers for classes (both explicit and automatically
712        generated) now use special opcodes instead of wrapping in ONCE brackets.
713    
714    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
715        (because it's meaningless), this was not happening when PCRE_CASELESS was
716        set. Not wrong, but inefficient.
717    
718    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
719        properties for \w, \d, etc) is present in a test regex. Otherwise if the
720        test contains no characters greater than 255, Perl doesn't realise it
721        should be using Unicode semantics.
722    
723    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
724        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
725        does in Unicode mode.
726    
727    32. Added the "forbid" facility to pcretest so that putting tests into the
728        wrong test files can sometimes be quickly detected.
729    
730    33. There is now a limit (default 250) on the depth of nesting of parentheses.
731        This limit is imposed to control the amount of system stack used at compile
732        time. It can be changed at build time by --with-parens-nest-limit=xxx or
733        the equivalent in CMake.
734    
735    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
736        errors. Perl warns for these when in warning mode, but PCRE has no facility
737        for giving warnings.
738    
739    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
740        However, this was not working for (?!) because it is optimized to (*FAIL),
741        for which PCRE does not allow quantifiers. The optimization is now disabled
742        when a quantifier follows (?!). I can't see any use for this, but it makes
743        things uniform.
744    
745    36. Perl no longer allows group names to start with digits, so I have made this
746        change also in PCRE. It simplifies the code a bit.
747    
748    37. In extended mode, Perl ignores spaces before a + that indicates a
749        possessive quantifier. PCRE allowed a space before the quantifier, but not
750        before the possessive +. It now does.
751    
752    38. The use of \K (reset reported match start) within a repeated possessive
753        group such as (a\Kb)*+ was not working.
754    
755    40. Document that the same character tables must be used at compile time and
756        run time, and that the facility to pass tables to pcre_exec() and
757        pcre_dfa_exec() is for use only with saved/restored patterns.
758    
759    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
760        features for Builds with MSVC:
761    
762        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
763           builds). The .rc files can be used to set FileDescription and many other
764           attributes.
765    
766        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
767           This allows higher-level build scripts which want .pdb files to avoid
768           hard-coding the exact files needed."
769    
770    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
771        mean "start of word" and "end of word", respectively, as a transition aid.
772    
773    43. A minimizing repeat of a class containing codepoints greater than 255 in
774        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
775        compiled to use the heap for recursion.
776    
777    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
778        is configured.
779    
780    
781    Version 8.33 28-May-2013
782    ------------------------
783    
784    1.  Added 'U' to some constants that are compared to unsigned integers, to
785        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
786        variables that are added to signed variables, to ensure the result is
787        signed and can be negated.
788    
789    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
790        CMake config files.
791    
792    3.  Revise the creation of config.h.generic so that all boolean macros are
793        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
794        overriding via -D on the command line possible.
795    
796    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
797        to unsigned int is reported to make a quite noticeable speed difference in
798        a specific Windows environment. Testing on Linux did also appear to show
799        some benefit (and it is clearly not harmful). Also fixed the definition of
800        Xop which should be unsigned.
801    
802    5.  Related to (4), changing the definition of the intermediate variable cc
803        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
804        improvements.
805    
806    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
807        unnecessary spaces.
808    
809    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
810        and later.
811    
812    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
813    
814    9.  Optimizing fast_forward_start_bits in JIT.
815    
816    10. Adding support for callouts in JIT, and fixing some issues revealed
817        during this work. Namely:
818    
819        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
820    
821        (b) Minimum length was not checked before the matching is started.
822    
823    11. The value of capture_last that is passed to callouts was incorrect in some
824        cases when there was a capture on one path that was subsequently abandoned
825        after a backtrack. Also, the capture_last value is now reset after a
826        recursion, since all captures are also reset in this case.
827    
828    12. The interpreter no longer returns the "too many substrings" error in the
829        case when an overflowing capture is in a branch that is subsequently
830        abandoned after a backtrack.
831    
832    13. In the pathological case when an offset vector of size 2 is used, pcretest
833        now prints out the matched string after a yield of 0 or 1.
834    
835    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
836        Only supported by the JIT compiler at the moment.
837    
838    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
839    
840    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
841        offset of the starting point of the matching process, provided the offsets
842        vector is large enough.
843    
844    17. The \A escape now records a lookbehind value of 1, though its execution
845        does not actually inspect the previous character. This is to ensure that,
846        in partial multi-segment matching, at least one character from the old
847        segment is retained when a new segment is processed. Otherwise, if there
848        are no lookbehinds in the pattern, \A might match incorrectly at the start
849        of a new segment.
850    
851    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
852    
853    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
854        gives some modest performance improvement in 8-bit mode.
855    
856    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
857        be expressed in certain programming languages using Universal Character
858        Names.
859    
860    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
861        which points out that "non characters" are not "characters that may not
862        appear in Unicode strings" but rather "characters that are reserved for
863        internal use and have only local meaning".
864    
865    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
866        there was a conditional group that depended on an assertion, if the
867        assertion was false, the callout that immediately followed the alternation
868        in the condition was skipped when pcre_exec() was used for matching.
869    
870    23. Allow an explicit callout to be inserted before an assertion that is the
871        condition for a conditional group, for compatibility with automatic
872        callouts, which always insert a callout at this point.
873    
874    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
875        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
876    
877    25. (*PRUNE) is now supported by the JIT compiler.
878    
879    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
880    
881    27. Fix the case where there are two or more SKIPs with arguments that may be
882        ignored.
883    
884    28. (*SKIP) is now supported by the JIT compiler.
885    
886    29. (*THEN) is now supported by the JIT compiler.
887    
888    30. Update RunTest with additional test selector options.
889    
890    31. The way PCRE handles backtracking verbs has been changed in two ways.
891    
892        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
893        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
894        backtracking. In some cases this makes it more Perl-compatible, but Perl's
895        rather obscure rules do not always do the same thing.
896    
897        (2) Previously, backtracking verbs were confined within assertions. This is
898        no longer the case for positive assertions, except for (*ACCEPT). Again,
899        this sometimes improves Perl compatibility, and sometimes does not.
900    
901    32. A number of tests that were in test 2 because Perl did things differently
902        have been moved to test 1, because either Perl or PCRE has changed, and
903        these tests are now compatible.
904    
905    32. Backtracking control verbs are now handled in the same way in JIT and
906        interpreter.
907    
908    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
909        contained a forward subroutine reference caused a compile error.
910    
911    34. Auto-detect and optimize limited repetitions in JIT.
912    
913    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
914        blocking (*UTF) etc.
915    
916    36. In the interpreter, maximizing pattern repetitions for characters and
917        character types now use tail recursion, which reduces stack usage.
918    
919    37. The value of the max lookbehind was not correctly preserved if a compiled
920        and saved regex was reloaded on a host of different endianness.
921    
922    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
923        of the compiled pattern block, expand the flags field from 16 to 32 bits
924        because it was almost full.
925    
926    39. Try madvise first before posix_madvise.
927    
928    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
929        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
930        original change locked out all matches of empty strings. This has been
931        changed so that one match of an empty string per line is recognized.
932        Subsequent searches on the same line (for colouring or for --only-matching,
933        for example) do not recognize empty strings.
934    
935    41. Applied a user patch to fix a number of spelling mistakes in comments.
936    
937    42. Data lines longer than 65536 caused pcretest to crash.
938    
939    43. Clarified the data type for length and startoffset arguments for pcre_exec
940        and pcre_dfa_exec in the function-specific man pages, where they were
941        explicitly stated to be in bytes, never having been updated. I also added
942        some clarification to the pcreapi man page.
943    
944    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
945        a segmentation fault.
946    
947    
948    Version 8.32 30-November-2012
949    -----------------------------
950    
951    1.  Improved JIT compiler optimizations for first character search and single
952        character iterators.
953    
954    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
955        Patch by Daniel Richard G.
956    
957    3.  Single character iterator optimizations in the JIT compiler.
958    
959    4.  Improved JIT compiler optimizations for character ranges.
960    
961    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
962        Reported by Giuseppe D'Angelo.
963    
964    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
965        of a line, was being set incorrectly in cases where .* appeared inside
966        atomic brackets at the start of a pattern, or where there was a subsequent
967        *PRUNE or *SKIP.
968    
969    7.  Improved instruction cache flush for POWER/PowerPC.
970        Patch by Daniel Richard G.
971    
972    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
973        grep:
974    
975        (a) There is now no limit to the number of patterns to be matched.
976    
977        (b) An error is given if a pattern is too long.
978    
979        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
980            are now supported.
981    
982        (d) --exclude-from and --include-from (multiple use) have been added.
983    
984        (e) Exclusions and inclusions now apply to all files and directories, not
985            just to those obtained from scanning a directory recursively.
986    
987        (f) Multiple uses of -f and --file-list are now supported.
988    
989        (g) In a Windows environment, the default for -d has been changed from
990            "read" (the GNU grep default) to "skip", because otherwise the presence
991            of a directory in the file list provokes an error.
992    
993        (h) The documentation has been revised and clarified in places.
994    
995    9.  Improve the matching speed of capturing brackets.
996    
997    10. Changed the meaning of \X so that it now matches a Unicode extended
998        grapheme cluster.
999    
1000    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1001        out POSIX threads when JIT support is configured.
1002    
1003    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1004    
1005    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1006        structure could go wrong in environments where size_t is not the same size
1007        as int.
1008    
1009    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1010        set.
1011    
1012    15. The EBCDIC support had decayed; later updates to the code had included
1013        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1014        general tidy up of EBCDIC-related issues, and the documentation was also
1015        not quite right. There is now a test that can be run on ASCII systems to
1016        check some of the EBCDIC-related things (but is it not a full test).
1017    
1018    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1019        in a small tidy to the code.
1020    
1021    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1022    
1023    18. If the --only-matching (-o) option in pcregrep is specified multiple
1024        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1025        substrings matched by the 1st and 2nd capturing parentheses. A separating
1026        string can be specified by --om-separator (default empty).
1027    
1028    19. Improving the first n character searches.
1029    
1030    20. Turn case lists for horizontal and vertical white space into macros so that
1031        they are defined only once.
1032    
1033    21. This set of changes together give more compatible Unicode case-folding
1034        behaviour for characters that have more than one other case when UCP
1035        support is available.
