/[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 1612 by ph10, Fri Nov 27 17:13:13 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    2.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
15        working correctly in UCP mode.
16    
17    
18    Version 8.38 23-November-2015
19    -----------------------------
20    
21    1.  If a group that contained a recursive back reference also contained a
22        forward reference subroutine call followed by a non-forward-reference
23        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
24        compile correct code, leading to undefined behaviour or an internally
25        detected error. This bug was discovered by the LLVM fuzzer.
26    
27    2.  Quantification of certain items (e.g. atomic back references) could cause
28        incorrect code to be compiled when recursive forward references were
29        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
30        This bug was discovered by the LLVM fuzzer.
31    
32    3.  A repeated conditional group whose condition was a reference by name caused
33        a buffer overflow if there was more than one group with the given name.
34        This bug was discovered by the LLVM fuzzer.
35    
36    4.  A recursive back reference by name within a group that had the same name as
37        another group caused a buffer overflow. For example:
38        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
39    
40    5.  A forward reference by name to a group whose number is the same as the
41        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
42        a buffer overflow at compile time. This bug was discovered by the LLVM
43        fuzzer.
44    
45    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
46        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
47    
48    7.  Another buffer overflow bug involved duplicate named groups with a
49        reference between their definition, with a group that reset capture
50        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
51        fixed by always allowing for more memory, even if not needed. (A proper fix
52        is implemented in PCRE2, but it involves more refactoring.)
53    
54    8.  There was no check for integer overflow in subroutine calls such as (?123).
55    
56    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
57        being treated as a literal 'l' instead of causing an error.
58    
59    10. There was a buffer overflow if pcre_exec() was called with an ovector of
60        size 1. This bug was found by american fuzzy lop.
61    
62    11. If a non-capturing group containing a conditional group that could match
63        an empty string was repeated, it was not identified as matching an empty
64        string itself. For example: /^(?:(?(1)x|)+)+$()/.
65    
66    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
67        \a and \e in test subject lines.
68    
69    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
70        instead of the EBCDIC value.
71    
72    14. The handling of \c in an EBCDIC environment has been revised so that it is
73        now compatible with the specification in Perl's perlebcdic page.
74    
75    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
76        ASCII/Unicode. This has now been added to the list of characters that are
77        recognized as white space in EBCDIC.
78    
79    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
80        error (correctly) when used outside a class, but did not give an error
81        within a class.
82    
83    17. \h within a class was incorrectly compiled in EBCDIC environments.
84    
85    18. A pattern with an unmatched closing parenthesis that contained a backward
86        assertion which itself contained a forward reference caused buffer
87        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
88    
89    19. JIT should return with error when the compiled pattern requires more stack
90        space than the maximum.
91    
92    20. A possessively repeated conditional group that could match an empty string,
93        for example, /(?(R))*+/, was incorrectly compiled.
94    
95    21. Fix infinite recursion in the JIT compiler when certain patterns such as
96        /(?:|a|){100}x/ are analysed.
97    
98    22. Some patterns with character classes involving [: and \\ were incorrectly
99        compiled and could cause reading from uninitialized memory or an incorrect
100        error diagnosis.
101    
102    23. Pathological patterns containing many nested occurrences of [: caused
103        pcre_compile() to run for a very long time.
104    
105    24. A conditional group with only one branch has an implicit empty alternative
106        branch and must therefore be treated as potentially matching an empty
107        string.
108    
109    25. If (?R was followed by - or + incorrect behaviour happened instead of a
110        diagnostic.
111    
112    26. Arrange to give up on finding the minimum matching length for overly
113        complex patterns.
114    
115    27. Similar to (4) above: in a pattern with duplicated named groups and an
116        occurrence of (?| it is possible for an apparently non-recursive back
117        reference to become recursive if a later named group with the relevant
118        number is encountered. This could lead to a buffer overflow. Wen Guanxing
119        from Venustech ADLAB discovered this bug.
120    
121    28. If pcregrep was given the -q option with -c or -l, or when handling a
122        binary file, it incorrectly wrote output to stdout.
123    
124    29. The JIT compiler did not restore the control verb head in case of *THEN
125        control verbs. This issue was found by Karl Skomski with a custom LLVM
126        fuzzer.
127    
128    30. Error messages for syntax errors following \g and \k were giving inaccurate
129        offsets in the pattern.
130    
131    31. Added a check for integer overflow in conditions (?(<digits>) and
132        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
133        fuzzer.
134    
135    32. Handling recursive references such as (?2) when the reference is to a group
136        later in the pattern uses code that is very hacked about and error-prone.
137        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
138        give an internal error if it is obvious that compiling has gone wrong.
139    
140    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
141        This issue was found by Karl Skomski with a custom LLVM fuzzer.
142    
143    34. The JIT compiler should restore the control chain for empty possessive
144        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
145    
146    35. Match limit check added to JIT recursion. This issue was found by Karl
147        Skomski with a custom LLVM fuzzer.
148    
149    36. Yet another case similar to 27 above has been circumvented by an
150        unconditional allocation of extra memory. This issue is fixed "properly" in
151        PCRE2 by refactoring the way references are handled. Wen Guanxing
152        from Venustech ADLAB discovered this bug.
153    
154    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
155        with a custom LLVM fuzzer.
156    
157    38. Fixed a corner case of range optimization in JIT.
158    
159    39. An incorrect error "overran compiling workspace" was given if there were
160        exactly enough group forward references such that the last one extended
161        into the workspace safety margin. The next one would have expanded the
162        workspace. The test for overflow was not including the safety margin.
163    
164    40. A match limit issue is fixed in JIT which was found by Karl Skomski
165        with a custom LLVM fuzzer.
166    
167    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
168        work under Windows. (Why has it taken so long for anyone to notice?)
169    
170    42. In a character class such as [\W\p{Any}] where both a negative-type escape
171        ("not a word character") and a property escape were present, the property
172        escape was being ignored.
173    
174    43. Fix crash caused by very long (*MARK) or (*THEN) names.
175    
176    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
177        by a single ASCII character in a class item, was incorrectly compiled in
178        UCP mode. The POSIX class got lost, but only if the single character
179        followed it.
180    
181    45. [:punct:] in UCP mode was matching some characters in the range 128-255
182        that should not have been matched.
183    
184    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
185        class, all characters with code points greater than 255 are in the class.
186        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
187        such as \w are turned into Unicode properties), wide characters were not
188        correctly handled, and could fail to match.
189    
190    
191    Version 8.37 28-April-2015
192    --------------------------
193    
194    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
195        for those parentheses to be closed with whatever has been captured so far.
196        However, it was failing to mark any other groups between the hightest
197        capture so far and the currrent group as "unset". Thus, the ovector for
198        those groups contained whatever was previously there. An example is the
199        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
200    
201    2.  If an assertion condition was quantified with a minimum of zero (an odd
202        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
203    
204    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
205        unrecognized modifier, a crash could occur.
206    
207    4.  An attempt to do global matching in pcretest with a zero-length ovector
208        caused a crash.
209    
210    5.  Fixed a memory leak during matching that could occur for a subpattern
211        subroutine call (recursive or otherwise) if the number of captured groups
212        that had to be saved was greater than ten.
213    
214    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
215        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
216        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
217    
218    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
219        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
220    
221    8. When a pattern is compiled, it remembers the highest back reference so that
222       when matching, if the ovector is too small, extra memory can be obtained to
223       use instead. A conditional subpattern whose condition is a check on a
224       capture having happened, such as, for example in the pattern
225       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
226       setting the highest backreference number. This mattered only if pcre_exec()
227       was called with an ovector that was too small to hold the capture, and there
228       was no other kind of back reference (a situation which is probably quite
229       rare). The effect of the bug was that the condition was always treated as
230       FALSE when the capture could not be consulted, leading to a incorrect
231       behaviour by pcre_exec(). This bug has been fixed.
232    
233    9. A reference to a duplicated named group (either a back reference or a test
234       for being set in a conditional) that occurred in a part of the pattern where
235       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
236       to be incorrectly calculated, leading to overwriting.
237    
238    10. A mutually recursive set of back references such as (\2)(\1) caused a
239        segfault at study time (while trying to find the minimum matching length).
240        The infinite loop is now broken (with the minimum length unset, that is,
241        zero).
242    
243    11. If an assertion that was used as a condition was quantified with a minimum
244        of zero, matching went wrong. In particular, if the whole group had
245        unlimited repetition and could match an empty string, a segfault was
246        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
247        assertions to be quantified, but not if they are being used as conditions,
248        so the above pattern is faulted by Perl. PCRE has now been changed so that
249        it also rejects such patterns.
250    
251    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
252        failed to allow the zero-repeat case if pcre2_exec() was called with an
253        ovector too small to capture the group.
254    
255    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
256        Red Hat Product Security:
257    
258        (a) A crash if /K and /F were both set with the option to save the compiled
259        pattern.
260    
261        (b) Another crash if the option to print captured substrings in a callout
262        was combined with setting a null ovector, for example \O\C+ as a subject
263        string.
264    
265    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
266        forward reference repeated a large (but limited) number of times within a
267        repeated outer group that has a zero minimum quantifier, caused incorrect
268        code to be compiled, leading to the error "internal error:
269        previously-checked referenced subpattern not found" when an incorrect
270        memory address was read. This bug was reported as "heap overflow",
271        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
272        CVE-2015-2325.
273    
274    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
275        call within a group that also contained a recursive back reference caused
276        incorrect code to be compiled. This bug was reported as "heap overflow",
277        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
278        number CVE-2015-2326.
279    
280    24. Computing the size of the JIT read-only data in advance has been a source
281        of various issues, and new ones are still appear unfortunately. To fix
282        existing and future issues, size computation is eliminated from the code,
283        and replaced by on-demand memory allocation.
284    
285    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
286        adjacent to the end of the range, and the range contained characters with
287        more than one other case, caused incorrect behaviour when compiled in UTF
288        mode. In that example, the range a-j was left out of the class.
289    
290    26. Fix JIT compilation of conditional blocks, which assertion
291        is converted to (*FAIL). E.g: /(?(?!))/.
292    
293    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
294        discovered by the LLVM fuzzer.
295    
296    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
297        when this assertion was used as a condition, for example (?(?!)a|b). In
298        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
299        error about an unsupported item.
300    
301    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
302        possessification code could take exponential time to complete. A recursion
303        depth limit of 1000 has been imposed to limit the resources used by this
304        optimization.
305    
306    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
307        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
308        because \S ensures they are all in the class. The code for doing this was
309        interacting badly with the code for computing the amount of space needed to
310        compile the pattern, leading to a buffer overflow. This bug was discovered
311        by the LLVM fuzzer.
312    
313    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
314        other kinds of group caused stack overflow at compile time. This bug was
315        discovered by the LLVM fuzzer.
316    
317    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
318        between a subroutine call and its quantifier was incorrectly compiled,
319        leading to buffer overflow or other errors. This bug was discovered by the
320        LLVM fuzzer.
321    
322    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
323        assertion after (?(. The code was failing to check the character after
324        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
325        was discovered by the LLVM fuzzer.
326    
327    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
328        a fixed maximum following a group that contains a subroutine reference was
329        incorrectly compiled and could trigger buffer overflow. This bug was
330        discovered by the LLVM fuzzer.
331    
332    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
333        caused a stack overflow instead of the diagnosis of a non-fixed length
334        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
335    
336    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
337        (e.g. /(?<=\Ka)/) could make pcregrep loop.
338    
339    37. There was a similar problem to 36 in pcretest for global matches.
340    
341    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
342        and a subsequent item in the pattern caused a non-match, backtracking over
343        the repeated \X did not stop, but carried on past the start of the subject,
344        causing reference to random memory and/or a segfault. There were also some
345        other cases where backtracking after \C could crash. This set of bugs was
346        discovered by the LLVM fuzzer.
347    
348    39. The function for finding the minimum length of a matching string could take
349        a very long time if mutual recursion was present many times in a pattern,
350        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
351        method has been implemented. This infelicity was discovered by the LLVM
352        fuzzer.
353    
354    40. Static linking against the PCRE library using the pkg-config module was
355        failing on missing pthread symbols.
356    
357    
358    Version 8.36 26-September-2014
359    ------------------------------
360    
361    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
362        -Wmissing-field-initializers and -Wunused-parameter.
363    
364    2.  The tests for quantifiers being too big (greater than 65535) were being
365        applied after reading the number, and stupidly assuming that integer
366        overflow would give a negative number. The tests are now applied as the
367        numbers are read.
368    
369    3.  Tidy code in pcre_exec.c where two branches that used to be different are
370        now the same.
371    
372    4.  The JIT compiler did not generate match limit checks for certain
373        bracketed expressions with quantifiers. This may lead to exponential
374        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
375        issue should be resolved now.
376    
377    5.  Fixed an issue, which occures when nested alternatives are optimized
378        with table jumps.
379    
380    6.  Inserted two casts and changed some ints to size_t in the light of some
381        reported 64-bit compiler warnings (Bugzilla 1477).
382    
383    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
384        an empty string, which sometimes were behaving incorrectly in the
385        interpreter (though correctly in the JIT matcher). This pcretest input is
386        an example:
387    
388          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
389          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
390    
391        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
392        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
393        for an empty string was breaking the inner loop and carrying on at a lower
394        level, when possessive repeated groups should always return to a higher
395        level as they have no backtrack points in them. The empty string test now
396        occurs at the outer level.
397    
398    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
399        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
400    
401    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
402        doesn't).
403    
404    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
405        to hold a pattern to be incorrectly computed (too small) when there were
406        named back references to duplicated names. This could cause "internal
407        error: code overflow" or "double free or corruption" or other memory
408        handling errors.
409    
410    11. When named subpatterns had the same prefixes, back references could be
411        confused. For example, in this pattern:
412    
413          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
414    
415        the reference to 'Name' was incorrectly treated as a reference to a
416        duplicate name.
417    
418    12. A pattern such as /^s?c/mi8 where the optional character has more than
419        one "other case" was incorrectly compiled such that it would only try to
420        match starting at "c".
421    
422    13. When a pattern starting with \s was studied, VT was not included in the
423        list of possible starting characters; this should have been part of the
424        8.34/18 patch.
425    
426    14. If a character class started [\Qx]... where x is any character, the class
427        was incorrectly terminated at the ].
428    
429    15. If a pattern that started with a caseless match for a character with more
430        than one "other case" was studied, PCRE did not set up the starting code
431        unit bit map for the list of possible characters. Now it does. This is an
432        optimization improvement, not a bug fix.
433    
434    16. The Unicode data tables have been updated to Unicode 7.0.0.
435    
436    17. Fixed a number of memory leaks in pcregrep.
437    
438    18. Avoid a compiler warning (from some compilers) for a function call with
439        a cast that removes "const" from an lvalue by using an intermediate
440        variable (to which the compiler does not object).
441    
442    19. Incorrect code was compiled if a group that contained an internal recursive
443        back reference was optional (had quantifier with a minimum of zero). This
444        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
445        caused segmentation faults because of stack overflows at compile time.
446    
447    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
448        group that is quantified with an indefinite repeat, caused a compile-time
449        loop which used up all the system stack and provoked a segmentation fault.
450        This was not the same bug as 19 above.
451    
452    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
453        Patch by Mike Frysinger.
454    
455    
456    Version 8.35 04-April-2014
457    --------------------------
458    
459    1.  A new flag is set, when property checks are present in an XCLASS.
460        When this flag is not set, PCRE can perform certain optimizations
461        such as studying these XCLASS-es.
462    
463    2.  The auto-possessification of character sets were improved: a normal
464        and an extended character set can be compared now. Furthermore
465        the JIT compiler optimizes more character set checks.
466    
467    3.  Got rid of some compiler warnings for potentially uninitialized variables
468        that show up only when compiled with -O2.
469    
470    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
471        of a match later then the end of the match. The pcretest program was not
472        handling the case sensibly - it was outputting from the start to the next
473        binary zero. It now reports this situation in a message, and outputs the
474        text from the end to the start.
475    
476    5.  Fast forward search is improved in JIT. Instead of the first three
477        characters, any three characters with fixed position can be searched.
478        Search order: first, last, middle.
479    
480    6.  Improve character range checks in JIT. Characters are read by an inprecise
481        function now, which returns with an unknown value if the character code is
482        above a certain threshold (e.g: 256). The only limitation is that the value
483        must be bigger than the threshold as well. This function is useful when
484        the characters above the threshold are handled in the same way.
485    
486    7.  The macros whose names start with RAWUCHAR are placeholders for a future
487        mode in which only the bottom 21 bits of 32-bit data items are used. To
488        make this more memorable for those maintaining the code, the names have
489        been changed to start with UCHAR21, and an extensive comment has been added
490        to their definition.
491    
492    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
493        to the export list in Makefile.am (they were accidentally omitted from the
494        8.34 tarball).
495    
496    9.  The informational output from pcretest used the phrase "starting byte set"
497        which is inappropriate for the 16-bit and 32-bit libraries. As the output
498        for "first char" and "need char" really means "non-UTF-char", I've changed
499        "byte" to "char", and slightly reworded the output. The documentation about
500        these values has also been (I hope) clarified.
501    
502    10. Another JIT related optimization: use table jumps for selecting the correct
503        backtracking path, when more than four alternatives are present inside a
504        bracket.
505    
506    11. Empty match is not possible, when the minimum length is greater than zero,
507        and there is no \K in the pattern. JIT should avoid empty match checks in
508        such cases.
509    
510    12. In a caseless character class with UCP support, when a character with more
511        than one alternative case was not the first character of a range, not all
512        the alternative cases were added to the class. For example, s and \x{17f}
513        are both alternative cases for S: the class [RST] was handled correctly,
514        but [R-T] was not.
515    
516    13. The configure.ac file always checked for pthread support when JIT was
517        enabled. This is not used in Windows, so I have put this test inside a
518        check for the presence of windows.h (which was already tested for).
519    
520    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
521        The algorithm provides a way to skip certain starting offsets, and usually
522        faster than linear prefix searches.
523    
524    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
525        as for 'fr_FR' and 'french'. For some reason, however, it then used the
526        Windows-specific input and output files, which have 'french' screwed in.
527        So this could never have worked. One of the problems with locales is that
528        they aren't always the same. I have now updated RunTest so that it checks
529        the output of the locale test (test 3) against three different output
530        files, and it allows the test to pass if any one of them matches. With luck
531        this should make the test pass on some versions of Solaris where it was
532        failing. Because of the uncertainty, the script did not used to stop if
533        test 3 failed; it now does. If further versions of a French locale ever
534        come to light, they can now easily be added.
535    
536    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
537        there was a message during ./configure, but it did not stop. This now
538        provokes an error. The invalid example in README has been corrected.
539        If a value less than the minimum is given, the minimum value has always
540        been used, but now a warning is given.
541    
542    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
543        was a bug in the test system, which is now fixed. Also, the list of various
544        configurations that are tested for each release did not have one with both
545        16/32 bits and --enable-bar-anycrlf. It now does.
546    
547    18. pcretest was missing "-C bsr" for displaying the \R default setting.
548    
549    19. Little endian PowerPC systems are supported now by the JIT compiler.
550    
551    20. The fast forward newline mechanism could enter to an infinite loop on
552        certain invalid UTF-8 input. Although we don't support these cases
553        this issue can be fixed by a performance optimization.
554    
555    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
556        not take account if existing stack usage. There is now a new global
557        variable called pcre_stack_guard that can be set to point to an external
558        function to check stack availability. It is called at the start of
559        processing every parenthesized group.
560    
561    22. A typo in the code meant that in ungreedy mode the max/min qualifier
562        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
563        not match "ab".
564    
565    23. When UTF was disabled, the JIT program reported some incorrect compile
566        errors. These messages are silenced now.
567    
568    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
569        compiler.
570    
571    25. Change all the temporary files used in RunGrepTest to be different to those
572        used by RunTest so that the tests can be run simultaneously, for example by
573        "make -j check".
574    
575    
576    Version 8.34 15-December-2013
577    -----------------------------
578    
579    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
580        executable memory. Patch inspired by Carsten Klein.
581    
582    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
583        this macro is never tested and has no effect, because the work to support
584        coverage involves only compiling and linking options and special targets in
585        the Makefile. The comment in config.h implied that defining the macro would
586        enable coverage support, which is totally false. There was also support for
587        setting this macro in the CMake files (my fault, I just copied it from
588        configure). SUPPORT_GCOV has now been removed.
589    
590    3.  Make a small performance improvement in strlen16() and strlen32() in
591        pcretest.
592    
593    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
594        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
595        these cases). There was also one in pcretest.c.
596    
597    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
598    
599    6.  In UTF mode, the code for checking whether a group could match an empty
600        string (which is used for indefinitely repeated groups to allow for
601        breaking an infinite loop) was broken when the group contained a repeated
602        negated single-character class with a character that occupied more than one
603        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
604        UTF-8 mode). The effect was undefined: the group might or might not be
605        deemed as matching an empty string, or the program might have crashed.
606    
607    7.  The code for checking whether a group could match an empty string was not
608        recognizing that \h, \H, \v, \V, and \R must match a character.
609    
610    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
611        an empty string. If it can, pcretest shows this in its information output.
612    
613    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
614        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
615        matched by pcre_exec() without using JIT:
616    
617        (a) If the rest of the pattern did not match after a maximal run of
618            grapheme clusters, the code for backing up to try with fewer of them
619            did not always back up over a full grapheme when characters that do not
620            have the modifier quality were involved, e.g. Hangul syllables.
621    
622        (b) If the match point in a subject started with modifier character, and
623            there was no match, the code could incorrectly back up beyond the match
624            point, and potentially beyond the first character in the subject,
625            leading to a segfault or an incorrect match result.
626    
627    10. A conditional group with an assertion condition could lead to PCRE
628        recording an incorrect first data item for a match if no other first data
629        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
630        first data item, and therefore matched "ca" after "c" instead of at the
631        start.
632    
633    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
634        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
635    
636    12. The source of pcregrep now includes z/OS-specific code so that it can be
637        compiled for z/OS as part of the special z/OS distribution.
638    
639    13. Added the -T and -TM options to pcretest.
640    
641    14. The code in pcre_compile.c for creating the table of named capturing groups
642        has been refactored. Instead of creating the table dynamically during the
643        actual compiling pass, the information is remembered during the pre-compile
644        pass (on the stack unless there are more than 20 named groups, in which
645        case malloc() is used) and the whole table is created before the actual
646        compile happens. This has simplified the code (it is now nearly 150 lines
647        shorter) and prepared the way for better handling of references to groups
648        with duplicate names.
649    
650    15. A back reference to a named subpattern when there is more than one of the
651        same name now checks them in the order in which they appear in the pattern.
652        The first one that is set is used for the reference. Previously only the
653        first one was inspected. This change makes PCRE more compatible with Perl.
654    
655    16. Unicode character properties were updated from Unicode 6.3.0.
656    
657    17. The compile-time code for auto-possessification has been refactored, based
658        on a patch by Zoltan Herczeg. It now happens after instead of during
659        compilation. The code is cleaner, and more cases are handled. The option
660        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
661        options in pcretest are provided to set it. It can also be set by
662        (*NO_AUTO_POSSESS) at the start of a pattern.
663    
664    18. The character VT has been added to the default ("C" locale) set of
665        characters that match \s and are generally treated as white space,
666        following this same change in Perl 5.18. There is now no difference between
667        "Perl space" and "POSIX space". Whether VT is treated as white space in
668        other locales depends on the locale.
669    
670    19. The code for checking named groups as conditions, either for being set or
671        for being recursed, has been refactored (this is related to 14 and 15
672        above). Processing unduplicated named groups should now be as fast at
673        numerical groups, and processing duplicated groups should be faster than
674        before.
675    
676    20. Two patches to the CMake build system, by Alexander Barkov:
677    
678          (1) Replace the "source" command by "." in CMakeLists.txt because
679              "source" is a bash-ism.
680    
681          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
682              without these the CMake build does not work on Solaris.
683    
684    21. Perl has changed its handling of \8 and \9. If there is no previously
685        encountered capturing group of those numbers, they are treated as the
686        literal characters 8 and 9 instead of a binary zero followed by the
687        literals. PCRE now does the same.