1036    
1037        (a) The Unicode property table now has offsets into a new table of sets of
1038            three or more characters that are case-equivalent. The MultiStage2.py
1039            script that generates these tables (the pcre_ucd.c file) now scans
1040            CaseFolding.txt instead of UnicodeData.txt for character case
1041            information.
1042    
1043        (b) The code for adding characters or ranges of characters to a character
1044            class has been abstracted into a generalized function that also handles
1045            case-independence. In UTF-mode with UCP support, this uses the new data
1046            to handle characters with more than one other case.
1047    
1048        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1049            whose other case is greater than 256 are now correctly matched
1050            caselessly. Previously, the high codepoint matched the low one, but not
1051            vice versa.
1052    
1053        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1054            of the new class addition function, using character lists defined as
1055            macros alongside the case definitions of 20 above.
1056    
1057        (e) Caseless back references now work with characters that have more than
1058            one other case.
1059    
1060        (f) General caseless matching of characters with more than one other case
1061            is supported.
1062    
1063    22. Unicode character properties were updated from Unicode 6.2.0
1064    
1065    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1066    
1067    24. Add support for 32-bit character strings, and UTF-32
1068    
1069    25. Major JIT compiler update (code refactoring and bugfixing).
1070        Experimental Sparc 32 support is added.
1071    
1072    26. Applied a modified version of Daniel Richard G's patch to create
1073        pcre.h.generic and config.h.generic by "make" instead of in the
1074        PrepareRelease script.
1075    
1076    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1077        pcre_compile.c when checking for a zero character.
1078    
1079    28. Introducing a native interface for JIT. Through this interface, the compiled
1080        machine code can be directly executed. The purpose of this interface is to
1081        provide fast pattern matching, so several sanity checks are not performed.
1082        However, feature tests are still performed. The new interface provides
1083        1.4x speedup compared to the old one.
1084    
1085    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1086        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1087        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1088    
1089    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1090        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1091        the "old" RFC 2279). Instead, it ended up passing a negative length to
1092        pcre_exec().
1093    
1094    31. Add support for GCC's visibility feature to hide internal functions.
1095    
1096    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1097        "unknown -C option" after outputting 0 or 1.
1098    
1099    33. There is now support for generating a code coverage report for the test
1100        suite in environments where gcc is the compiler and lcov is installed. This
1101        is mainly for the benefit of the developers.
1102    
1103    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1104        unaddressable using valgrind annotations, allowing valgrind to detect
1105        invalid memory accesses. This is mainly for the benefit of the developers.
1106    
1107    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1108    
1109    26. Give configure error if --enable-cpp but no C++ compiler found.
1110    
1111    
1112    Version 8.31 06-July-2012
1113    -------------------------
1114    
1115    1.  Fixing a wrong JIT test case and some compiler warnings.
1116    
1117    2.  Removed a bashism from the RunTest script.
1118    
1119    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1120        to unsigned type, result still unsigned" that was given by an MS compiler
1121        on encountering the code "-sizeof(xxx)".
1122    
1123    4.  Partial matching support is added to the JIT compiler.
1124    
1125    5.  Fixed several bugs concerned with partial matching of items that consist
1126        of more than one character:
1127    
1128        (a) /^(..)\1/ did not partially match "aba" because checking references was
1129            done on an "all or nothing" basis. This also applied to repeated
1130            references.
1131    
1132        (b) \R did not give a hard partial match if \r was found at the end of the
1133            subject.
1134    
1135        (c) \X did not give a hard partial match after matching one or more
1136            characters at the end of the subject.
1137    
1138        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1139            a partial match for the string "\r".
1140    
1141        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1142            a partial match for a CR character at the end of the subject string.
1143    
1144    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1145        running pcretest, the text "(JIT)" added to the output whenever JIT is
1146        actually used to run the match.
1147    
1148    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1149        or /S+[+] with a digit between 1 and 7.
1150    
1151    8.  OP_NOT now supports any UTF character not just single-byte ones.
1152    
1153    9.  (*MARK) control verb is now supported by the JIT compiler.
1154    
1155    10. The command "./RunTest list" lists the available tests without actually
1156        running any of them. (Because I keep forgetting what they all are.)
1157    
1158    11. Add PCRE_INFO_MAXLOOKBEHIND.
1159    
1160    12. Applied a (slightly modified) user-supplied patch that improves performance
1161        when the heap is used for recursion (compiled with --disable-stack-for-
1162        recursion). Instead of malloc and free for each heap frame each time a
1163        logical recursion happens, frames are retained on a chain and re-used where
1164        possible. This sometimes gives as much as 30% improvement.
1165    
1166    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1167        call.
1168    
1169    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1170    
1171    15. It is now possible to link pcretest with libedit as an alternative to
1172        libreadline.
1173    
1174    16. (*COMMIT) control verb is now supported by the JIT compiler.
1175    
1176    17. The Unicode data tables have been updated to Unicode 6.1.0.
1177    
1178    18. Added --file-list option to pcregrep.
1179    
1180    19. Added binary file support to pcregrep, including the -a, --binary-files,
1181        -I, and --text options.
1182    
1183    20. The madvise function is renamed for posix_madvise for QNX compatibility
1184        reasons. Fixed by Giuseppe D'Angelo.
1185    
1186    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1187    
1188    22. Changed the option for creating symbolic links for 16-bit man pages from
1189        -s to -sf so that re-installing does not cause issues.
1190    
1191    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1192    
1193    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1194        matches in certain environments (the workspace was not being correctly
1195        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1196        some of the workspace data at the beginning of a restart.
1197    
1198    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1199        was not doing so when it should - probably a typo introduced by SVN 528
1200        (change 8.10/14).
1201    
1202    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1203        \w+ when the character tables indicated that \x{c4} was a word character.
1204        There were several related cases, all because the tests for doing a table
1205        lookup were testing for characters less than 127 instead of 255.
1206    
1207    27. If a pattern contains capturing parentheses that are not used in a match,
1208        their slots in the ovector are set to -1. For those that are higher than
1209        any matched groups, this happens at the end of processing. In the case when
1210        there were back references that the ovector was too small to contain
1211        (causing temporary malloc'd memory to be used during matching), and the
1212        highest capturing number was not used, memory off the end of the ovector
1213        was incorrectly being set to -1. (It was using the size of the temporary
1214        memory instead of the true size.)
1215    
1216    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1217        ovector size, it uses memory at the end of the block that it has got.
1218    
1219    29. Check for an overlong MARK name and give an error at compile time. The
1220        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1221    
1222    30. JIT compiler update.
1223    
1224    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1225        Rill for the patch.
1226    
1227    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1228    
1229    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1230    
1231    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1232        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1233        was enabled.
1234    
1235    35. Improve JIT code generation for greedy plus quantifier.
1236    
1237    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1238        1 to "aa" instead of to an empty string. The bug affected repeated groups
1239        that could potentially match an empty string.
1240    
1241    37. Optimizing single character iterators in JIT.
1242    
1243    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1244        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1245        codepoints that are too big for the mode are faulted, and in a UTF mode,
1246        disallowed codepoints are also faulted.
1247    
1248    39. If PCRE was compiled with UTF support, in three places in the DFA
1249        matcher there was code that should only have been obeyed in UTF mode, but
1250        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1251        processing when bytes with values greater than 127 were present. In 16-bit
1252        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1253        both cases the values are those that cannot be the first data item in a UTF
1254        character. The three items that might have provoked this were recursions,
1255        possessively repeated groups, and atomic groups.
1256    
1257    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1258        and pcregrep, because some OS require shared objects to be explicitly
1259        passed to ld, causing the link step to fail if they are not.
1260    
1261    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1262        mode, patterns that started with \h* or \R* might be incorrectly matched.
1263    
1264    
1265    Version 8.30 04-February-2012
1266    -----------------------------
1267    
1268    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1269        name is defined in ctype.h.
1270    
1271    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1272        only in quite long subpatterns.
1273    
1274    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1275        since it was replaced by pcre_fullinfo() in February 2000.
1276    
1277    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1278        match a (*MARK), and the match failed at the start of the subject, a
1279        reference to memory before the start of the subject could occur. This bug
1280        was introduced by fix 17 of release 8.21.
1281    
1282    5.  A reference to an unset group with zero minimum repetition was giving
1283        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1284        /(another)?(\1?)test/ matched against "hello world test". This bug was
1285        introduced in release 8.13.
1286    
1287    6.  Add support for 16-bit character strings (a large amount of work involving
1288        many changes and refactorings).
1289    
1290    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1291        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1292        from a file.
1293    
1294    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1295        rounding is not applied in this particular case).
1296    
1297    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1298        if they appear, or are escaped, in patterns.
1299    
1300    10. Get rid of a number of -Wunused-but-set-variable warnings.
1301    
1302    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1303        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1304        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1305        also returns the mark "x". This bug applied to capturing parentheses,
1306        non-capturing parentheses, and atomic parentheses. It also applied to some
1307        assertions.
1308    
1309    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1310        information out of configure.ac instead of relying on pcre.h.generic, which
1311        is not stored in the repository.
1312    
1313    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1314        -lreadline.
1315    
1316    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1317    
1318    15. Applied Graycode's patch to put the top-level frame on the stack rather
1319        than the heap when not using the stack for recursion. This gives a
1320        performance improvement in many cases when recursion is not deep.
1321    
1322    16. Experimental code added to "pcretest -C" to output the stack frame size.
1323    
1324    
1325    Version 8.21 12-Dec-2011
1326    ------------------------
1327    
1328    1.  Updating the JIT compiler.
1329    
1330    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1331        are added as well.
1332    
1333    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1334        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1335        calling _pcre_jit_exec. Some extra comments are added.
1336    
1337    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1338        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1339        was introduced by change 18 for 8.20.
1340    
1341    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1342        ECMA-262 standard.
1343    
1344    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1345        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1346        This bug was probably introduced by change 9 of 8.13.
1347    
1348    7.  While fixing 6 above, I noticed that a number of other items were being
1349        incorrectly rejected as "not fixed length". This arose partly because newer
1350        opcodes had not been added to the fixed-length checking code. I have (a)
1351        corrected the bug and added tests for these items, and (b) arranged for an
1352        error to occur if an unknown opcode is encountered while checking for fixed
1353        length instead of just assuming "not fixed length". The items that were
1354        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1355        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1356        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1357    
1358    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1359        being incorrectly compiled and would have given unpredicatble results.