688    
689    22. Following Perl, added \o{} to specify codepoints in octal, making it
690        possible to specify values greater than 0777 and also making them
691        unambiguous.
692    
693    23. Perl now gives an error for missing closing braces after \x{... instead of
694        treating the string as literal. PCRE now does the same.
695    
696    24. RunTest used to grumble if an inappropriate test was selected explicitly,
697        but just skip it when running all tests. This make it awkward to run ranges
698        of tests when one of them was inappropriate. Now it just skips any
699        inappropriate tests, as it always did when running all tests.
700    
701    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
702        character types such as \d or \w, too many callouts were inserted, and the
703        data that they returned was rubbish.
704    
705    26. In UCP mode, \s was not matching two of the characters that Perl matches,
706        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
707        were matched by \h. The code has now been refactored so that the lists of
708        the horizontal and vertical whitespace characters used for \h and \v (which
709        are defined only in one place) are now also used for \s.
710    
711    27. Add JIT support for the 64 bit TileGX architecture.
712        Patch by Jiong Wang (Tilera Corporation).
713    
714    28. Possessive quantifiers for classes (both explicit and automatically
715        generated) now use special opcodes instead of wrapping in ONCE brackets.
716    
717    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
718        (because it's meaningless), this was not happening when PCRE_CASELESS was
719        set. Not wrong, but inefficient.
720    
721    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
722        properties for \w, \d, etc) is present in a test regex. Otherwise if the
723        test contains no characters greater than 255, Perl doesn't realise it
724        should be using Unicode semantics.
725    
726    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
727        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
728        does in Unicode mode.
729    
730    32. Added the "forbid" facility to pcretest so that putting tests into the
731        wrong test files can sometimes be quickly detected.
732    
733    33. There is now a limit (default 250) on the depth of nesting of parentheses.
734        This limit is imposed to control the amount of system stack used at compile
735        time. It can be changed at build time by --with-parens-nest-limit=xxx or
736        the equivalent in CMake.
737    
738    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
739        errors. Perl warns for these when in warning mode, but PCRE has no facility
740        for giving warnings.
741    
742    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
743        However, this was not working for (?!) because it is optimized to (*FAIL),
744        for which PCRE does not allow quantifiers. The optimization is now disabled
745        when a quantifier follows (?!). I can't see any use for this, but it makes
746        things uniform.
747    
748    36. Perl no longer allows group names to start with digits, so I have made this
749        change also in PCRE. It simplifies the code a bit.
750    
751    37. In extended mode, Perl ignores spaces before a + that indicates a
752        possessive quantifier. PCRE allowed a space before the quantifier, but not
753        before the possessive +. It now does.
754    
755    38. The use of \K (reset reported match start) within a repeated possessive
756        group such as (a\Kb)*+ was not working.
757    
758    40. Document that the same character tables must be used at compile time and
759        run time, and that the facility to pass tables to pcre_exec() and
760        pcre_dfa_exec() is for use only with saved/restored patterns.
761    
762    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
763        features for Builds with MSVC:
764    
765        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
766           builds). The .rc files can be used to set FileDescription and many other
767           attributes.
768    
769        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
770           This allows higher-level build scripts which want .pdb files to avoid
771           hard-coding the exact files needed."
772    
773    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
774        mean "start of word" and "end of word", respectively, as a transition aid.
775    
776    43. A minimizing repeat of a class containing codepoints greater than 255 in
777        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
778        compiled to use the heap for recursion.
779    
780    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
781        is configured.
782    
783    
784    Version 8.33 28-May-2013
785    ------------------------
786    
787    1.  Added 'U' to some constants that are compared to unsigned integers, to
788        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
789        variables that are added to signed variables, to ensure the result is
790        signed and can be negated.
791    
792    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
793        CMake config files.
794    
795    3.  Revise the creation of config.h.generic so that all boolean macros are
796        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
797        overriding via -D on the command line possible.
798    
799    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
800        to unsigned int is reported to make a quite noticeable speed difference in
801        a specific Windows environment. Testing on Linux did also appear to show
802        some benefit (and it is clearly not harmful). Also fixed the definition of
803        Xop which should be unsigned.
804    
805    5.  Related to (4), changing the definition of the intermediate variable cc
806        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
807        improvements.
808    
809    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
810        unnecessary spaces.
811    
812    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
813        and later.
814    
815    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
816    
817    9.  Optimizing fast_forward_start_bits in JIT.
818    
819    10. Adding support for callouts in JIT, and fixing some issues revealed
820        during this work. Namely:
821    
822        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
823    
824        (b) Minimum length was not checked before the matching is started.
825    
826    11. The value of capture_last that is passed to callouts was incorrect in some
827        cases when there was a capture on one path that was subsequently abandoned
828        after a backtrack. Also, the capture_last value is now reset after a
829        recursion, since all captures are also reset in this case.
830    
831    12. The interpreter no longer returns the "too many substrings" error in the
832        case when an overflowing capture is in a branch that is subsequently
833        abandoned after a backtrack.
834    
835    13. In the pathological case when an offset vector of size 2 is used, pcretest
836        now prints out the matched string after a yield of 0 or 1.
837    
838    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
839        Only supported by the JIT compiler at the moment.
840    
841    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
842    
843    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
844        offset of the starting point of the matching process, provided the offsets
845        vector is large enough.
846    
847    17. The \A escape now records a lookbehind value of 1, though its execution
848        does not actually inspect the previous character. This is to ensure that,
849        in partial multi-segment matching, at least one character from the old
850        segment is retained when a new segment is processed. Otherwise, if there
851        are no lookbehinds in the pattern, \A might match incorrectly at the start
852        of a new segment.
853    
854    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
855    
856    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
857        gives some modest performance improvement in 8-bit mode.
858    
859    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
860        be expressed in certain programming languages using Universal Character
861        Names.
862    
863    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
864        which points out that "non characters" are not "characters that may not
865        appear in Unicode strings" but rather "characters that are reserved for
866        internal use and have only local meaning".
867    
868    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
869        there was a conditional group that depended on an assertion, if the
870        assertion was false, the callout that immediately followed the alternation
871        in the condition was skipped when pcre_exec() was used for matching.
872    
873    23. Allow an explicit callout to be inserted before an assertion that is the
874        condition for a conditional group, for compatibility with automatic
875        callouts, which always insert a callout at this point.
876    
877    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
878        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
879    
880    25. (*PRUNE) is now supported by the JIT compiler.
881    
882    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
883    
884    27. Fix the case where there are two or more SKIPs with arguments that may be
885        ignored.
886    
887    28. (*SKIP) is now supported by the JIT compiler.
888    
889    29. (*THEN) is now supported by the JIT compiler.
890    
891    30. Update RunTest with additional test selector options.
892    
893    31. The way PCRE handles backtracking verbs has been changed in two ways.
894    
895        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
896        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
897        backtracking. In some cases this makes it more Perl-compatible, but Perl's
898        rather obscure rules do not always do the same thing.
899    
900        (2) Previously, backtracking verbs were confined within assertions. This is
901        no longer the case for positive assertions, except for (*ACCEPT). Again,
902        this sometimes improves Perl compatibility, and sometimes does not.
903    
904    32. A number of tests that were in test 2 because Perl did things differently
905        have been moved to test 1, because either Perl or PCRE has changed, and
906        these tests are now compatible.
907    
908    32. Backtracking control verbs are now handled in the same way in JIT and
909        interpreter.
910    
911    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
912        contained a forward subroutine reference caused a compile error.
913    
914    34. Auto-detect and optimize limited repetitions in JIT.
915    
916    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
917        blocking (*UTF) etc.
918    
919    36. In the interpreter, maximizing pattern repetitions for characters and
920        character types now use tail recursion, which reduces stack usage.
921    
922    37. The value of the max lookbehind was not correctly preserved if a compiled
923        and saved regex was reloaded on a host of different endianness.
924    
925    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
926        of the compiled pattern block, expand the flags field from 16 to 32 bits
927        because it was almost full.
928    
929    39. Try madvise first before posix_madvise.
930    
931    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
932        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
933        original change locked out all matches of empty strings. This has been
934        changed so that one match of an empty string per line is recognized.
935        Subsequent searches on the same line (for colouring or for --only-matching,
936        for example) do not recognize empty strings.
937    
938    41. Applied a user patch to fix a number of spelling mistakes in comments.
939    
940    42. Data lines longer than 65536 caused pcretest to crash.
941    
942    43. Clarified the data type for length and startoffset arguments for pcre_exec
943        and pcre_dfa_exec in the function-specific man pages, where they were
944        explicitly stated to be in bytes, never having been updated. I also added
945        some clarification to the pcreapi man page.
946    
947    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
948        a segmentation fault.
949    
950    
951    Version 8.32 30-November-2012
952    -----------------------------
953    
954    1.  Improved JIT compiler optimizations for first character search and single
955        character iterators.
956    
957    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
958        Patch by Daniel Richard G.
959    
960    3.  Single character iterator optimizations in the JIT compiler.
961    
962    4.  Improved JIT compiler optimizations for character ranges.
963    
964    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
965        Reported by Giuseppe D'Angelo.
966    
967    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
968        of a line, was being set incorrectly in cases where .* appeared inside
969        atomic brackets at the start of a pattern, or where there was a subsequent
970        *PRUNE or *SKIP.
971    
972    7.  Improved instruction cache flush for POWER/PowerPC.
973        Patch by Daniel Richard G.
974    
975    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
976        grep:
977    
978        (a) There is now no limit to the number of patterns to be matched.
979    
980        (b) An error is given if a pattern is too long.
981    
982        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
983            are now supported.
984    
985        (d) --exclude-from and --include-from (multiple use) have been added.
986    
987        (e) Exclusions and inclusions now apply to all files and directories, not
988            just to those obtained from scanning a directory recursively.
989    
990        (f) Multiple uses of -f and --file-list are now supported.
991    
992        (g) In a Windows environment, the default for -d has been changed from
993            "read" (the GNU grep default) to "skip", because otherwise the presence
994            of a directory in the file list provokes an error.
995    
996        (h) The documentation has been revised and clarified in places.
997    
998    9.  Improve the matching speed of capturing brackets.
999    
1000    10. Changed the meaning of \X so that it now matches a Unicode extended
1001        grapheme cluster.
1002    
1003    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1004        out POSIX threads when JIT support is configured.
1005    
1006    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1007    
1008    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1009        structure could go wrong in environments where size_t is not the same size
1010        as int.
1011    
1012    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1013        set.
1014    
1015    15. The EBCDIC support had decayed; later updates to the code had included
1016        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1017        general tidy up of EBCDIC-related issues, and the documentation was also
1018        not quite right. There is now a test that can be run on ASCII systems to
1019        check some of the EBCDIC-related things (but is it not a full test).
1020    
1021    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1022        in a small tidy to the code.
1023    
1024    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1025    
1026    18. If the --only-matching (-o) option in pcregrep is specified multiple
1027        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1028        substrings matched by the 1st and 2nd capturing parentheses. A separating
1029        string can be specified by --om-separator (default empty).
1030    
1031    19. Improving the first n character searches.
1032    
1033    20. Turn case lists for horizontal and vertical white space into macros so that
1034        they are defined only once.
1035    
1036    21. This set of changes together give more compatible Unicode case-folding
1037        behaviour for characters that have more than one other case when UCP
1038        support is available.
1039    
1040        (a) The Unicode property table now has offsets into a new table of sets of
1041            three or more characters that are case-equivalent. The MultiStage2.py
1042            script that generates these tables (the pcre_ucd.c file) now scans
1043            CaseFolding.txt instead of UnicodeData.txt for character case
1044            information.
1045    
1046        (b) The code for adding characters or ranges of characters to a character
1047            class has been abstracted into a generalized function that also handles
1048            case-independence. In UTF-mode with UCP support, this uses the new data
1049            to handle characters with more than one other case.
1050    
1051        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1052            whose other case is greater than 256 are now correctly matched
1053            caselessly. Previously, the high codepoint matched the low one, but not
1054            vice versa.
1055    
1056        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1057            of the new class addition function, using character lists defined as
1058            macros alongside the case definitions of 20 above.
1059    
1060        (e) Caseless back references now work with characters that have more than
1061            one other case.
1062    
1063        (f) General caseless matching of characters with more than one other case
1064            is supported.
1065    
1066    22. Unicode character properties were updated from Unicode 6.2.0
1067    
1068    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1069    
1070    24. Add support for 32-bit character strings, and UTF-32
1071    
1072    25. Major JIT compiler update (code refactoring and bugfixing).
1073        Experimental Sparc 32 support is added.
1074    
1075    26. Applied a modified version of Daniel Richard G's patch to create
1076        pcre.h.generic and config.h.generic by "make" instead of in the
1077        PrepareRelease script.
1078    
1079    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1080        pcre_compile.c when checking for a zero character.
1081    
1082    28. Introducing a native interface for JIT. Through this interface, the compiled
1083        machine code can be directly executed. The purpose of this interface is to
1084        provide fast pattern matching, so several sanity checks are not performed.
1085        However, feature tests are still performed. The new interface provides
1086        1.4x speedup compared to the old one.
1087    
1088    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1089        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1090        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1091    
1092    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1093        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1094        the "old" RFC 2279). Instead, it ended up passing a negative length to
1095        pcre_exec().
1096    
1097    31. Add support for GCC's visibility feature to hide internal functions.
1098    
1099    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1100        "unknown -C option" after outputting 0 or 1.
1101    
1102    33. There is now support for generating a code coverage report for the test
1103        suite in environments where gcc is the compiler and lcov is installed. This
1104        is mainly for the benefit of the developers.
1105    
1106    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1107        unaddressable using valgrind annotations, allowing valgrind to detect
1108        invalid memory accesses. This is mainly for the benefit of the developers.
1109    
1110    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1111    
1112    26. Give configure error if --enable-cpp but no C++ compiler found.
1113    
1114    
1115    Version 8.31 06-July-2012
1116    -------------------------
1117    
1118    1.  Fixing a wrong JIT test case and some compiler warnings.
1119    
1120    2.  Removed a bashism from the RunTest script.
1121    
1122    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1123        to unsigned type, result still unsigned" that was given by an MS compiler
1124        on encountering the code "-sizeof(xxx)".
1125    
1126    4.  Partial matching support is added to the JIT compiler.
1127    
1128    5.  Fixed several bugs concerned with partial matching of items that consist
1129        of more than one character:
1130    
1131        (a) /^(..)\1/ did not partially match "aba" because checking references was
1132            done on an "all or nothing" basis. This also applied to repeated
1133            references.
1134    
1135        (b) \R did not give a hard partial match if \r was found at the end of the
1136            subject.
1137    
1138        (c) \X did not give a hard partial match after matching one or more
1139            characters at the end of the subject.
1140    
1141        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1142            a partial match for the string "\r".
1143    
1144        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1145            a partial match for a CR character at the end of the subject string.
1146    
1147    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1148        running pcretest, the text "(JIT)" added to the output whenever JIT is
1149        actually used to run the match.
1150    
1151    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1152        or /S+[+] with a digit between 1 and 7.
1153    
1154    8.  OP_NOT now supports any UTF character not just single-byte ones.
1155    
1156    9.  (*MARK) control verb is now supported by the JIT compiler.
1157    
1158    10. The command "./RunTest list" lists the available tests without actually
1159        running any of them. (Because I keep forgetting what they all are.)
1160    
1161    11. Add PCRE_INFO_MAXLOOKBEHIND.
1162    
1163    12. Applied a (slightly modified) user-supplied patch that improves performance
1164        when the heap is used for recursion (compiled with --disable-stack-for-
1165        recursion). Instead of malloc and free for each heap frame each time a
1166        logical recursion happens, frames are retained on a chain and re-used where
1167        possible. This sometimes gives as much as 30% improvement.
1168    
1169    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1170        call.
1171    
1172    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1173    
1174    15. It is now possible to link pcretest with libedit as an alternative to
1175        libreadline.
1176    
1177    16. (*COMMIT) control verb is now supported by the JIT compiler.
1178    
1179    17. The Unicode data tables have been updated to Unicode 6.1.0.
1180    
1181    18. Added --file-list option to pcregrep.
1182    
1183    19. Added binary file support to pcregrep, including the -a, --binary-files,
1184        -I, and --text options.
1185    
1186    20. The madvise function is renamed for posix_madvise for QNX compatibility
1187        reasons. Fixed by Giuseppe D'Angelo.
1188    
1189    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1190    
1191    22. Changed the option for creating symbolic links for 16-bit man pages from
1192        -s to -sf so that re-installing does not cause issues.
1193    
1194    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1195    
1196    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1197        matches in certain environments (the workspace was not being correctly
1198        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1199        some of the workspace data at the beginning of a restart.
1200    
1201    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1202        was not doing so when it should - probably a typo introduced by SVN 528
1203        (change 8.10/14).
1204    
1205    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1206        \w+ when the character tables indicated that \x{c4} was a word character.
1207        There were several related cases, all because the tests for doing a table
1208        lookup were testing for characters less than 127 instead of 255.
1209    
1210    27. If a pattern contains capturing parentheses that are not used in a match,
1211        their slots in the ovector are set to -1. For those that are higher than
1212        any matched groups, this happens at the end of processing. In the case when
1213        there were back references that the ovector was too small to contain
1214        (causing temporary malloc'd memory to be used during matching), and the
1215        highest capturing number was not used, memory off the end of the ovector
1216        was incorrectly being set to -1. (It was using the size of the temporary
1217        memory instead of the true size.)
1218    
1219    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1220        ovector size, it uses memory at the end of the block that it has got.
1221    
1222    29. Check for an overlong MARK name and give an error at compile time. The
1223        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1224    
1225    30. JIT compiler update.
1226    
1227    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1228        Rill for the patch.
1229    
1230    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1231    
1232    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1233    
1234    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1235        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1236        was enabled.
1237    
1238    35. Improve JIT code generation for greedy plus quantifier.
1239    
1240    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1241        1 to "aa" instead of to an empty string. The bug affected repeated groups
1242        that could potentially match an empty string.
1243    
1244    37. Optimizing single character iterators in JIT.
1245    
1246    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1247        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1248        codepoints that are too big for the mode are faulted, and in a UTF mode,
1249        disallowed codepoints are also faulted.
1250    
1251    39. If PCRE was compiled with UTF support, in three places in the DFA
1252        matcher there was code that should only have been obeyed in UTF mode, but
1253        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1254        processing when bytes with values greater than 127 were present. In 16-bit
1255        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1256        both cases the values are those that cannot be the first data item in a UTF
1257        character. The three items that might have provoked this were recursions,
1258        possessively repeated groups, and atomic groups.
1259    
1260    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1261        and pcregrep, because some OS require shared objects to be explicitly
1262        passed to ld, causing the link step to fail if they are not.
1263    
1264    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1265        mode, patterns that started with \h* or \R* might be incorrectly matched.
1266    
1267    
1268    Version 8.30 04-February-2012
1269    -----------------------------
1270    
1271    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1272        name is defined in ctype.h.
1273    
1274    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1275        only in quite long subpatterns.
1276    
1277    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1278        since it was replaced by pcre_fullinfo() in February 2000.
1279    
1280    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1281        match a (*MARK), and the match failed at the start of the subject, a
1282        reference to memory before the start of the subject could occur. This bug
1283        was introduced by fix 17 of release 8.21.
1284    
1285    5.  A reference to an unset group with zero minimum repetition was giving
1286        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1287        /(another)?(\1?)test/ matched against "hello world test". This bug was
1288        introduced in release 8.13.
1289    
1290    6.  Add support for 16-bit character strings (a large amount of work involving
1291        many changes and refactorings).
1292    
1293    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1294        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1295        from a file.
1296    
1297    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1298        rounding is not applied in this particular case).
1299    
1300    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1301        if they appear, or are escaped, in patterns.
1302    
1303    10. Get rid of a number of -Wunused-but-set-variable warnings.
1304    
1305    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1306        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1307        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1308        also returns the mark "x". This bug applied to capturing parentheses,
1309        non-capturing parentheses, and atomic parentheses. It also applied to some
1310        assertions.
1311    
1312    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1313        information out of configure.ac instead of relying on pcre.h.generic, which
1314        is not stored in the repository.
1315    
1316    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1317        -lreadline.
1318    
1319    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1320    
1321    15. Applied Graycode's patch to put the top-level frame on the stack rather
1322        than the heap when not using the stack for recursion. This gives a
1323        performance improvement in many cases when recursion is not deep.
1324    
1325    16. Experimental code added to "pcretest -C" to output the stack frame size.
1326    
1327    
1328    Version 8.21 12-Dec-2011
1329    ------------------------
1330    
1331    1.  Updating the JIT compiler.
1332    
1333    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1334        are added as well.
1335    
1336    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1337        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1338        calling _pcre_jit_exec. Some extra comments are added.
1339    
1340    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1341        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1342        was introduced by change 18 for 8.20.
1343    
1344    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1345        ECMA-262 standard.
1346    
1347    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1348        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1349        This bug was probably introduced by change 9 of 8.13.
1350    
1351    7.  While fixing 6 above, I noticed that a number of other items were being
1352        incorrectly rejected as "not fixed length". This arose partly because newer
1353        opcodes had not been added to the fixed-length checking code. I have (a)
1354        corrected the bug and added tests for these items, and (b) arranged for an
1355        error to occur if an unknown opcode is encountered while checking for fixed
1356        length instead of just assuming "not fixed length". The items that were
1357        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1358        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1359        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1360    
1361    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1362        being incorrectly compiled and would have given unpredicatble results.
1363    
1364    9.  A possessively repeated subpattern with minimum repeat count greater than
1365        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1366        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1367        the first (A) could occur when it should not.
1368    
1369    10. Add a cast and remove a redundant test from the code.
1370    
1371    11. JIT should use pcre_malloc/pcre_free for allocation.
1372    
1373    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1374        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1375        is anything other than /usr, -L is still shown).
1376    
1377    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1378    
1379    14. Perl does not support \N without a following name in a [] class; PCRE now
1380        also gives an error.
1381    
1382    15. If a forward reference was repeated with an upper limit of around 2000,
1383        it caused the error "internal error: overran compiling workspace". The
1384        maximum number of forward references (including repeats) was limited by the
1385        internal workspace, and dependent on the LINK_SIZE. The code has been
1386        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1387        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1388        of around 200,000 forward references. While doing this, I also speeded up
1389        the filling in of repeated forward references.
1390    
1391    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1392        incorrectly expecting the subject to contain another "a" after the start.
1393    
1394    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1395        in the match, the SKIP should be ignored. This was not happening; instead
1396        the SKIP was being treated as NOMATCH. For patterns such as
1397        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1398        tested.
1399    
1400    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1401        now much more compatible with Perl, in particular in cases where the result
1402        is a non-match for a non-anchored pattern. For example, if
1403        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1404        "m", where previously it did not return a name. A side effect of this
1405        change is that for partial matches, the last encountered mark name is
1406        returned, as for non matches. A number of tests that were previously not
1407        Perl-compatible have been moved into the Perl-compatible test files. The
1408        refactoring has had the pleasing side effect of removing one argument from
1409        the match() function, thus reducing its stack requirements.
1410    
1411    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1412        subsequent uses of /S (without +) incorrectly behaved like /S+.
1413    
1414    21. Retrieve executable code size support for the JIT compiler and fixing
1415        some warnings.
1416    
1417    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1418        not work when the shorter character appeared right at the end of the
1419        subject string.
1420    
1421    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1422        systems.
1423    
1424    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1425        output it when the /M option is used in pcretest.
1426    
1427    25. The CheckMan script was not being included in the distribution. Also, added
1428        an explicit "perl" to run Perl scripts from the PrepareRelease script
1429        because this is reportedly needed in Windows.
1430    
1431    26. If study data was being save in a file and studying had not found a set of
1432        "starts with" bytes for the pattern, the data written to the file (though
1433        never used) was taken from uninitialized memory and so caused valgrind to
1434        complain.
1435    
1436    27. Updated RunTest.bat as provided by Sheri Pierce.
1437    
1438    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1439    
1440    29. Computation of memory usage for the table of capturing group names was
1441        giving an unnecessarily large value.