1360    
1361    9.  A possessively repeated subpattern with minimum repeat count greater than
1362        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1363        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1364        the first (A) could occur when it should not.
1365    
1366    10. Add a cast and remove a redundant test from the code.
1367    
1368    11. JIT should use pcre_malloc/pcre_free for allocation.
1369    
1370    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1371        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1372        is anything other than /usr, -L is still shown).
1373    
1374    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1375    
1376    14. Perl does not support \N without a following name in a [] class; PCRE now
1377        also gives an error.
1378    
1379    15. If a forward reference was repeated with an upper limit of around 2000,
1380        it caused the error "internal error: overran compiling workspace". The
1381        maximum number of forward references (including repeats) was limited by the
1382        internal workspace, and dependent on the LINK_SIZE. The code has been
1383        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1384        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1385        of around 200,000 forward references. While doing this, I also speeded up
1386        the filling in of repeated forward references.
1387    
1388    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1389        incorrectly expecting the subject to contain another "a" after the start.
1390    
1391    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1392        in the match, the SKIP should be ignored. This was not happening; instead
1393        the SKIP was being treated as NOMATCH. For patterns such as
1394        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1395        tested.
1396    
1397    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1398        now much more compatible with Perl, in particular in cases where the result
1399        is a non-match for a non-anchored pattern. For example, if
1400        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1401        "m", where previously it did not return a name. A side effect of this
1402        change is that for partial matches, the last encountered mark name is
1403        returned, as for non matches. A number of tests that were previously not
1404        Perl-compatible have been moved into the Perl-compatible test files. The
1405        refactoring has had the pleasing side effect of removing one argument from
1406        the match() function, thus reducing its stack requirements.
1407    
1408    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1409        subsequent uses of /S (without +) incorrectly behaved like /S+.
1410    
1411    21. Retrieve executable code size support for the JIT compiler and fixing
1412        some warnings.
1413    
1414    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1415        not work when the shorter character appeared right at the end of the
1416        subject string.
1417    
1418    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1419        systems.
1420    
1421    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1422        output it when the /M option is used in pcretest.
1423    
1424    25. The CheckMan script was not being included in the distribution. Also, added
1425        an explicit "perl" to run Perl scripts from the PrepareRelease script
1426        because this is reportedly needed in Windows.
1427    
1428    26. If study data was being save in a file and studying had not found a set of
1429        "starts with" bytes for the pattern, the data written to the file (though
1430        never used) was taken from uninitialized memory and so caused valgrind to
1431        complain.
1432    
1433    27. Updated RunTest.bat as provided by Sheri Pierce.
1434    
1435    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1436    
1437    29. Computation of memory usage for the table of capturing group names was
1438        giving an unnecessarily large value.
1439    
1440    
1441    Version 8.20 21-Oct-2011
1442    ------------------------
1443    
1444    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1445        a POSIX class. After further experiments with Perl, which convinced me that
1446        Perl has bugs and confusions, a closing square bracket is no longer allowed
1447        in a POSIX name. This bug also affected patterns with classes that started
1448        with full stops.
1449    
1450    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1451        captured substring, but while checking the failing first alternative,
1452        substring 1 is temporarily captured. If the output vector supplied to
1453        pcre_exec() was not big enough for this capture, the yield of the function
1454        was still zero ("insufficient space for captured substrings"). This cannot
1455        be totally fixed without adding another stack variable, which seems a lot
1456        of expense for a edge case. However, I have improved the situation in cases
1457        such as /(a)(b)x|abc/ matched against "abc", where the return code
1458        indicates that fewer than the maximum number of slots in the ovector have
1459        been set.
1460    
1461    3.  Related to (2) above: when there are more back references in a pattern than
1462        slots in the output vector, pcre_exec() uses temporary memory during
1463        matching, and copies in the captures as far as possible afterwards. It was
1464        using the entire output vector, but this conflicts with the specification
1465        that only 2/3 is used for passing back captured substrings. Now it uses
1466        only the first 2/3, for compatibility. This is, of course, another edge
1467        case.
1468    
1469    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1470        main code base, and can be used by building with --enable-jit. When this is
1471        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1472        runtime --no-jit option is given.
1473    
1474    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1475        ovector, the return from the function was zero, implying that there were
1476        other matches that did not fit. The correct "exactly full" value is now
1477        returned.
1478    
1479    6.  If a subpattern that was called recursively or as a subroutine contained
1480        (*PRUNE) or any other control that caused it to give a non-standard return,
1481        invalid errors such as "Error -26 (nested recursion at the same subject
1482        position)" or even infinite loops could occur.
1483    
1484    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1485        computing the minimum length on reaching *ACCEPT, and so ended up with the
1486        wrong value of 1 rather than 0. Further investigation indicates that
1487        computing a minimum subject length in the presence of *ACCEPT is difficult
1488        (think back references, subroutine calls), and so I have changed the code
1489        so that no minimum is registered for a pattern that contains *ACCEPT.
1490    
1491    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1492        it was not handled as intended. [But see 16 below.]
1493    
1494    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1495        Sheri Pierce.
1496    
1497    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1498        the first byte in a match must be "a".
1499    
1500    11. Change 17 for 8.13 increased the recursion depth for patterns like
1501        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1502        pattern contains any instances of (*THEN). If it does not, the old
1503        optimizations are restored. It would be nice to do this on a per-group
1504        basis, but at the moment that is not feasible.
1505    
1506    12. In some environments, the output of pcretest -C is CRLF terminated. This
1507        broke RunTest's code that checks for the link size. A single white space
1508        character after the value is now allowed for.
1509    
1510    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1511        For "fr", it uses the Windows-specific input and output files.
1512    
1513    14. If (*THEN) appeared in a group that was called recursively or as a
1514        subroutine, it did not work as intended. [But see next item.]
1515    
1516    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1517        pattern fragments (but not containing any | characters). If A and B are
1518        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1519        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1520        D. In other words, Perl considers parentheses that do not contain any |
1521        characters to be part of a surrounding alternative, whereas PCRE was
1522        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1523        differently. PCRE now behaves in the same way as Perl, except in the case
1524        of subroutine/recursion calls such as (?1) which have in any case always
1525        been different (but PCRE had them first :-).
1526    
1527    16. Related to 15 above: Perl does not treat the | in a conditional group as
1528        creating alternatives. Such a group is treated in the same way as an
1529        ordinary group without any | characters when processing (*THEN). PCRE has
1530        been changed to match Perl's behaviour.
1531    
1532    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1533        RunGrepTest script failed.
1534    
1535    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1536        inevitable for groups that contain captures, but it can lead to a lot of
1537        stack use in large patterns. The old behaviour has been restored for atomic
1538        groups that do not contain any capturing parentheses.
1539    
1540    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1541        suppress the check for a minimum subject length at run time. (If it was
1542        given to pcre_exec() or pcre_dfa_exec() it did work.)
1543    
1544    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1545        fail to work when decoding hex characters in data strings in EBCDIC
1546        environments.
1547    
1548    21. It appears that in at least one Mac OS environment, the isxdigit() function
1549        is implemented as a macro that evaluates to its argument more than once,
1550        contravening the C 90 Standard (I haven't checked a later standard). There
1551        was an instance in pcretest which caused it to go wrong when processing
1552        \x{...} escapes in subject strings. The has been rewritten to avoid using
1553        things like p++ in the argument of isxdigit().
1554    
1555    
1556    Version 8.13 16-Aug-2011
1557    ------------------------
1558    
1559    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1560    
1561    2.  Two minor typos in pcre_internal.h have been fixed.
1562    
1563    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1564        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1565        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1566    
1567    4.  There were a number of related bugs in the code for matching backrefences
1568        caselessly in UTF-8 mode when codes for the characters concerned were
1569        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1570        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1571        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1572        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1573        2-byte code at the end of the subject (it thought there wasn't enough data
1574        left).
1575    
1576    5.  Comprehensive information about what went wrong is now returned by
1577        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1578        as the output vector has at least 2 elements. The offset of the start of
1579        the failing character and a reason code are placed in the vector.
1580    
1581    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1582        now returned is for the first byte of the failing character, instead of the
1583        last byte inspected. This is an incompatible change, but I hope it is small
1584        enough not to be a problem. It makes the returned offset consistent with
1585        pcre_exec() and pcre_dfa_exec().
1586    
1587    7.  pcretest now gives a text phrase as well as the error number when
1588        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1589        failure, the offset and reason code are output.
1590    
1591    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1592        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1593        back over a single character (\n). This seems wrong (because it treated the
1594        two characters as a single entity when going forwards), conflicts with the
1595        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1596        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1597        has been changed.
1598    
1599    9.  Some internal refactoring has changed the processing so that the handling
1600        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1601        time (the PCRE_DOTALL option was changed this way some time ago: version
1602        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1603        which was always a bit of a fudge. It also means that there is one less
1604        argument for the match() function, which reduces its stack requirements
1605        slightly. This change also fixes an incompatibility with Perl: the pattern
1606        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1607    
1608    10. More internal refactoring has drastically reduced the number of recursive
1609        calls to match() for possessively repeated groups such as (abc)++ when
1610        using pcre_exec().
1611    
1612    11. While implementing 10, a number of bugs in the handling of groups were
1613        discovered and fixed:
1614    
1615        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1616        (a|)*(?1) gave a compile-time internal error.
1617        ((a|)+)+  did not notice that the outer group could match an empty string.
1618        (^a|^)+   was not marked as anchored.
1619        (.*a|.*)+ was not marked as matching at start or after a newline.
1620    
1621    12. Yet more internal refactoring has removed another argument from the match()
1622        function. Special calls to this function are now indicated by setting a
1623        value in a variable in the "match data" data block.
1624    
1625    13. Be more explicit in pcre_study() instead of relying on "default" for
1626        opcodes that mean there is no starting character; this means that when new
1627        ones are added and accidentally left out of pcre_study(), testing should
1628        pick them up.
1629    
1630    14. The -s option of pcretest has been documented for ages as being an old
1631        synonym of -m (show memory usage). I have changed it to mean "force study
1632        for every regex", that is, assume /S for every regex. This is similar to -i
1633        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1634        using it. It makes it easier to run collections of tests with and without
1635        study enabled, and thereby test pcre_study() more easily. All the standard
1636        tests are now run with and without -s (but some patterns can be marked as
1637        "never study" - see 20 below).