1442    
1443    
1444    Version 8.20 21-Oct-2011
1445    ------------------------
1446    
1447    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1448        a POSIX class. After further experiments with Perl, which convinced me that
1449        Perl has bugs and confusions, a closing square bracket is no longer allowed
1450        in a POSIX name. This bug also affected patterns with classes that started
1451        with full stops.
1452    
1453    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1454        captured substring, but while checking the failing first alternative,
1455        substring 1 is temporarily captured. If the output vector supplied to
1456        pcre_exec() was not big enough for this capture, the yield of the function
1457        was still zero ("insufficient space for captured substrings"). This cannot
1458        be totally fixed without adding another stack variable, which seems a lot
1459        of expense for a edge case. However, I have improved the situation in cases
1460        such as /(a)(b)x|abc/ matched against "abc", where the return code
1461        indicates that fewer than the maximum number of slots in the ovector have
1462        been set.
1463    
1464    3.  Related to (2) above: when there are more back references in a pattern than
1465        slots in the output vector, pcre_exec() uses temporary memory during
1466        matching, and copies in the captures as far as possible afterwards. It was
1467        using the entire output vector, but this conflicts with the specification
1468        that only 2/3 is used for passing back captured substrings. Now it uses
1469        only the first 2/3, for compatibility. This is, of course, another edge
1470        case.
1471    
1472    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1473        main code base, and can be used by building with --enable-jit. When this is
1474        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1475        runtime --no-jit option is given.
1476    
1477    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1478        ovector, the return from the function was zero, implying that there were
1479        other matches that did not fit. The correct "exactly full" value is now
1480        returned.
1481    
1482    6.  If a subpattern that was called recursively or as a subroutine contained
1483        (*PRUNE) or any other control that caused it to give a non-standard return,
1484        invalid errors such as "Error -26 (nested recursion at the same subject
1485        position)" or even infinite loops could occur.
1486    
1487    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1488        computing the minimum length on reaching *ACCEPT, and so ended up with the
1489        wrong value of 1 rather than 0. Further investigation indicates that
1490        computing a minimum subject length in the presence of *ACCEPT is difficult
1491        (think back references, subroutine calls), and so I have changed the code
1492        so that no minimum is registered for a pattern that contains *ACCEPT.
1493    
1494    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1495        it was not handled as intended. [But see 16 below.]
1496    
1497    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1498        Sheri Pierce.
1499    
1500    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1501        the first byte in a match must be "a".
1502    
1503    11. Change 17 for 8.13 increased the recursion depth for patterns like
1504        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1505        pattern contains any instances of (*THEN). If it does not, the old
1506        optimizations are restored. It would be nice to do this on a per-group
1507        basis, but at the moment that is not feasible.
1508    
1509    12. In some environments, the output of pcretest -C is CRLF terminated. This
1510        broke RunTest's code that checks for the link size. A single white space
1511        character after the value is now allowed for.
1512    
1513    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1514        For "fr", it uses the Windows-specific input and output files.
1515    
1516    14. If (*THEN) appeared in a group that was called recursively or as a
1517        subroutine, it did not work as intended. [But see next item.]
1518    
1519    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1520        pattern fragments (but not containing any | characters). If A and B are
1521        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1522        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1523        D. In other words, Perl considers parentheses that do not contain any |
1524        characters to be part of a surrounding alternative, whereas PCRE was
1525        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1526        differently. PCRE now behaves in the same way as Perl, except in the case
1527        of subroutine/recursion calls such as (?1) which have in any case always
1528        been different (but PCRE had them first :-).
1529    
1530    16. Related to 15 above: Perl does not treat the | in a conditional group as
1531        creating alternatives. Such a group is treated in the same way as an
1532        ordinary group without any | characters when processing (*THEN). PCRE has
1533        been changed to match Perl's behaviour.
1534    
1535    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1536        RunGrepTest script failed.
1537    
1538    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1539        inevitable for groups that contain captures, but it can lead to a lot of
1540        stack use in large patterns. The old behaviour has been restored for atomic
1541        groups that do not contain any capturing parentheses.
1542    
1543    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1544        suppress the check for a minimum subject length at run time. (If it was
1545        given to pcre_exec() or pcre_dfa_exec() it did work.)
1546    
1547    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1548        fail to work when decoding hex characters in data strings in EBCDIC
1549        environments.
1550    
1551    21. It appears that in at least one Mac OS environment, the isxdigit() function
1552        is implemented as a macro that evaluates to its argument more than once,
1553        contravening the C 90 Standard (I haven't checked a later standard). There
1554        was an instance in pcretest which caused it to go wrong when processing
1555        \x{...} escapes in subject strings. The has been rewritten to avoid using
1556        things like p++ in the argument of isxdigit().
1557    
1558    
1559    Version 8.13 16-Aug-2011
1560    ------------------------
1561    
1562    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1563    
1564    2.  Two minor typos in pcre_internal.h have been fixed.
1565    
1566    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1567        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1568        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1569    
1570    4.  There were a number of related bugs in the code for matching backrefences
1571        caselessly in UTF-8 mode when codes for the characters concerned were
1572        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1573        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1574        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1575        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1576        2-byte code at the end of the subject (it thought there wasn't enough data
1577        left).
1578    
1579    5.  Comprehensive information about what went wrong is now returned by
1580        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1581        as the output vector has at least 2 elements. The offset of the start of
1582        the failing character and a reason code are placed in the vector.
1583    
1584    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1585        now returned is for the first byte of the failing character, instead of the
1586        last byte inspected. This is an incompatible change, but I hope it is small
1587        enough not to be a problem. It makes the returned offset consistent with
1588        pcre_exec() and pcre_dfa_exec().
1589    
1590    7.  pcretest now gives a text phrase as well as the error number when
1591        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1592        failure, the offset and reason code are output.
1593    
1594    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1595        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1596        back over a single character (\n). This seems wrong (because it treated the
1597        two characters as a single entity when going forwards), conflicts with the
1598        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1599        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1600        has been changed.
1601    
1602    9.  Some internal refactoring has changed the processing so that the handling
1603        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1604        time (the PCRE_DOTALL option was changed this way some time ago: version
1605        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1606        which was always a bit of a fudge. It also means that there is one less
1607        argument for the match() function, which reduces its stack requirements
1608        slightly. This change also fixes an incompatibility with Perl: the pattern
1609        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1610    
1611    10. More internal refactoring has drastically reduced the number of recursive
1612        calls to match() for possessively repeated groups such as (abc)++ when
1613        using pcre_exec().
1614    
1615    11. While implementing 10, a number of bugs in the handling of groups were
1616        discovered and fixed:
1617    
1618        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1619        (a|)*(?1) gave a compile-time internal error.
1620        ((a|)+)+  did not notice that the outer group could match an empty string.
1621        (^a|^)+   was not marked as anchored.
1622        (.*a|.*)+ was not marked as matching at start or after a newline.
1623    
1624    12. Yet more internal refactoring has removed another argument from the match()
1625        function. Special calls to this function are now indicated by setting a
1626        value in a variable in the "match data" data block.
1627    
1628    13. Be more explicit in pcre_study() instead of relying on "default" for
1629        opcodes that mean there is no starting character; this means that when new
1630        ones are added and accidentally left out of pcre_study(), testing should
1631        pick them up.
1632    
1633    14. The -s option of pcretest has been documented for ages as being an old
1634        synonym of -m (show memory usage). I have changed it to mean "force study
1635        for every regex", that is, assume /S for every regex. This is similar to -i
1636        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1637        using it. It makes it easier to run collections of tests with and without
1638        study enabled, and thereby test pcre_study() more easily. All the standard
1639        tests are now run with and without -s (but some patterns can be marked as
1640        "never study" - see 20 below).
1641    
1642    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1643        restoration of the capturing data to the outer values was not happening
1644        correctly.
1645    
1646    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1647        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1648        pattern had matched an empty string, and so incorrectly returned a no
1649        match.
1650    
1651    17. There was optimizing code for the last branch of non-capturing parentheses,
1652        and also for the obeyed branch of a conditional subexpression, which used
1653        tail recursion to cut down on stack usage. Unfortunately, now that there is
1654        the possibility of (*THEN) occurring in these branches, tail recursion is
1655        no longer possible because the return has to be checked for (*THEN). These
1656        two optimizations have therefore been removed. [But see 8.20/11 above.]
1657    
1658    18. If a pattern containing \R was studied, it was assumed that \R always
1659        matched two bytes, thus causing the minimum subject length to be
1660        incorrectly computed because \R can also match just one byte.
1661    
1662    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1663        was incorrectly computed.
1664    
1665    20. If /S is present twice on a test pattern in pcretest input, it now
1666        *disables* studying, thereby overriding the use of -s on the command line
1667        (see 14 above). This is necessary for one or two tests to keep the output
1668        identical in both cases.
1669    
1670    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1671        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1672    
1673    22. When an atomic group that contained a capturing parenthesis was
1674        successfully matched, but the branch in which it appeared failed, the
1675        capturing was not being forgotten if a higher numbered group was later
1676        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1677        group 1 to "a", when in fact it should be unset. This applied to multi-
1678        branched capturing and non-capturing groups, repeated or not, and also to
1679        positive assertions (capturing in negative assertions does not happen
1680        in PCRE) and also to nested atomic groups.
1681    
1682    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1683        subject after a captured substring, to make it easier to tell which of a
1684        number of identical substrings has been captured.
1685    
1686    24. The way atomic groups are processed by pcre_exec() has been changed so that
1687        if they are repeated, backtracking one repetition now resets captured
1688        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1689        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1690        "aaa". Previously, it would have been "a". As part of this code
1691        refactoring, the way recursive calls are handled has also been changed.
1692    
1693    25. If an assertion condition captured any substrings, they were not passed
1694        back unless some other capturing happened later. For example, if
1695        (?(?=(a))a) was matched against "a", no capturing was returned.
1696    
1697    26. When studying a pattern that contained subroutine calls or assertions,
1698        the code for finding the minimum length of a possible match was handling
1699        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1700        group 1 called group 2 while simultaneously a separate group 2 called group
1701        1). A stack overflow occurred in this case. I have fixed this by limiting
1702        the recursion depth to 10.
1703    
1704    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1705        Fortmann. This supports explicit test numbers on the command line, and has
1706        argument validation and error reporting.
1707    
1708    28. An instance of \X with an unlimited repeat could fail if at any point the
1709        first character it looked at was a mark character.
1710    
1711    29. Some minor code refactoring concerning Unicode properties and scripts
1712        should reduce the stack requirement of match() slightly.
1713    
1714    30. Added the '=' option to pcretest to check the setting of unused capturing
1715        slots at the end of the pattern, which are documented as being -1, but are
1716        not included in the return count.
1717    
1718    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1719        compiled something random. Now it gives a compile-time error (as does
1720        Perl).
1721    
1722    32. A *MARK encountered during the processing of a positive assertion is now
1723        recorded and passed back (compatible with Perl).
1724    
1725    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1726        had alternative anchored branches, the search for a second match in a line
1727        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1728        matched the line "0102" twice. The same bug affected patterns that started
1729        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1730        twice.
1731    
1732    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1733        does, and because of capturing effects, quantifying parenthesized
1734        assertions may at times be useful. Quantifiers are now allowed for
1735        parenthesized assertions.
1736    
1737    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1738    
1739    36. \g was being checked for fancy things in a character class, when it should
1740        just be a literal "g".
1741    
1742    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1743        appearance of a nested POSIX class supersedes an apparent external class.
1744        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1745        unescaped square brackets may also appear as part of class names. For
1746        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1747        more like Perl. (But see 8.20/1 above.)
1748    
1749    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1750        was because it thought it was \N{name}, which is not supported).
1751    
1752    39. Add minix to OS list not supporting the -S option in pcretest.
1753    
1754    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1755        cannot analyze patterns in sufficient detail to catch mutual recursions
1756        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1757        subgroup is called recursively as a subpattern for a second time at the
1758        same position in the subject string. In previous releases this might have
1759        been caught by the recursion limit, or it might have run out of stack.
1760    
1761    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1762        happen only once. PCRE was, however incorrectly giving a compile time error
1763        "recursive call could loop indefinitely" because it cannot analyze the
1764        pattern in sufficient detail. The compile time test no longer happens when
1765        PCRE is compiling a conditional subpattern, but actual runaway loops are
1766        now caught at runtime (see 40 above).
1767    
1768    42. It seems that Perl allows any characters other than a closing parenthesis
1769        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1770        has been changed to be the same.
1771    
1772    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1773        as not to get warnings when autogen.sh is called. Also changed
1774        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1775    
1776    44. To help people who use pcregrep to scan files containing exceedingly long
1777        lines, the following changes have been made:
1778    
1779        (a) The default value of the buffer size parameter has been increased from
1780            8K to 20K. (The actual buffer used is three times this size.)
1781    
1782        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1783            PCRE is built.
1784    
1785        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1786            to be set at run time.
1787    
1788        (d) Numerical values in pcregrep options can be followed by K or M, for
1789            example --buffer-size=50K.
1790    
1791        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1792            given and the return code is set to 2.
1793    
1794    45. Add a pointer to the latest mark to the callout data block.
1795    
1796    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1797        partial match of an empty string instead of no match. This was specific to
1798        the use of ".".
1799    
1800    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1801        complete match instead of a partial match. This bug was dependent on both
1802        the PCRE_UTF8 and PCRE_DOTALL options being set.
1803    
1804    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1805        starting byte set, because \b was not being ignored.
1806    
1807    
1808    Version 8.12 15-Jan-2011
1809    ------------------------
1810    
1811    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1812        checks for such things as part of the documentation building process.
1813    
1814    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1815        --match-limit and --recursion-limit options (added for 8.11). In
1816        particular, this made one of the standard tests fail. (The integer value
1817        went into the wrong half of a long int.)
1818    
1819    3.  If the --colour option was given to pcregrep with -v (invert match), it
1820        did strange things, either producing crazy output, or crashing. It should,
1821        of course, ignore a request for colour when reporting lines that do not
1822        match.
1823    
1824    4.  Another pcregrep bug caused similar problems if --colour was specified with
1825        -M (multiline) and the pattern match finished with a line ending.
1826    
1827    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1828        matched in multiline mode, the following line was shown as part of the
1829        match. This seems wrong, so I have changed it.
1830    
1831    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1832        the check for further matches in the same line (so they could be coloured)
1833        to overrun the end of the current line. If another match was found, it was
1834        incorrectly shown (and then shown again when found in the next line).
1835    
1836    7.  If pcregrep was compiled under Windows, there was a reference to the
1837        function pcregrep_exit() before it was defined. I am assuming this was
1838        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1839        reported by a user. I've moved the definition above the reference.
1840    
1841    
1842    Version 8.11 10-Dec-2010
1843    ------------------------
1844    
1845    1.  (*THEN) was not working properly if there were untried alternatives prior
1846        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1847        backtracked to try for "b" instead of moving to the next alternative branch
1848        at the same level (in this case, to look for "c"). The Perl documentation
1849        is clear that when (*THEN) is backtracked onto, it goes to the "next
1850        alternative in the innermost enclosing group".
1851    
1852    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1853        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1854        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1855        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1856        (*THEN).
1857    
1858    3.  If \s appeared in a character class, it removed the VT character from
1859        the class, even if it had been included by some previous item, for example
1860        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1861        of \s, but is part of the POSIX "space" class.)
1862    
1863    4.  A partial match never returns an empty string (because you can always
1864        match an empty string at the end of the subject); however the checking for
1865        an empty string was starting at the "start of match" point. This has been
1866        changed to the "earliest inspected character" point, because the returned
1867        data for a partial match starts at this character. This means that, for
1868        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1869        (previously it gave "no match").
1870    
1871    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1872        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1873        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1874        has an implication that the given string is incomplete (because a partial
1875        match is preferred over a full match). For this reason, these items now
1876        give a partial match in this situation. [Aside: previously, the one case
1877        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1878        match rather than a full match, which was wrong by the old rules, but is
1879        now correct.]
1880    
1881    6.  There was a bug in the handling of #-introduced comments, recognized when
1882        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1883        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1884        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1885        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1886        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1887        places in pcre_compile().
1888    
1889    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1890        comments when looking ahead for named forward references to subpatterns,
1891        the only newline sequence it recognized was NL. It now handles newlines
1892        according to the set newline convention.
1893    
1894    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1895        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1896        cater for a lack of strerror(). These oversights have been fixed.
1897    
1898    9.  Added --match-limit and --recursion-limit to pcregrep.
1899    
1900    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1901    
1902    11. When the -o option was used, pcregrep was setting a return code of 1, even
1903        when matches were found, and --line-buffered was not being honoured.
1904    
1905    12. Added an optional parentheses number to the -o and --only-matching options
1906        of pcregrep.
1907    
1908    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1909        can match an empty string. The code to do it in pcretest and pcredemo
1910        needed fixing:
1911    
1912        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1913            only one byte after an empty string match just before CRLF (this case
1914            just got forgotten; "any" and "anycrlf" were OK).
1915    
1916        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1917            mode when an empty string match preceded an ASCII character followed by
1918            a non-ASCII character. (The code for advancing by one character rather
1919            than one byte was nonsense.)
1920    
1921        (c) The pcredemo.c sample program did not have any code at all to handle
1922            the cases when CRLF is a valid newline sequence.
1923    
1924    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1925        as a starting offset was within the subject string. There is now a new
1926        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1927        negative or greater than the length of the string. In order to test this,
1928        pcretest is extended to allow the setting of negative starting offsets.
1929    
1930    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1931        starting offset points to the beginning of a UTF-8 character was
1932        unnecessarily clumsy. I tidied it up.
1933    
1934    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1935        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1936    
1937    17. Nobody had reported that the --include_dir option, which was added in
1938        release 7.7 should have been called --include-dir (hyphen, not underscore)
1939        for compatibility with GNU grep. I have changed it to --include-dir, but
1940        left --include_dir as an undocumented synonym, and the same for
1941        --exclude-dir, though that is not available in GNU grep, at least as of
1942        release 2.5.4.
1943    
1944    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1945        characters from a string of bytes) have been redefined so as not to use
1946        loops, in order to improve performance in some environments. At the same
1947        time, I abstracted some of the common code into auxiliary macros to save
1948        repetition (this should not affect the compiled code).
1949    
1950    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1951        compile-time error is now given if \c is not followed by an ASCII
1952        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1953        different, and any byte value is allowed.)
1954    
1955    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1956        START_OPTIMIZE option, which is now allowed at compile time - but just
1957        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1958        to pcregrep and other applications that have no direct access to PCRE
1959        options. The new /Y option in pcretest sets this option when calling
1960        pcre_compile().
1961    
1962    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1963        back references. Groups containing recursive back references were forced to
1964        be atomic by that change, but in the case of named groups, the amount of
1965        memory required was incorrectly computed, leading to "Failed: internal
1966        error: code overflow". This has been fixed.
1967    
1968    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1969        pcretest.c, to avoid build problems in some Borland environments.
1970    
1971    
1972    Version 8.10 25-Jun-2010
1973    ------------------------
1974    
1975    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1976        THEN.
1977    
1978    2.  (*ACCEPT) was not working when inside an atomic group.
1979    
1980    3.  Inside a character class, \B is treated as a literal by default, but
1981        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1982        causes the error). The code is unchanged, but I tidied the documentation.
1983    
1984    4.  Inside a character class, PCRE always treated \R and \X as literals,
1985        whereas Perl faults them if its -w option is set. I have changed PCRE so
1986        that it faults them when PCRE_EXTRA is set.
1987    
1988    5.  Added support for \N, which always matches any character other than
1989        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1990    
1991    6.  When compiling pcregrep with newer versions of gcc which may have
1992        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1993        declared with attribute warn_unused_result" were given. Just casting the
1994        result to (void) does not stop the warnings; a more elaborate fudge is
1995        needed. I've used a macro to implement this.
1996    
1997    7.  Minor change to pcretest.c to avoid a compiler warning.
1998    
1999    8.  Added four artifical Unicode properties to help with an option to make
2000        \s etc use properties (see next item). The new properties are: Xan
2001        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2002    
2003    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2004        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2005        this option. Modified pcretest to add /W to test this facility. Added
2006        REG_UCP to make it available via the POSIX interface.
2007    
2008    10. Added --line-buffered to pcregrep.
2009    
2010    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2011        studied, and the match started with a letter with a code point greater than
2012        127 whose first byte was different to the first byte of the other case of
2013        the letter, the other case of this starting letter was not recognized
2014        (#976).
2015    
2016    12. If a pattern that was studied started with a repeated Unicode property
2017        test, for example, \p{Nd}+, there was the theoretical possibility of
2018        setting up an incorrect bitmap of starting bytes, but fortunately it could
2019        not have actually happened in practice until change 8 above was made (it
2020        added property types that matched character-matching opcodes).
2021    
2022    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2023        possible starting bytes for non-anchored patterns.
2024    
2025    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2026        \R, and also a number of cases that involve Unicode properties, both
2027        explicit and implicit when PCRE_UCP is set.
2028    
2029    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2030        input, it could crash or give wrong results if characters with values
2031        greater than 0xc0 were present in the subject string. (Detail: it assumed
2032        UTF-8 input when processing these items.)
2033    
2034    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2035        size_t is 64-bit (#991).
2036    
2037    17. Added a check for running out of memory when PCRE is compiled with
2038        --disable-stack-for-recursion (#990).
2039    
2040    18. If the last data line in a file for pcretest does not have a newline on
2041        the end, a newline was missing in the output.
2042    
2043    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2044        less than 128) in its various bitmaps. However, there is a facility for
2045        generating tables according to the current locale when PCRE is compiled. It
2046        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2047        characters, are recognized by isspace() and therefore were getting set in
2048        these tables, and indeed these tables seem to approximate to ISO 8859. This
2049        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2050        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2051        which of course cannot start UTF-8 characters. I have changed the code so
2052        that only real ASCII characters (less than 128) and the correct starting
2053        bytes for UTF-8 encodings are set for characters greater than 127 when in
2054        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2055        altogether.)
2056    
2057    20. Added the /T option to pcretest so as to be able to run tests with non-
2058        standard character tables, thus making it possible to include the tests
2059        used for 19 above in the standard set of tests.
2060    
2061    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2062        reference to a subpattern the other side of a comment that contains an
2063        opening parenthesis caused either an internal compiling error, or a
2064        reference to the wrong subpattern.
2065    
2066    
2067    Version 8.02 19-Mar-2010
2068    ------------------------
2069    
2070    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2071    
2072    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2073        configured.
2074    
2075    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2076        original author of that file, following a query about its status.
2077    
2078    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2079        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2080    
2081    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2082        quantifier applied to a forward-referencing subroutine call, could compile
2083        incorrect code or give the error "internal error: previously-checked
2084        referenced subpattern not found".
2085    
2086    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2087        variables to point to external functions. For these systems, therefore,
2088        pcre_malloc etc. are now initialized to local functions that call the
2089        relevant global functions.
2090    
2091    7.  There were two entries missing in the vectors called coptable and poptable
2092        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2093        I've fixed the data, and added a kludgy way of testing at compile time that
2094        the lengths are correct (equal to the number of opcodes).
2095    
2096    8.  Following on from 7, I added a similar kludge to check the length of the
2097        eint vector in pcreposix.c.
2098    
2099    9.  Error texts for pcre_compile() are held as one long string to avoid too
2100        much relocation at load time. To find a text, the string is searched,
2101        counting zeros. There was no check for running off the end of the string,
2102        which could happen if a new error number was added without updating the
2103        string.
2104    
2105    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2106    
2107    11. \K was not working if it appeared in an atomic group or in a group that
2108        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2109        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2110        the assertion is positive, but not if it is negative.
2111    
2112    12. Change 11 fortuitously reduced the size of the stack frame used in the
2113        "match()" function of pcre_exec.c by one pointer. Forthcoming
2114        implementation of support for (*MARK) will need an extra pointer on the
2115        stack; I have reserved it now, so that the stack frame size does not
2116        decrease.
2117    
2118    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2119        item in branch that calls a recursion is a subroutine call - as in the
2120        second branch in the above example - was incorrectly given the compile-
2121        time error "recursive call could loop indefinitely" because pcre_compile()
2122        was not correctly checking the subroutine for matching a non-empty string.