1638    
1639    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1640        restoration of the capturing data to the outer values was not happening
1641        correctly.
1642    
1643    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1644        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1645        pattern had matched an empty string, and so incorrectly returned a no
1646        match.
1647    
1648    17. There was optimizing code for the last branch of non-capturing parentheses,
1649        and also for the obeyed branch of a conditional subexpression, which used
1650        tail recursion to cut down on stack usage. Unfortunately, now that there is
1651        the possibility of (*THEN) occurring in these branches, tail recursion is
1652        no longer possible because the return has to be checked for (*THEN). These
1653        two optimizations have therefore been removed. [But see 8.20/11 above.]
1654    
1655    18. If a pattern containing \R was studied, it was assumed that \R always
1656        matched two bytes, thus causing the minimum subject length to be
1657        incorrectly computed because \R can also match just one byte.
1658    
1659    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1660        was incorrectly computed.
1661    
1662    20. If /S is present twice on a test pattern in pcretest input, it now
1663        *disables* studying, thereby overriding the use of -s on the command line
1664        (see 14 above). This is necessary for one or two tests to keep the output
1665        identical in both cases.
1666    
1667    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1668        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1669    
1670    22. When an atomic group that contained a capturing parenthesis was
1671        successfully matched, but the branch in which it appeared failed, the
1672        capturing was not being forgotten if a higher numbered group was later
1673        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1674        group 1 to "a", when in fact it should be unset. This applied to multi-
1675        branched capturing and non-capturing groups, repeated or not, and also to
1676        positive assertions (capturing in negative assertions does not happen
1677        in PCRE) and also to nested atomic groups.
1678    
1679    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1680        subject after a captured substring, to make it easier to tell which of a
1681        number of identical substrings has been captured.
1682    
1683    24. The way atomic groups are processed by pcre_exec() has been changed so that
1684        if they are repeated, backtracking one repetition now resets captured
1685        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1686        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1687        "aaa". Previously, it would have been "a". As part of this code
1688        refactoring, the way recursive calls are handled has also been changed.
1689    
1690    25. If an assertion condition captured any substrings, they were not passed
1691        back unless some other capturing happened later. For example, if
1692        (?(?=(a))a) was matched against "a", no capturing was returned.
1693    
1694    26. When studying a pattern that contained subroutine calls or assertions,
1695        the code for finding the minimum length of a possible match was handling
1696        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1697        group 1 called group 2 while simultaneously a separate group 2 called group
1698        1). A stack overflow occurred in this case. I have fixed this by limiting
1699        the recursion depth to 10.
1700    
1701    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1702        Fortmann. This supports explicit test numbers on the command line, and has
1703        argument validation and error reporting.
1704    
1705    28. An instance of \X with an unlimited repeat could fail if at any point the
1706        first character it looked at was a mark character.
1707    
1708    29. Some minor code refactoring concerning Unicode properties and scripts
1709        should reduce the stack requirement of match() slightly.
1710    
1711    30. Added the '=' option to pcretest to check the setting of unused capturing
1712        slots at the end of the pattern, which are documented as being -1, but are
1713        not included in the return count.
1714    
1715    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1716        compiled something random. Now it gives a compile-time error (as does
1717        Perl).
1718    
1719    32. A *MARK encountered during the processing of a positive assertion is now
1720        recorded and passed back (compatible with Perl).
1721    
1722    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1723        had alternative anchored branches, the search for a second match in a line
1724        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1725        matched the line "0102" twice. The same bug affected patterns that started
1726        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1727        twice.
1728    
1729    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1730        does, and because of capturing effects, quantifying parenthesized
1731        assertions may at times be useful. Quantifiers are now allowed for
1732        parenthesized assertions.
1733    
1734    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1735    
1736    36. \g was being checked for fancy things in a character class, when it should
1737        just be a literal "g".
1738    
1739    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1740        appearance of a nested POSIX class supersedes an apparent external class.
1741        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1742        unescaped square brackets may also appear as part of class names. For
1743        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1744        more like Perl. (But see 8.20/1 above.)
1745    
1746    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1747        was because it thought it was \N{name}, which is not supported).
1748    
1749    39. Add minix to OS list not supporting the -S option in pcretest.
1750    
1751    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1752        cannot analyze patterns in sufficient detail to catch mutual recursions
1753        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1754        subgroup is called recursively as a subpattern for a second time at the
1755        same position in the subject string. In previous releases this might have
1756        been caught by the recursion limit, or it might have run out of stack.
1757    
1758    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1759        happen only once. PCRE was, however incorrectly giving a compile time error
1760        "recursive call could loop indefinitely" because it cannot analyze the
1761        pattern in sufficient detail. The compile time test no longer happens when
1762        PCRE is compiling a conditional subpattern, but actual runaway loops are
1763        now caught at runtime (see 40 above).
1764    
1765    42. It seems that Perl allows any characters other than a closing parenthesis
1766        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1767        has been changed to be the same.
1768    
1769    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1770        as not to get warnings when autogen.sh is called. Also changed
1771        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1772    
1773    44. To help people who use pcregrep to scan files containing exceedingly long
1774        lines, the following changes have been made:
1775    
1776        (a) The default value of the buffer size parameter has been increased from
1777            8K to 20K. (The actual buffer used is three times this size.)
1778    
1779        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1780            PCRE is built.
1781    
1782        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1783            to be set at run time.
1784    
1785        (d) Numerical values in pcregrep options can be followed by K or M, for
1786            example --buffer-size=50K.
1787    
1788        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1789            given and the return code is set to 2.
1790    
1791    45. Add a pointer to the latest mark to the callout data block.
1792    
1793    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1794        partial match of an empty string instead of no match. This was specific to
1795        the use of ".".
1796    
1797    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1798        complete match instead of a partial match. This bug was dependent on both
1799        the PCRE_UTF8 and PCRE_DOTALL options being set.
1800    
1801    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1802        starting byte set, because \b was not being ignored.
1803    
1804    
1805    Version 8.12 15-Jan-2011
1806    ------------------------
1807    
1808    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1809        checks for such things as part of the documentation building process.
1810    
1811    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1812        --match-limit and --recursion-limit options (added for 8.11). In
1813        particular, this made one of the standard tests fail. (The integer value
1814        went into the wrong half of a long int.)
1815    
1816    3.  If the --colour option was given to pcregrep with -v (invert match), it
1817        did strange things, either producing crazy output, or crashing. It should,
1818        of course, ignore a request for colour when reporting lines that do not
1819        match.
1820    
1821    4.  Another pcregrep bug caused similar problems if --colour was specified with
1822        -M (multiline) and the pattern match finished with a line ending.
1823    
1824    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1825        matched in multiline mode, the following line was shown as part of the
1826        match. This seems wrong, so I have changed it.
1827    
1828    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1829        the check for further matches in the same line (so they could be coloured)
1830        to overrun the end of the current line. If another match was found, it was
1831        incorrectly shown (and then shown again when found in the next line).
1832    
1833    7.  If pcregrep was compiled under Windows, there was a reference to the
1834        function pcregrep_exit() before it was defined. I am assuming this was
1835        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1836        reported by a user. I've moved the definition above the reference.
1837    
1838    
1839    Version 8.11 10-Dec-2010
1840    ------------------------
1841    
1842    1.  (*THEN) was not working properly if there were untried alternatives prior
1843        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1844        backtracked to try for "b" instead of moving to the next alternative branch
1845        at the same level (in this case, to look for "c"). The Perl documentation
1846        is clear that when (*THEN) is backtracked onto, it goes to the "next
1847        alternative in the innermost enclosing group".
1848    
1849    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1850        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1851        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1852        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1853        (*THEN).
1854    
1855    3.  If \s appeared in a character class, it removed the VT character from
1856        the class, even if it had been included by some previous item, for example
1857        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1858        of \s, but is part of the POSIX "space" class.)
1859    
1860    4.  A partial match never returns an empty string (because you can always
1861        match an empty string at the end of the subject); however the checking for
1862        an empty string was starting at the "start of match" point. This has been
1863        changed to the "earliest inspected character" point, because the returned
1864        data for a partial match starts at this character. This means that, for
1865        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1866        (previously it gave "no match").
1867    
1868    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1869        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1870        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1871        has an implication that the given string is incomplete (because a partial
1872        match is preferred over a full match). For this reason, these items now
1873        give a partial match in this situation. [Aside: previously, the one case
1874        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1875        match rather than a full match, which was wrong by the old rules, but is
1876        now correct.]
1877    
1878    6.  There was a bug in the handling of #-introduced comments, recognized when
1879        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1880        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1881        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1882        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1883        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1884        places in pcre_compile().
1885    
1886    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1887        comments when looking ahead for named forward references to subpatterns,
1888        the only newline sequence it recognized was NL. It now handles newlines
1889        according to the set newline convention.
1890    
1891    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1892        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1893        cater for a lack of strerror(). These oversights have been fixed.
1894    
1895    9.  Added --match-limit and --recursion-limit to pcregrep.
1896    
1897    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1898    
1899    11. When the -o option was used, pcregrep was setting a return code of 1, even
1900        when matches were found, and --line-buffered was not being honoured.
1901    
1902    12. Added an optional parentheses number to the -o and --only-matching options
1903        of pcregrep.
1904    
1905    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1906        can match an empty string. The code to do it in pcretest and pcredemo
1907        needed fixing:
1908    
1909        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1910            only one byte after an empty string match just before CRLF (this case
1911            just got forgotten; "any" and "anycrlf" were OK).
1912    
1913        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1914            mode when an empty string match preceded an ASCII character followed by
1915            a non-ASCII character. (The code for advancing by one character rather
1916            than one byte was nonsense.)
1917    
1918        (c) The pcredemo.c sample program did not have any code at all to handle
1919            the cases when CRLF is a valid newline sequence.
1920    
1921    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1922        as a starting offset was within the subject string. There is now a new
1923        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1924        negative or greater than the length of the string. In order to test this,
1925        pcretest is extended to allow the setting of negative starting offsets.
1926    
1927    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1928        starting offset points to the beginning of a UTF-8 character was
1929        unnecessarily clumsy. I tidied it up.