2123    
2124    14. The checks for overrunning compiling workspace could trigger after an
2125        overrun had occurred. This is a "should never occur" error, but it can be
2126        triggered by pathological patterns such as hundreds of nested parentheses.
2127        The checks now trigger 100 bytes before the end of the workspace.
2128    
2129    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2130    
2131    
2132    Version 8.01 19-Jan-2010
2133    ------------------------
2134    
2135    1.  If a pattern contained a conditional subpattern with only one branch (in
2136        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2137        computed the wrong minimum data length (which is of course zero for such
2138        subpatterns). This could cause incorrect "no match" results.
2139    
2140    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2141        the pattern is reset in the first branch, pcre_compile() failed with
2142        "internal error: code overflow at offset...". This happened only when
2143        the reset was to the original external option setting. (An optimization
2144        abstracts leading options settings into an external setting, which was the
2145        cause of this.)
2146    
2147    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2148        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2149        assertion pattern did not match (meaning that the assertion was true), it
2150        was incorrectly treated as false if the SKIP had been reached during the
2151        matching. This also applied to assertions used as conditions.
2152    
2153    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2154        assertion subpattern, including such a pattern used as a condition,
2155        unpredictable results occurred, instead of the error return
2156        PCRE_ERROR_DFA_UITEM.
2157    
2158    5.  The C++ GlobalReplace function was not working like Perl for the special
2159        situation when an empty string is matched. It now does the fancy magic
2160        stuff that is necessary.
2161    
2162    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2163        removed. (These were left over from very, very early versions of PCRE.)
2164    
2165    7.  Some cosmetic changes to the code to make life easier when compiling it
2166        as part of something else:
2167    
2168        (a) Change DEBUG to PCRE_DEBUG.
2169    
2170        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2171            called "current" as "current_branch", to prevent a collision with the
2172            Linux macro when compiled as a kernel module.
2173    
2174        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2175            prevent a collision with the Linux macro when compiled as a kernel
2176            module.
2177    
2178    8.  In pcre_compile() there are some checks for integer overflows that used to
2179        cast potentially large values to (double). This has been changed to that
2180        when building, a check for int64_t is made, and if it is found, it is used
2181        instead, thus avoiding the use of floating point arithmetic. (There is no
2182        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2183        double.
2184    
2185    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2186        2005 (difference between two addresses compared to an unsigned value).
2187    
2188    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2189        custom one, because of the following reported problem in Windows:
2190    
2191          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2192              under Win32.
2193          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2194              therefore missing the function definition.
2195          - The compiler thus generates a "C" signature for the test function.
2196          - The linker fails to find the "C" function.
2197          - PCRE fails to configure if asked to do so against libbz2.
2198    
2199    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2200        messages were output:
2201    
2202          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2203          rerunning libtoolize, to keep the correct libtool macros in-tree.
2204          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2205    
2206        I have done both of these things.
2207    
2208    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2209        most of the time, it *can* run out if it is given a pattern that contains a
2210        runaway infinite recursion. I updated the discussion in the pcrestack man
2211        page.
2212    
2213    13. Now that we have gone to the x.xx style of version numbers, the minor
2214        version may start with zero. Using 08 or 09 is a bad idea because users
2215        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2216        interpreted as invalid octal numbers. I've updated the previous comment in
2217        configure.ac, and also added a check that gives an error if 08 or 09 are
2218        used.
2219    
2220    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2221        causing partial matching to fail when the end of the subject matched \W
2222        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2223    
2224    15. There were some discrepancies between the declarations in pcre_internal.h
2225        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2226        their definitions. The declarations used "const uschar *" and the
2227        definitions used USPTR. Even though USPTR is normally defined as "const
2228        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2229        reported that: "This difference in casting confuses some C++ compilers, for
2230        example, SunCC recognizes above declarations as different functions and
2231        generates broken code for hbpcre." I have changed the declarations to use
2232        USPTR.
2233    
2234    16. GNU libtool is named differently on some systems. The autogen.sh script now
2235        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2236        (FreeBSD).
2237    
2238    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2239        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2240        comment: "Figure out how to create a longlong from a string: strtoll and
2241        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2242        instance, but it only takes 2 args instead of 3!"
2243    
2244    18. A subtle bug concerned with back references has been fixed by a change of
2245        specification, with a corresponding code fix. A pattern such as
2246        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2247        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2248        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2249        same bug. Such groups have to be quantified to be useful, or contained
2250        inside another quantified group. (If there's no repetition, the reference
2251        can never match.) The problem arises because, having left the group and
2252        moved on to the rest of the pattern, a later failure that backtracks into
2253        the group uses the captured value from the final iteration of the group
2254        rather than the correct earlier one. I have fixed this in PCRE by forcing
2255        any group that contains a reference to itself to be an atomic group; that
2256        is, there cannot be any backtracking into it once it has completed. This is
2257        similar to recursive and subroutine calls.
2258    
2259    
2260    Version 8.00 19-Oct-09
2261    ----------------------
2262    
2263    1.  The table for translating pcre_compile() error codes into POSIX error codes
2264        was out-of-date, and there was no check on the pcre_compile() error code
2265        being within the table. This could lead to an OK return being given in
2266        error.
2267    
2268    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2269        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2270        in a Windows environment.
2271    
2272    3.  The pcregrep --count option prints the count for each file even when it is
2273        zero, as does GNU grep. However, pcregrep was also printing all files when
2274        --files-with-matches was added. Now, when both options are given, it prints
2275        counts only for those files that have at least one match. (GNU grep just
2276        prints the file name in this circumstance, but including the count seems
2277        more useful - otherwise, why use --count?) Also ensured that the
2278        combination -clh just lists non-zero counts, with no names.
2279    
2280    4.  The long form of the pcregrep -F option was incorrectly implemented as
2281        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2282        but it seems right to fix it, and I didn't think it was worth preserving
2283        the old behaviour.
2284    
2285    5.  The command line items --regex=pattern and --regexp=pattern were not
2286        recognized by pcregrep, which required --regex pattern or --regexp pattern
2287        (with a space rather than an '='). The man page documented the '=' forms,
2288        which are compatible with GNU grep; these now work.
2289    
2290    6.  No libpcreposix.pc file was created for pkg-config; there was just
2291        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2292    
2293    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2294        when UCP support is not needed, by modifying the Python script that
2295        generates it from Unicode data files. This should not matter if the module
2296        is correctly used as a library, but I received one complaint about 50K of
2297        unwanted data. My guess is that the person linked everything into his
2298        program rather than using a library. Anyway, it does no harm.
2299    
2300    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2301        was a minimum greater than 1 for a wide character in a possessive
2302        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2303        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2304        character. Chaos in the form of incorrect output or a compiling loop could
2305        result.
2306    
2307    9.  The restrictions on what a pattern can contain when partial matching is
2308        requested for pcre_exec() have been removed. All patterns can now be
2309        partially matched by this function. In addition, if there are at least two
2310        slots in the offset vector, the offset of the earliest inspected character
2311        for the match and the offset of the end of the subject are set in them when
2312        PCRE_ERROR_PARTIAL is returned.
2313    
2314    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2315        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2316        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2317        and may be more useful for multi-segment matching.
2318    
2319    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2320        used to be given if ever the end of the subject was reached; now it is
2321        given only if matching could not proceed because another character was
2322        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2323        string "Z", which now yields "no match" instead of "partial match". In the
2324        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2325        final character ended with (*FAIL).
2326    
2327    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2328        if the pattern had a "must contain" character that was already found in the
2329        earlier partial match, unless partial matching was again requested. For
2330        example, with the pattern /dog.(body)?/, the "must contain" character is
2331        "g". If the first part-match was for the string "dog", restarting with
2332        "sbody" failed. This bug has been fixed.
2333    
2334    13. The string returned by pcre_dfa_exec() after a partial match has been
2335        changed so that it starts at the first inspected character rather than the
2336        first character of the match. This makes a difference only if the pattern
2337        starts with a lookbehind assertion or \b or \B (\K is not supported by
2338        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2339        matching functions compatible, and I think it's the right thing to do.
2340    
2341    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2342        so that the demonstration program is easily available in environments where
2343        PCRE has not been installed from source.
2344    
2345    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2346        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2347        library.
2348    
2349    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2350        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2351        is not the first non-POSIX option to be added. Clearly some people find
2352        these options useful.
2353    
2354    17. If a caller to the POSIX matching function regexec() passes a non-zero
2355        value for nmatch with a NULL value for pmatch, the value of
2356        nmatch is forced to zero.
2357    
2358    18. RunGrepTest did not have a test for the availability of the -u option of
2359        the diff command, as RunTest does. It now checks in the same way as
2360        RunTest, and also checks for the -b option.
2361    
2362    19. If an odd number of negated classes containing just a single character
2363        interposed, within parentheses, between a forward reference to a named
2364        subpattern and the definition of the subpattern, compilation crashed with
2365        an internal error, complaining that it could not find the referenced
2366        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2367        [The bug was that it was starting one character too far in when skipping
2368        over the character class, thus treating the ] as data rather than
2369        terminating the class. This meant it could skip too much.]
2370    
2371    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2372        /g option in pcretest when the pattern contains \K, which makes it possible
2373        to have an empty string match not at the start, even when the pattern is
2374        anchored. Updated pcretest and pcredemo to use this option.
2375    
2376    21. If the maximum number of capturing subpatterns in a recursion was greater
2377        than the maximum at the outer level, the higher number was returned, but
2378        with unset values at the outer level. The correct (outer level) value is
2379        now given.
2380    
2381    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2382        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2383        make it do so. The string so far is captured, making this feature
2384        compatible with Perl.
2385    
2386    23. The tests have been re-organized, adding tests 11 and 12, to make it
2387        possible to check the Perl 5.10 features against Perl 5.10.
2388    
2389    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2390        pattern matches a fixed length string. PCRE did not allow this; now it
2391        does. Neither allows recursion.
2392    
2393    25. I finally figured out how to implement a request to provide the minimum
2394        length of subject string that was needed in order to match a given pattern.
2395        (It was back references and recursion that I had previously got hung up
2396        on.) This code has now been added to pcre_study(); it finds a lower bound
2397        to the length of subject needed. It is not necessarily the greatest lower
2398        bound, but using it to avoid searching strings that are too short does give
2399        some useful speed-ups. The value is available to calling programs via
2400        pcre_fullinfo().
2401    
2402    26. While implementing 25, I discovered to my embarrassment that pcretest had
2403        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2404        study optimizations had never been tested with that matching function.
2405        Oops. What is worse, even when it was passed study data, there was a bug in
2406        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2407        were also very few tests of studied patterns with pcre_dfa_exec().
2408    
2409    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2410        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2411        on the other side of the coin, they are no longer allowed to have different
2412        names, because these cannot be distinguished in PCRE, and this has caused
2413        confusion. (This is a difference from Perl.)
2414    
2415    28. When duplicate subpattern names are present (necessarily with different
2416        numbers, as required by 27 above), and a test is made by name in a
2417        conditional pattern, either for a subpattern having been matched, or for
2418        recursion in such a pattern, all the associated numbered subpatterns are
2419        tested, and the overall condition is true if the condition is true for any
2420        one of them. This is the way Perl works, and is also more like the way
2421        testing by number works.
2422    
2423    
2424    Version 7.9 11-Apr-09
2425    ---------------------
2426    
2427    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2428        (pcretest), all targets were linked against these libraries. This included
2429        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2430        libraries. This caused unwanted dependencies to be created. This problem
2431        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2432        pcretest is linked with readline.
2433    
2434    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2435        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2436        moved outside it again, because FALSE and TRUE are already defined in AIX,
2437        but BOOL is not.
2438    
2439    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2440        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2441    
2442    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2443        hyphens) following file names and line numbers when outputting matching
2444        lines. This is not true; no spaces are inserted. I have also clarified the
2445        wording for the --colour (or --color) option.
2446    
2447    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2448        was not coloured; this is different to GNU grep, so I have changed it to be
2449        the same.
2450    
2451    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2452        each matching line was coloured. Now it goes on to look for further matches
2453        of any of the test patterns, which is the same behaviour as GNU grep.