1930    
1931    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1932        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1933    
1934    17. Nobody had reported that the --include_dir option, which was added in
1935        release 7.7 should have been called --include-dir (hyphen, not underscore)
1936        for compatibility with GNU grep. I have changed it to --include-dir, but
1937        left --include_dir as an undocumented synonym, and the same for
1938        --exclude-dir, though that is not available in GNU grep, at least as of
1939        release 2.5.4.
1940    
1941    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1942        characters from a string of bytes) have been redefined so as not to use
1943        loops, in order to improve performance in some environments. At the same
1944        time, I abstracted some of the common code into auxiliary macros to save
1945        repetition (this should not affect the compiled code).
1946    
1947    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1948        compile-time error is now given if \c is not followed by an ASCII
1949        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1950        different, and any byte value is allowed.)
1951    
1952    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1953        START_OPTIMIZE option, which is now allowed at compile time - but just
1954        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1955        to pcregrep and other applications that have no direct access to PCRE
1956        options. The new /Y option in pcretest sets this option when calling
1957        pcre_compile().
1958    
1959    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1960        back references. Groups containing recursive back references were forced to
1961        be atomic by that change, but in the case of named groups, the amount of
1962        memory required was incorrectly computed, leading to "Failed: internal
1963        error: code overflow". This has been fixed.
1964    
1965    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1966        pcretest.c, to avoid build problems in some Borland environments.
1967    
1968    
1969    Version 8.10 25-Jun-2010
1970    ------------------------
1971    
1972    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1973        THEN.
1974    
1975    2.  (*ACCEPT) was not working when inside an atomic group.
1976    
1977    3.  Inside a character class, \B is treated as a literal by default, but
1978        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1979        causes the error). The code is unchanged, but I tidied the documentation.
1980    
1981    4.  Inside a character class, PCRE always treated \R and \X as literals,
1982        whereas Perl faults them if its -w option is set. I have changed PCRE so
1983        that it faults them when PCRE_EXTRA is set.
1984    
1985    5.  Added support for \N, which always matches any character other than
1986        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1987    
1988    6.  When compiling pcregrep with newer versions of gcc which may have
1989        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1990        declared with attribute warn_unused_result" were given. Just casting the
1991        result to (void) does not stop the warnings; a more elaborate fudge is
1992        needed. I've used a macro to implement this.
1993    
1994    7.  Minor change to pcretest.c to avoid a compiler warning.
1995    
1996    8.  Added four artifical Unicode properties to help with an option to make
1997        \s etc use properties (see next item). The new properties are: Xan
1998        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1999    
2000    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2001        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2002        this option. Modified pcretest to add /W to test this facility. Added
2003        REG_UCP to make it available via the POSIX interface.
2004    
2005    10. Added --line-buffered to pcregrep.
2006    
2007    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2008        studied, and the match started with a letter with a code point greater than
2009        127 whose first byte was different to the first byte of the other case of
2010        the letter, the other case of this starting letter was not recognized
2011        (#976).
2012    
2013    12. If a pattern that was studied started with a repeated Unicode property
2014        test, for example, \p{Nd}+, there was the theoretical possibility of
2015        setting up an incorrect bitmap of starting bytes, but fortunately it could
2016        not have actually happened in practice until change 8 above was made (it
2017        added property types that matched character-matching opcodes).
2018    
2019    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2020        possible starting bytes for non-anchored patterns.
2021    
2022    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2023        \R, and also a number of cases that involve Unicode properties, both
2024        explicit and implicit when PCRE_UCP is set.
2025    
2026    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2027        input, it could crash or give wrong results if characters with values
2028        greater than 0xc0 were present in the subject string. (Detail: it assumed
2029        UTF-8 input when processing these items.)
2030    
2031    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2032        size_t is 64-bit (#991).
2033    
2034    17. Added a check for running out of memory when PCRE is compiled with
2035        --disable-stack-for-recursion (#990).
2036    
2037    18. If the last data line in a file for pcretest does not have a newline on
2038        the end, a newline was missing in the output.
2039    
2040    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2041        less than 128) in its various bitmaps. However, there is a facility for
2042        generating tables according to the current locale when PCRE is compiled. It
2043        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2044        characters, are recognized by isspace() and therefore were getting set in
2045        these tables, and indeed these tables seem to approximate to ISO 8859. This
2046        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2047        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2048        which of course cannot start UTF-8 characters. I have changed the code so
2049        that only real ASCII characters (less than 128) and the correct starting
2050        bytes for UTF-8 encodings are set for characters greater than 127 when in
2051        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2052        altogether.)
2053    
2054    20. Added the /T option to pcretest so as to be able to run tests with non-
2055        standard character tables, thus making it possible to include the tests
2056        used for 19 above in the standard set of tests.
2057    
2058    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2059        reference to a subpattern the other side of a comment that contains an
2060        opening parenthesis caused either an internal compiling error, or a
2061        reference to the wrong subpattern.
2062    
2063    
2064    Version 8.02 19-Mar-2010
2065    ------------------------
2066    
2067    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2068    
2069    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2070        configured.
2071    
2072    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2073        original author of that file, following a query about its status.
2074    
2075    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2076        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2077    
2078    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2079        quantifier applied to a forward-referencing subroutine call, could compile
2080        incorrect code or give the error "internal error: previously-checked
2081        referenced subpattern not found".
2082    
2083    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2084        variables to point to external functions. For these systems, therefore,
2085        pcre_malloc etc. are now initialized to local functions that call the
2086        relevant global functions.
2087    
2088    7.  There were two entries missing in the vectors called coptable and poptable
2089        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2090        I've fixed the data, and added a kludgy way of testing at compile time that
2091        the lengths are correct (equal to the number of opcodes).
2092    
2093    8.  Following on from 7, I added a similar kludge to check the length of the
2094        eint vector in pcreposix.c.
2095    
2096    9.  Error texts for pcre_compile() are held as one long string to avoid too
2097        much relocation at load time. To find a text, the string is searched,
2098        counting zeros. There was no check for running off the end of the string,
2099        which could happen if a new error number was added without updating the
2100        string.
2101    
2102    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2103    
2104    11. \K was not working if it appeared in an atomic group or in a group that
2105        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2106        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2107        the assertion is positive, but not if it is negative.
2108    
2109    12. Change 11 fortuitously reduced the size of the stack frame used in the
2110        "match()" function of pcre_exec.c by one pointer. Forthcoming
2111        implementation of support for (*MARK) will need an extra pointer on the
2112        stack; I have reserved it now, so that the stack frame size does not
2113        decrease.
2114    
2115    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2116        item in branch that calls a recursion is a subroutine call - as in the
2117        second branch in the above example - was incorrectly given the compile-
2118        time error "recursive call could loop indefinitely" because pcre_compile()
2119        was not correctly checking the subroutine for matching a non-empty string.
2120    
2121    14. The checks for overrunning compiling workspace could trigger after an
2122        overrun had occurred. This is a "should never occur" error, but it can be
2123        triggered by pathological patterns such as hundreds of nested parentheses.
2124        The checks now trigger 100 bytes before the end of the workspace.
2125    
2126    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2127    
2128    
2129    Version 8.01 19-Jan-2010
2130    ------------------------
2131    
2132    1.  If a pattern contained a conditional subpattern with only one branch (in
2133        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2134        computed the wrong minimum data length (which is of course zero for such
2135        subpatterns). This could cause incorrect "no match" results.
2136    
2137    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2138        the pattern is reset in the first branch, pcre_compile() failed with
2139        "internal error: code overflow at offset...". This happened only when
2140        the reset was to the original external option setting. (An optimization
2141        abstracts leading options settings into an external setting, which was the
2142        cause of this.)
2143    
2144    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2145        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2146        assertion pattern did not match (meaning that the assertion was true), it
2147        was incorrectly treated as false if the SKIP had been reached during the
2148        matching. This also applied to assertions used as conditions.
2149    
2150    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2151        assertion subpattern, including such a pattern used as a condition,
2152        unpredictable results occurred, instead of the error return
2153        PCRE_ERROR_DFA_UITEM.
2154    
2155    5.  The C++ GlobalReplace function was not working like Perl for the special
2156        situation when an empty string is matched. It now does the fancy magic
2157        stuff that is necessary.
2158    
2159    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2160        removed. (These were left over from very, very early versions of PCRE.)
2161    
2162    7.  Some cosmetic changes to the code to make life easier when compiling it
2163        as part of something else:
2164    
2165        (a) Change DEBUG to PCRE_DEBUG.
2166    
2167        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2168            called "current" as "current_branch", to prevent a collision with the
2169            Linux macro when compiled as a kernel module.
2170    
2171        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2172            prevent a collision with the Linux macro when compiled as a kernel
2173            module.
2174    
2175    8.  In pcre_compile() there are some checks for integer overflows that used to
2176        cast potentially large values to (double). This has been changed to that
2177        when building, a check for int64_t is made, and if it is found, it is used
2178        instead, thus avoiding the use of floating point arithmetic. (There is no
2179        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2180        double.
2181    
2182    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2183        2005 (difference between two addresses compared to an unsigned value).
2184    
2185    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2186        custom one, because of the following reported problem in Windows:
2187    
2188          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2189              under Win32.
2190          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2191              therefore missing the function definition.
2192          - The compiler thus generates a "C" signature for the test function.
2193          - The linker fails to find the "C" function.
2194          - PCRE fails to configure if asked to do so against libbz2.
2195    
2196    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2197        messages were output:
2198    
2199          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2200          rerunning libtoolize, to keep the correct libtool macros in-tree.
2201          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2202    
2203        I have done both of these things.
2204    
2205    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2206        most of the time, it *can* run out if it is given a pattern that contains a
2207        runaway infinite recursion. I updated the discussion in the pcrestack man
2208        page.
2209    
2210    13. Now that we have gone to the x.xx style of version numbers, the minor
2211        version may start with zero. Using 08 or 09 is a bad idea because users
2212        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2213        interpreted as invalid octal numbers. I've updated the previous comment in
2214        configure.ac, and also added a check that gives an error if 08 or 09 are
2215        used.
2216    
2217    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2218        causing partial matching to fail when the end of the subject matched \W
2219        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2220    
2221    15. There were some discrepancies between the declarations in pcre_internal.h
2222        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2223        their definitions. The declarations used "const uschar *" and the
2224        definitions used USPTR. Even though USPTR is normally defined as "const
2225        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2226        reported that: "This difference in casting confuses some C++ compilers, for
2227        example, SunCC recognizes above declarations as different functions and
2228        generates broken code for hbpcre." I have changed the declarations to use
2229        USPTR.
2230    
2231    16. GNU libtool is named differently on some systems. The autogen.sh script now
2232        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2233        (FreeBSD).
2234    
2235    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2236        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2237        comment: "Figure out how to create a longlong from a string: strtoll and
2238        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2239        instance, but it only takes 2 args instead of 3!"
2240    
2241    18. A subtle bug concerned with back references has been fixed by a change of
2242        specification, with a corresponding code fix. A pattern such as
2243        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2244        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2245        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2246        same bug. Such groups have to be quantified to be useful, or contained
2247        inside another quantified group. (If there's no repetition, the reference
2248        can never match.) The problem arises because, having left the group and
2249        moved on to the rest of the pattern, a later failure that backtracks into
2250        the group uses the captured value from the final iteration of the group
2251        rather than the correct earlier one. I have fixed this in PCRE by forcing
2252        any group that contains a reference to itself to be an atomic group; that
2253        is, there cannot be any backtracking into it once it has completed. This is
2254        similar to recursive and subroutine calls.
2255    
2256    
2257    Version 8.00 19-Oct-09
2258    ----------------------
2259    
2260    1.  The table for translating pcre_compile() error codes into POSIX error codes
2261        was out-of-date, and there was no check on the pcre_compile() error code
2262        being within the table. This could lead to an OK return being given in
2263        error.
2264    
2265    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2266        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2267        in a Windows environment.
2268    
2269    3.  The pcregrep --count option prints the count for each file even when it is
2270        zero, as does GNU grep. However, pcregrep was also printing all files when
2271        --files-with-matches was added. Now, when both options are given, it prints
2272        counts only for those files that have at least one match. (GNU grep just
2273        prints the file name in this circumstance, but including the count seems
2274        more useful - otherwise, why use --count?) Also ensured that the
2275        combination -clh just lists non-zero counts, with no names.
2276    
2277    4.  The long form of the pcregrep -F option was incorrectly implemented as
2278        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2279        but it seems right to fix it, and I didn't think it was worth preserving
2280        the old behaviour.
2281    
2282    5.  The command line items --regex=pattern and --regexp=pattern were not
2283        recognized by pcregrep, which required --regex pattern or --regexp pattern
2284        (with a space rather than an '='). The man page documented the '=' forms,
2285        which are compatible with GNU grep; these now work.
2286    
2287    6.  No libpcreposix.pc file was created for pkg-config; there was just
2288        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2289    
2290    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2291        when UCP support is not needed, by modifying the Python script that
2292        generates it from Unicode data files. This should not matter if the module
2293        is correctly used as a library, but I received one complaint about 50K of
2294        unwanted data. My guess is that the person linked everything into his
2295        program rather than using a library. Anyway, it does no harm.
2296    
2297    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2298        was a minimum greater than 1 for a wide character in a possessive
2299        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2300        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2301        character. Chaos in the form of incorrect output or a compiling loop could
2302        result.
2303    
2304    9.  The restrictions on what a pattern can contain when partial matching is
2305        requested for pcre_exec() have been removed. All patterns can now be
2306        partially matched by this function. In addition, if there are at least two
2307        slots in the offset vector, the offset of the earliest inspected character
2308        for the match and the offset of the end of the subject are set in them when
2309        PCRE_ERROR_PARTIAL is returned.
2310    
2311    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2312        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2313        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2314        and may be more useful for multi-segment matching.
2315    
2316    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2317        used to be given if ever the end of the subject was reached; now it is
2318        given only if matching could not proceed because another character was
2319        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2320        string "Z", which now yields "no match" instead of "partial match". In the
2321        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2322        final character ended with (*FAIL).
2323    
2324    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2325        if the pattern had a "must contain" character that was already found in the
2326        earlier partial match, unless partial matching was again requested. For
2327        example, with the pattern /dog.(body)?/, the "must contain" character is
2328        "g". If the first part-match was for the string "dog", restarting with
2329        "sbody" failed. This bug has been fixed.
2330    
2331    13. The string returned by pcre_dfa_exec() after a partial match has been
2332        changed so that it starts at the first inspected character rather than the
2333        first character of the match. This makes a difference only if the pattern
2334        starts with a lookbehind assertion or \b or \B (\K is not supported by
2335        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2336        matching functions compatible, and I think it's the right thing to do.
2337    
2338    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2339        so that the demonstration program is easily available in environments where
2340        PCRE has not been installed from source.
2341    
2342    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2343        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2344        library.
2345    
2346    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2347        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2348        is not the first non-POSIX option to be added. Clearly some people find
2349        these options useful.
2350    
2351    17. If a caller to the POSIX matching function regexec() passes a non-zero
2352        value for nmatch with a NULL value for pmatch, the value of
2353        nmatch is forced to zero.
2354    
2355    18. RunGrepTest did not have a test for the availability of the -u option of
2356        the diff command, as RunTest does. It now checks in the same way as
2357        RunTest, and also checks for the -b option.
2358    
2359    19. If an odd number of negated classes containing just a single character
2360        interposed, within parentheses, between a forward reference to a named
2361        subpattern and the definition of the subpattern, compilation crashed with
2362        an internal error, complaining that it could not find the referenced
2363        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2364        [The bug was that it was starting one character too far in when skipping
2365        over the character class, thus treating the ] as data rather than
2366        terminating the class. This meant it could skip too much.]
2367    
2368    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2369        /g option in pcretest when the pattern contains \K, which makes it possible
2370        to have an empty string match not at the start, even when the pattern is
2371        anchored. Updated pcretest and pcredemo to use this option.
2372    
2373    21. If the maximum number of capturing subpatterns in a recursion was greater
2374        than the maximum at the outer level, the higher number was returned, but
2375        with unset values at the outer level. The correct (outer level) value is
2376        now given.
2377    
2378    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2379        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2380        make it do so. The string so far is captured, making this feature
2381        compatible with Perl.
2382    
2383    23. The tests have been re-organized, adding tests 11 and 12, to make it
2384        possible to check the Perl 5.10 features against Perl 5.10.
2385    
2386    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2387        pattern matches a fixed length string. PCRE did not allow this; now it
2388        does. Neither allows recursion.
2389    
2390    25. I finally figured out how to implement a request to provide the minimum
2391        length of subject string that was needed in order to match a given pattern.
2392        (It was back references and recursion that I had previously got hung up
2393        on.) This code has now been added to pcre_study(); it finds a lower bound
2394        to the length of subject needed. It is not necessarily the greatest lower
2395        bound, but using it to avoid searching strings that are too short does give
2396        some useful speed-ups. The value is available to calling programs via
2397        pcre_fullinfo().
2398    
2399    26. While implementing 25, I discovered to my embarrassment that pcretest had
2400        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2401        study optimizations had never been tested with that matching function.
2402        Oops. What is worse, even when it was passed study data, there was a bug in
2403        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2404        were also very few tests of studied patterns with pcre_dfa_exec().
2405    
2406    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2407        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2408        on the other side of the coin, they are no longer allowed to have different
2409        names, because these cannot be distinguished in PCRE, and this has caused
2410        confusion. (This is a difference from Perl.)
2411    
2412    28. When duplicate subpattern names are present (necessarily with different
2413        numbers, as required by 27 above), and a test is made by name in a
2414        conditional pattern, either for a subpattern having been matched, or for
2415        recursion in such a pattern, all the associated numbered subpatterns are
2416        tested, and the overall condition is true if the condition is true for any
2417        one of them. This is the way Perl works, and is also more like the way
2418        testing by number works.
2419    
2420    
2421    Version 7.9 11-Apr-09
2422    ---------------------
2423    
2424    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2425        (pcretest), all targets were linked against these libraries. This included
2426        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2427        libraries. This caused unwanted dependencies to be created. This problem
2428        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2429        pcretest is linked with readline.
2430    
2431    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2432        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2433        moved outside it again, because FALSE and TRUE are already defined in AIX,
2434        but BOOL is not.
2435    
2436    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2437        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2438    
2439    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2440        hyphens) following file names and line numbers when outputting matching
2441        lines. This is not true; no spaces are inserted. I have also clarified the
2442        wording for the --colour (or --color) option.
2443    
2444    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2445        was not coloured; this is different to GNU grep, so I have changed it to be
2446        the same.
2447    
2448    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2449        each matching line was coloured. Now it goes on to look for further matches
2450        of any of the test patterns, which is the same behaviour as GNU grep.
2451    
2452    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2453        doesn't make sense to accept an empty string match in pcregrep, so I have
2454        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2455        seems to be how GNU grep behaves. [But see later change 40 for release
2456        8.33.]
2457    
2458    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2459        start or after a newline", because the conditional assertion was not being
2460        correctly handled. The rule now is that both the assertion and what follows
2461        in the first alternative must satisfy the test.
2462    
2463    9.  If auto-callout was enabled in a pattern with a conditional group whose
2464        condition was an assertion, PCRE could crash during matching, both with
2465        pcre_exec() and pcre_dfa_exec().
2466    
2467    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2468        used for matching.
2469    
2470    11. Unicode property support in character classes was not working for
2471        characters (bytes) greater than 127 when not in UTF-8 mode.
2472    
2473    12. Added the -M command line option to pcretest.
2474    
2475    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2476    
2477    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2478    
2479    16. Added comments and documentation about mis-use of no_arg in the C++
2480        wrapper.
2481    
2482    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2483        from Martin Jerabek that uses macro names for all relevant character and
2484        string constants.
2485    
2486    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2487        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2488        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2489        these, but not everybody uses configure.
2490    
2491    19. A conditional group that had only one branch was not being correctly
2492        recognized as an item that could match an empty string. This meant that an
2493        enclosing group might also not be so recognized, causing infinite looping
2494        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2495        with the subject "ab", where knowledge that the repeated group can match
2496        nothing is needed in order to break the loop.
2497    
2498    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2499        exec(), but without supplying a callout function, matching went wrong.
2500    
2501    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2502        leak if the size of the offset vector was greater than 30. When the vector
2503        is smaller, the saved offsets during recursion go onto a local stack
2504        vector, but for larger vectors malloc() is used. It was failing to free
2505        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2506        error, in fact).
2507    
2508    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2509        heapframe that is used only when UTF-8 support is enabled. This caused no
2510        problem, but was untidy.
2511    
2512    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2513        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2514        included within another project.
2515    
2516    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2517        slightly modified by me:
2518    
2519          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2520              not building pcregrep.
2521    
2522          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2523              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2524    
2525    25. Forward references, both numeric and by name, in patterns that made use of
2526        duplicate group numbers, could behave incorrectly or give incorrect errors,
2527        because when scanning forward to find the reference group, PCRE was not
2528        taking into account the duplicate group numbers. A pattern such as
2529        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2530    
2531    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2532        the feature of a custom pointer more persuasive (as requested by a user).
2533    
2534    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2535        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2536        pre-defined.
2537    
2538    28. Added support for (*UTF8) at the start of a pattern.
2539    
2540    29. Arrange for flags added by the "release type" setting in CMake to be shown
2541        in the configuration summary.
2542    
2543    
2544    Version 7.8 05-Sep-08
2545    ---------------------
2546    
2547    1.  Replaced UCP searching code with optimized version as implemented for Ad
2548        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2549        stage table and inline lookup instead of a function, giving speed ups of 2
2550        to 5 times on some simple patterns that I tested. Permission was given to
2551        distribute the MultiStage2.py script that generates the tables (it's not in
2552        the tarball, but is in the Subversion repository).
2553    
2554    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2555        scripts.
2556    
2557    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2558        a group with a zero qualifier. The result of the study could be incorrect,
2559        or the function might crash, depending on the pattern.
2560    
2561    4.  Caseless matching was not working for non-ASCII characters in back
2562        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2563        It now works when Unicode Property Support is available.
2564    
2565    5.  In pcretest, an escape such as \x{de} in the data was always generating
2566        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2567        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2568        truncation.
2569    
2570    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2571    
2572    7.  Added two (int) casts to pcregrep when printing the difference of two
2573        pointers, in case they are 64-bit values.
2574    
2575    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2576        test 2 if it fails.
2577    
2578    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2579        and a #define of that name to empty if it is not externally set. This is to
2580        allow users of MSVC to set it if necessary.
2581    
2582    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2583        the convenience functions in the pcre_get.c source file.
2584    
2585    11. An option change at the start of a pattern that had top-level alternatives
2586        could cause overwriting and/or a crash. This command provoked a crash in
2587        some environments:
2588    
2589          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2590    
2591        This potential security problem was recorded as CVE-2008-2371.
2592    
2593    12. For a pattern where the match had to start at the beginning or immediately
2594        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2595        pcre_dfa_exec() could read past the end of the passed subject if there was
2596        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2597        pcretest so that it places the subject at the end of its malloc-ed buffer.
2598    
2599    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2600        exec() might read past the end of the data buffer in UTF-8 mode.
2601    
2602    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2603        the data contained the byte 0x85 as part of a UTF-8 character within its
2604        first line. This applied both to normal and DFA matching.
2605    
2606    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2607        /^[^d]*?$/8 failed to match "abc".
2608    
2609    16. Added a missing copyright notice to pcrecpp_internal.h.
2610    
2611    17. Make it more clear in the documentation that values returned from
2612        pcre_exec() in ovector are byte offsets, not character counts.
2613    
2614    18. Tidied a few places to stop certain compilers from issuing warnings.
2615    
2616    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2617        supplied by Stefan Weber. I made a further small update for 7.8 because
2618        there is a change of source arrangements: the pcre_searchfuncs.c module is
2619        replaced by pcre_ucd.c.
2620    
2621    
2622    Version 7.7 07-May-08
2623    ---------------------
2624    
2625    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2626        a string to a long long, pretend we don't even have a long long." This is
2627        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2628    
2629    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2630        pre-7.6 versions, which defined a global no_arg variable instead of putting
2631        it in the RE class. (See also #8 below.)
2632    
2633    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2634        Lopes.
2635    
2636    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2637    
2638        (1) The include/exclude patterns were being applied to the whole pathnames
2639            of files, instead of just to the final components.
2640    
2641        (2) If there was more than one level of directory, the subdirectories were
2642            skipped unless they satisfied the include/exclude conditions. This is
2643            inconsistent with GNU grep (and could even be seen as contrary to the
2644            pcregrep specification - which I improved to make it absolutely clear).
2645            The action now is always to scan all levels of directory, and just
2646            apply the include/exclude patterns to regular files.
2647    
2648    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2649        --exclude_dir in the tests to avoid scanning .svn directories.
2650    
2651    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2652        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2653        doesn't support NULs in patterns.
2654    
2655    7.  Added some missing "const"s to declarations of static tables in
2656        pcre_compile.c and pcre_dfa_exec.c.
2657    
2658    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2659        caused by fix #2  above. (Subsequently also a second patch to fix the
2660        first patch. And a third patch - this was a messy problem.)
2661    
2662    9.  Applied Craig's patch to remove the use of push_back().
2663    
2664    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2665        matching function regexec().
2666    
2667    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2668        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2669        references. PCRE supports relative numbers with this syntax (I don't think
2670        Oniguruma does).
2671    
2672    12. Previously, a group with a zero repeat such as (...){0} was completely
2673        omitted from the compiled regex. However, this means that if the group
2674        was called as a subroutine from elsewhere in the pattern, things went wrong
2675        (an internal error was given). Such groups are now left in the compiled
2676        pattern, with a new opcode that causes them to be skipped at execution
2677        time.
2678    
2679    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2680        to the way PCRE behaves:
2681    
2682        (a) A lone ] character is dis-allowed (Perl treats it as data).
2683    
2684        (b) A back reference to an unmatched subpattern matches an empty string
2685            (Perl fails the current match path).
2686    
2687        (c) A data ] in a character class must be notated as \] because if the
2688            first data character in a class is ], it defines an empty class. (In
2689            Perl it is not possible to have an empty class.) The empty class []
2690            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2691            The negative empty class [^] matches any one character, independently
2692            of the DOTALL setting.
2693    
2694    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2695        non-existent subpattern following a character class starting with ']' and
2696        containing () gave an internal compiling error instead of "reference to
2697        non-existent subpattern". Fortunately, when the pattern did exist, the
2698        compiled code was correct. (When scanning forwards to check for the
2699        existence of the subpattern, it was treating the data ']' as terminating
2700        the class, so got the count wrong. When actually compiling, the reference
2701        was subsequently set up correctly.)
2702    
2703    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2704        it was being rejected as not supported by pcre_dfa_exec(), even though
2705        other assertions are supported. I have made pcre_dfa_exec() support
2706        (*FAIL).
2707    
2708    16. The implementation of 13c above involved the invention of a new opcode,
2709        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2710        cannot be changed at match time, I realized I could make a small
2711        improvement to matching performance by compiling OP_ALLANY instead of
2712        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2713        on the OP_ANY path.
2714    
2715    17. Compiling pcretest on Windows with readline support failed without the
2716        following two fixes: (1) Make the unistd.h include conditional on
2717        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2718    
2719    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2720        ncurses library to be included for pcretest when ReadLine support is
2721        requested, but also to allow for it to be overridden. This patch came from
2722        Daniel Bergström.
2723    
2724    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2725        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2726        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2727        spotting this.
2728    
2729    
2730    Version 7.6 28-Jan-08
2731    ---------------------
2732    
2733    1.  A character class containing a very large number of characters with
2734        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2735        overflow.
2736    
2737    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2738        HAVE_LONG_LONG is not defined.
2739    
2740    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2741        bring them up to date and include new features. This patch includes:
2742    
2743        - Fixed PH's badly added libz and libbz2 support.
2744        - Fixed a problem with static linking.
2745        - Added pcredemo. [But later removed - see 7 below.]
2746        - Fixed dftables problem and added an option.
2747        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2748            HAVE_LONG_LONG.
2749        - Added readline support for pcretest.
2750        - Added an listing of the option settings after cmake has run.
2751    
2752    4.  A user submitted a patch to Makefile that makes it easy to create
2753        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2754        Makefile.am that cause it to include this special target, without
2755        affecting anything else. Note that the same mingw target plus all
2756        the other distribution libraries and programs are now supported
2757        when configuring with CMake (see 6 below) instead of with
2758        Configure/Make.
2759    
2760    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2761        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2762        exported in the Windows port". It has not yet been confirmed that the patch
2763        solves the problem, but it does no harm.
2764    
2765    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2766        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2767        with CMake, and also correct the comment about stack recursion.
2768    
2769    7.  Remove the automatic building of pcredemo from the ./configure system and
2770        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2771        of a program that users should build themselves after PCRE is installed, so
2772        building it automatically is not really right. What is more, it gave
2773        trouble in some build environments.
2774    
2775    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2776    
2777    
2778    Version 7.5 10-Jan-08
2779    ---------------------
2780    
2781    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2782        values in parens when parsing an RE using the C++ wrapper."
2783    
2784    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2785        Characters greater than 255 were excluded from the class instead of being
2786        included.
2787    
2788    3.  The same bug as (2) above applied to negated POSIX classes such as
2789        [:^space:].
2790    
2791    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2792        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2793        I have changed it.
2794    
2795    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2796        first named subpattern) and a construct such as (?&a) would reference the
2797        first named subpattern whose name started with "a" (in other words, the
2798        length check was missing). Both these problems are fixed. "Subpattern name
2799        expected" is now given for (?&) (a zero-length name), and this patch also
2800        makes it give the same error for \k'' (previously it complained that that
2801        was a reference to a non-existent subpattern).
2802    
2803    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2804        this is right because (?- can be followed by option settings as well as by
2805        digits. I have, however, made the messages clearer.
2806    
2807    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2808        than the number used in the conditional) now cause a compile-time error.
2809        This is actually not compatible with Perl, which accepts such patterns, but
2810        treats the conditional as always being FALSE (as PCRE used to), but it
2811        seems to me that giving a diagnostic is better.
2812    
2813    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2814        and messages.
2815    
2816    9.  Fix two occurrences of "backslash" in comments that should have been
2817        "backspace".
2818    
2819    10. Remove two redundant lines of code that can never be obeyed (their function
2820        was moved elsewhere).
2821    
2822    11. The program that makes PCRE's Unicode character property table had a bug
2823        which caused it to generate incorrect table entries for sequences of
2824        characters that have the same character type, but are in different scripts.
2825        It amalgamated them into a single range, with the script of the first of
2826        them. In other words, some characters were in the wrong script. There were
2827        thirteen such cases, affecting characters in the following ranges:
2828    
2829          U+002b0 - U+002c1
2830          U+0060c - U+0060d
2831          U+0061e - U+00612
2832          U+0064b - U+0065e
2833          U+0074d - U+0076d
2834          U+01800 - U+01805
2835          U+01d00 - U+01d77
2836          U+01d9b - U+01dbf
2837          U+0200b - U+0200f
2838          U+030fc - U+030fe
2839          U+03260 - U+0327f
2840          U+0fb46 - U+0fbb1
2841          U+10450 - U+1049d
2842    
2843    12. The -o option (show only the matching part of a line) for pcregrep was not
2844        compatible with GNU grep in that, if there was more than one match in a
2845        line, it showed only the first of them. It now behaves in the same way as
2846        GNU grep.
2847    
2848    13. If the -o and -v options were combined for pcregrep, it printed a blank
2849        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2850        does the same. The return code can be used to tell if there were any
2851        non-matching lines.
2852    
2853    14. Added --file-offsets and --line-offsets to pcregrep.
2854    
2855    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2856        infinitely looping recursion. The bug was that positive lookaheads were not
2857        being skipped when checking for a possible empty match (negative lookaheads
2858        and both kinds of lookbehind were skipped).
2859    
2860    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2861        inclusion of <windows.h> to before rather than after the definition of
2862        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2863    
2864    17. Specifying a possessive quantifier with a specific limit for a Unicode
2865        character property caused pcre_compile() to compile bad code, which led at
2866        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2867        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2868        caused the error; without that there was no problem.
2869    
2870    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2871    
2872    19. Added --enable-pcretest-libreadline.
2873    
2874    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2875        RE::GlobalReplace(). As a result, the number of replacements returned was
2876        double what it should be. I removed one of the increments, but Craig sent a
2877        later patch that removed the other one (the right fix) and added unit tests
2878        that check the return values (which was not done before).
2879    
2880    21. Several CMake things:
2881    
2882        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2883            the names libpcre and libpcreposix, not just pcre and pcreposix.
2884    
2885        (2) The above change means that pcretest and pcregrep are now correctly
2886            linked with the newly-built libraries, not previously installed ones.
2887    
2888        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2889    
2890    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2891        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2892        UTF-8 newline character). The key issue is that the pattern starts .*;
2893        this means that the match must be either at the beginning, or after a
2894        newline. The bug was in the code for advancing after a failed match and
2895        checking that the new position followed a newline. It was not taking
2896        account of UTF-8 characters correctly.
2897    
2898    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2899        character classes. PCRE was not treating the sequence [:...:] as a
2900        character class unless the ... were all letters. Perl, however, seems to
2901        allow any characters between [: and :], though of course it rejects as
2902        unknown any "names" that contain non-letters, because all the known class
2903        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2904        for example, whereas PCRE did not - it did not recognize a POSIX character
2905        class. This seemed a bit dangerous, so the code has been changed to be
2906        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2907        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2908        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2909        Perl does, and where it didn't before.
2910    
2911    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2912        Windows environments %n is disabled by default.
2913    
2914    
2915    Version 7.4 21-Sep-07
2916    ---------------------
2917    
2918    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2919        means that a class such as [\s] counted as "explicit reference to CR or
2920        LF". That isn't really right - the whole point of the change was to try to
2921        help when there was an actual mention of one of the two characters. So now
2922        the change happens only if \r or \n (or a literal CR or LF) character is
2923        encountered.
2924    
2925    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2926        of both had grown to the point where there were only 3 bits left.
2927        Fortunately, there was spare space in the data structure, and so I have
2928        moved the internal flags into a new 16-bit field to free up more option
2929        bits.
2930    
2931    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2932        but did not set the internal JCHANGED flag - either of these is enough to
2933        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2934        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2935        start sets both bits.
2936    
2937    4.  Added options (at build time, compile time, exec time) to change \R from
2938        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2939    
2940    5.  doc/pcresyntax.html was missing from the distribution.
2941    
2942    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2943        compatibility, even though it is no longer used.
2944    
2945    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2946        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2947        windows.h file is present (where different names are used). [This was
2948        reversed later after testing - see 16 below.]
2949    
2950    8.  Changed all #include <config.h> to #include "config.h". There were also
2951        some further <pcre.h> cases that I changed to "pcre.h".
2952    
2953    9.  When pcregrep was used with the --colour option, it missed the line ending
2954        sequence off the lines that it output.
2955    
2956    10. It was pointed out to me that arrays of string pointers cause lots of
2957        relocations when a shared library is dynamically loaded. A technique of
2958        using a single long string with a table of offsets can drastically reduce
2959        these. I have refactored PCRE in four places to do this. The result is
2960        dramatic:
2961    
2962          Originally:                          290
2963          After changing UCP table:            187
2964          After changing error message table:   43
2965          After changing table of "verbs"       36
2966          After changing table of Posix names   22
2967    
2968        Thanks to the folks working on Gregex for glib for this insight.
2969    
2970    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2971        unicode-properties was also set.
2972    
2973    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2974    
2975    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2976        checked only for CRLF.
2977    
2978    14. Added casts to pcretest.c to avoid compiler warnings.
2979    
2980    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2981    
2982    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2983        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2984        entirely. This removes changes made in 7 above.
2985    
2986    17. The CMake files have been updated, and there is now more information about
2987        building with CMake in the NON-UNIX-USE document.
2988    
2989    
2990    Version 7.3 28-Aug-07
2991  ---------------------  ---------------------
2992    
2993   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 29  Version 7.3 05-Jul-07 Line 3015  Version 7.3 05-Jul-07
3015   4. Debugging: adjusted the names of opcodes for different kinds of parentheses   4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3016      in debug output.      in debug output.
3017    
3018   5. Arrange to use "%I64d" instead of "%lld" for long printing in the pcrecpp   5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3019      unittest when running under MinGW.      long printing in the pcrecpp unittest when running under MinGW.
3020    
3021   6. ESC_K was left out of the EBCDIC table.   6. ESC_K was left out of the EBCDIC table.
3022    
# Line 53  Version 7.3 05-Jul-07 Line 3039  Version 7.3 05-Jul-07
3039      pcreposix and pcrecpp parts of the library. These were overlooked when this      pcreposix and pcrecpp parts of the library. These were overlooked when this
3040      problem was solved for the main library.      problem was solved for the main library.
3041    
3042     9. There were some crude static tests to avoid integer overflow when computing
3043        the size of patterns that contain repeated groups with explicit upper
3044        limits. As the maximum quantifier is 65535, the maximum group length was
3045        set at 30,000 so that the product of these two numbers did not overflow a
3046        32-bit integer. However, it turns out that people want to use groups that
3047        are longer than 30,000 bytes (though not repeat them that many times).
3048        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3049        made it possible to implement the integer overflow checks in a much more
3050        dynamic way, which I have now done. The artificial limitation on group
3051        length has been removed - we now have only the limit on the total length of
3052        the compiled pattern, which depends on the LINK_SIZE setting.
3053    
3054    10. Fixed a bug in the documentation for get/copy named substring when
3055        duplicate names are permitted. If none of the named substrings are set, the
3056        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3057        empty string.
3058    
3059    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3060        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3061        because the ] is interpreted as the first data character and the
3062        terminating ] is not found. PCRE has been made compatible with Perl in this
3063        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3064        cause memory overwriting.
3065    
3066    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3067        string has been matched (to stop an infinite loop). It was not recognizing
3068        a conditional subpattern that could match an empty string if that
3069        subpattern was within another subpattern. For example, it looped when
3070        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3071        condition was not nested. This bug has been fixed.
3072    
3073    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3074        past the start of the subject in the presence of bytes with the top bit
3075        set, for example "\x8aBCD".
3076    
3077    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3078        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3079    
3080    14. Optimized (?!) to (*FAIL).
3081    
3082    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3083        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3084        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3085        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3086        does: it's just the validity check that is more restrictive.
3087    
3088    16. Inserted checks for integer overflows during escape sequence (backslash)
3089        processing, and also fixed erroneous offset values for syntax errors during
3090        backslash processing.
3091    
3092    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3093        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3094    
3095    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3096        caused an overrun.
3097    
3098    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3099        something other than just ASCII characters) inside a group that had an
3100        unlimited repeat caused a loop at compile time (while checking to see
3101        whether the group could match an empty string).
3102    
3103    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3104        [\P{Any}] did so. (Error in the code for printing property names.)
3105    
3106    21. An orphan \E inside a character class could cause a crash.
3107    
3108    22. A repeated capturing bracket such as (A)? could cause a wild memory
3109        reference during compilation.
3110    
3111    23. There are several functions in pcre_compile() that scan along a compiled
3112        expression for various reasons (e.g. to see if it's fixed length for look
3113        behind). There were bugs in these functions when a repeated \p or \P was
3114        present in the pattern. These operators have additional parameters compared
3115        with \d, etc, and these were not being taken into account when moving along
3116        the compiled data. Specifically:
3117    
3118        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3119            length.
3120    
3121        (b) An item such as \pL+ within a repeated group could cause crashes or
3122            loops.
3123    
3124        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3125            "reference to non-existent subpattern" error.
3126    
3127        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3128    
3129    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3130        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3131    
3132    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3133    
3134    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3135        character were causing crashes (broken optimization).
3136    
3137    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3138        \p or \P) caused a compile-time loop.
3139    
3140    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3141        break. For example, the unstudied pattern [\r\n]A does not match the string
3142        "\r\nA" because change 7.0/46 below moves the current point on by two
3143        characters after failing to match at the start. However, the pattern \nA
3144        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3145        the same is true. There doesn't seem any very clean way out of this, but
3146        what I have chosen to do makes the common cases work: PCRE now takes note
3147        of whether there can be an explicit match for \r or \n anywhere in the
3148        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3149        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3150        pattern has explicit CR or LF references.
3151    
3152    29. Added (*CR) etc for changing newline setting at start of pattern.
3153    
3154    
3155  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3156  ---------------------  ---------------------

Legend:
Removed from v.200  
changed lines
  Added in v.1611

  ViewVC Help
Powered by ViewVC 1.1.5