2454    
2455    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2456        doesn't make sense to accept an empty string match in pcregrep, so I have
2457        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2458        seems to be how GNU grep behaves. [But see later change 40 for release
2459        8.33.]
2460    
2461    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2462        start or after a newline", because the conditional assertion was not being
2463        correctly handled. The rule now is that both the assertion and what follows
2464        in the first alternative must satisfy the test.
2465    
2466    9.  If auto-callout was enabled in a pattern with a conditional group whose
2467        condition was an assertion, PCRE could crash during matching, both with
2468        pcre_exec() and pcre_dfa_exec().
2469    
2470    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2471        used for matching.
2472    
2473    11. Unicode property support in character classes was not working for
2474        characters (bytes) greater than 127 when not in UTF-8 mode.
2475    
2476    12. Added the -M command line option to pcretest.
2477    
2478    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2479    
2480    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2481    
2482    16. Added comments and documentation about mis-use of no_arg in the C++
2483        wrapper.
2484    
2485    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2486        from Martin Jerabek that uses macro names for all relevant character and
2487        string constants.
2488    
2489    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2490        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2491        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2492        these, but not everybody uses configure.
2493    
2494    19. A conditional group that had only one branch was not being correctly
2495        recognized as an item that could match an empty string. This meant that an
2496        enclosing group might also not be so recognized, causing infinite looping
2497        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2498        with the subject "ab", where knowledge that the repeated group can match
2499        nothing is needed in order to break the loop.
2500    
2501    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2502        exec(), but without supplying a callout function, matching went wrong.
2503    
2504    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2505        leak if the size of the offset vector was greater than 30. When the vector
2506        is smaller, the saved offsets during recursion go onto a local stack
2507        vector, but for larger vectors malloc() is used. It was failing to free
2508        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2509        error, in fact).
2510    
2511    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2512        heapframe that is used only when UTF-8 support is enabled. This caused no
2513        problem, but was untidy.
2514    
2515    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2516        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2517        included within another project.
2518    
2519    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2520        slightly modified by me:
2521    
2522          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2523              not building pcregrep.
2524    
2525          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2526              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2527    
2528    25. Forward references, both numeric and by name, in patterns that made use of
2529        duplicate group numbers, could behave incorrectly or give incorrect errors,
2530        because when scanning forward to find the reference group, PCRE was not
2531        taking into account the duplicate group numbers. A pattern such as
2532        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2533    
2534    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2535        the feature of a custom pointer more persuasive (as requested by a user).
2536    
2537    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2538        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2539        pre-defined.
2540    
2541    28. Added support for (*UTF8) at the start of a pattern.
2542    
2543    29. Arrange for flags added by the "release type" setting in CMake to be shown
2544        in the configuration summary.
2545    
2546    
2547    Version 7.8 05-Sep-08
2548    ---------------------
2549    
2550    1.  Replaced UCP searching code with optimized version as implemented for Ad
2551        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2552        stage table and inline lookup instead of a function, giving speed ups of 2
2553        to 5 times on some simple patterns that I tested. Permission was given to
2554        distribute the MultiStage2.py script that generates the tables (it's not in
2555        the tarball, but is in the Subversion repository).
2556    
2557    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2558        scripts.
2559    
2560    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2561        a group with a zero qualifier. The result of the study could be incorrect,
2562        or the function might crash, depending on the pattern.
2563    
2564    4.  Caseless matching was not working for non-ASCII characters in back
2565        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2566        It now works when Unicode Property Support is available.
2567    
2568    5.  In pcretest, an escape such as \x{de} in the data was always generating
2569        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2570        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2571        truncation.
2572    
2573    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2574    
2575    7.  Added two (int) casts to pcregrep when printing the difference of two
2576        pointers, in case they are 64-bit values.
2577    
2578    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2579        test 2 if it fails.
2580    
2581    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2582        and a #define of that name to empty if it is not externally set. This is to
2583        allow users of MSVC to set it if necessary.
2584    
2585    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2586        the convenience functions in the pcre_get.c source file.
2587    
2588    11. An option change at the start of a pattern that had top-level alternatives
2589        could cause overwriting and/or a crash. This command provoked a crash in
2590        some environments:
2591    
2592          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2593    
2594        This potential security problem was recorded as CVE-2008-2371.
2595    
2596    12. For a pattern where the match had to start at the beginning or immediately
2597        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2598        pcre_dfa_exec() could read past the end of the passed subject if there was
2599        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2600        pcretest so that it places the subject at the end of its malloc-ed buffer.
2601    
2602    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2603        exec() might read past the end of the data buffer in UTF-8 mode.
2604    
2605    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2606        the data contained the byte 0x85 as part of a UTF-8 character within its
2607        first line. This applied both to normal and DFA matching.
2608    
2609    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2610        /^[^d]*?$/8 failed to match "abc".
2611    
2612    16. Added a missing copyright notice to pcrecpp_internal.h.
2613    
2614    17. Make it more clear in the documentation that values returned from
2615        pcre_exec() in ovector are byte offsets, not character counts.
2616    
2617    18. Tidied a few places to stop certain compilers from issuing warnings.
2618    
2619    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2620        supplied by Stefan Weber. I made a further small update for 7.8 because
2621        there is a change of source arrangements: the pcre_searchfuncs.c module is
2622        replaced by pcre_ucd.c.
2623    
2624    
2625    Version 7.7 07-May-08
2626    ---------------------
2627    
2628    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2629        a string to a long long, pretend we don't even have a long long." This is
2630        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2631    
2632    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2633        pre-7.6 versions, which defined a global no_arg variable instead of putting
2634        it in the RE class. (See also #8 below.)
2635    
2636    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2637        Lopes.
2638    
2639    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2640    
2641        (1) The include/exclude patterns were being applied to the whole pathnames
2642            of files, instead of just to the final components.
2643    
2644        (2) If there was more than one level of directory, the subdirectories were
2645            skipped unless they satisfied the include/exclude conditions. This is
2646            inconsistent with GNU grep (and could even be seen as contrary to the
2647            pcregrep specification - which I improved to make it absolutely clear).
2648            The action now is always to scan all levels of directory, and just
2649            apply the include/exclude patterns to regular files.
2650    
2651    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2652        --exclude_dir in the tests to avoid scanning .svn directories.
2653    
2654    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2655        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2656        doesn't support NULs in patterns.
2657    
2658    7.  Added some missing "const"s to declarations of static tables in
2659        pcre_compile.c and pcre_dfa_exec.c.
2660    
2661    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2662        caused by fix #2  above. (Subsequently also a second patch to fix the
2663        first patch. And a third patch - this was a messy problem.)
2664    
2665    9.  Applied Craig's patch to remove the use of push_back().
2666    
2667    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2668        matching function regexec().
2669    
2670    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2671        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2672        references. PCRE supports relative numbers with this syntax (I don't think
2673        Oniguruma does).
2674    
2675    12. Previously, a group with a zero repeat such as (...){0} was completely
2676        omitted from the compiled regex. However, this means that if the group
2677        was called as a subroutine from elsewhere in the pattern, things went wrong
2678        (an internal error was given). Such groups are now left in the compiled
2679        pattern, with a new opcode that causes them to be skipped at execution
2680        time.
2681    
2682    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2683        to the way PCRE behaves:
2684    
2685        (a) A lone ] character is dis-allowed (Perl treats it as data).
2686    
2687        (b) A back reference to an unmatched subpattern matches an empty string
2688            (Perl fails the current match path).
2689    
2690        (c) A data ] in a character class must be notated as \] because if the
2691            first data character in a class is ], it defines an empty class. (In
2692            Perl it is not possible to have an empty class.) The empty class []
2693            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2694            The negative empty class [^] matches any one character, independently
2695            of the DOTALL setting.
2696    
2697    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2698        non-existent subpattern following a character class starting with ']' and
2699        containing () gave an internal compiling error instead of "reference to
2700        non-existent subpattern". Fortunately, when the pattern did exist, the
2701        compiled code was correct. (When scanning forwards to check for the
2702        existence of the subpattern, it was treating the data ']' as terminating
2703        the class, so got the count wrong. When actually compiling, the reference
2704        was subsequently set up correctly.)
2705    
2706    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2707        it was being rejected as not supported by pcre_dfa_exec(), even though
2708        other assertions are supported. I have made pcre_dfa_exec() support
2709        (*FAIL).
2710    
2711    16. The implementation of 13c above involved the invention of a new opcode,
2712        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2713        cannot be changed at match time, I realized I could make a small
2714        improvement to matching performance by compiling OP_ALLANY instead of
2715        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2716        on the OP_ANY path.
2717    
2718    17. Compiling pcretest on Windows with readline support failed without the
2719        following two fixes: (1) Make the unistd.h include conditional on
2720        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2721    
2722    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2723        ncurses library to be included for pcretest when ReadLine support is
2724        requested, but also to allow for it to be overridden. This patch came from
2725        Daniel Bergström.
2726    
2727    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2728        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2729        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2730        spotting this.
2731    
2732    
2733    Version 7.6 28-Jan-08
2734    ---------------------
2735    
2736    1.  A character class containing a very large number of characters with
2737        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2738        overflow.
2739    
2740    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2741        HAVE_LONG_LONG is not defined.
2742    
2743    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2744        bring them up to date and include new features. This patch includes:
2745    
2746        - Fixed PH's badly added libz and libbz2 support.
2747        - Fixed a problem with static linking.
2748        - Added pcredemo. [But later removed - see 7 below.]
2749        - Fixed dftables problem and added an option.
2750        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2751            HAVE_LONG_LONG.
2752        - Added readline support for pcretest.
2753        - Added an listing of the option settings after cmake has run.
2754    
2755    4.  A user submitted a patch to Makefile that makes it easy to create
2756        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2757        Makefile.am that cause it to include this special target, without
2758        affecting anything else. Note that the same mingw target plus all
2759        the other distribution libraries and programs are now supported
2760        when configuring with CMake (see 6 below) instead of with
2761        Configure/Make.
2762    
2763    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2764        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2765        exported in the Windows port". It has not yet been confirmed that the patch
2766        solves the problem, but it does no harm.
2767    
2768    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2769        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2770        with CMake, and also correct the comment about stack recursion.
2771    
2772    7.  Remove the automatic building of pcredemo from the ./configure system and
2773        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2774        of a program that users should build themselves after PCRE is installed, so
2775        building it automatically is not really right. What is more, it gave
2776        trouble in some build environments.
2777    
2778    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2779    
2780    
2781    Version 7.5 10-Jan-08
2782    ---------------------
2783    
2784    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2785        values in parens when parsing an RE using the C++ wrapper."
2786    
2787    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2788        Characters greater than 255 were excluded from the class instead of being
2789        included.
2790    
2791    3.  The same bug as (2) above applied to negated POSIX classes such as
2792        [:^space:].
2793    
2794    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2795        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2796        I have changed it.
2797    
2798    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2799        first named subpattern) and a construct such as (?&a) would reference the
2800        first named subpattern whose name started with "a" (in other words, the
2801        length check was missing). Both these problems are fixed. "Subpattern name
2802        expected" is now given for (?&) (a zero-length name), and this patch also
2803        makes it give the same error for \k'' (previously it complained that that
2804        was a reference to a non-existent subpattern).
2805    
2806    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2807        this is right because (?- can be followed by option settings as well as by
2808        digits. I have, however, made the messages clearer.
2809    
2810    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2811        than the number used in the conditional) now cause a compile-time error.
2812        This is actually not compatible with Perl, which accepts such patterns, but
2813        treats the conditional as always being FALSE (as PCRE used to), but it
2814        seems to me that giving a diagnostic is better.
2815    
2816    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2817        and messages.
2818    
2819    9.  Fix two occurrences of "backslash" in comments that should have been
2820        "backspace".
2821    
2822    10. Remove two redundant lines of code that can never be obeyed (their function
2823        was moved elsewhere).
2824    
2825    11. The program that makes PCRE's Unicode character property table had a bug
2826        which caused it to generate incorrect table entries for sequences of
2827        characters that have the same character type, but are in different scripts.
2828        It amalgamated them into a single range, with the script of the first of
2829        them. In other words, some characters were in the wrong script. There were
2830        thirteen such cases, affecting characters in the following ranges:
2831    
2832          U+002b0 - U+002c1
2833          U+0060c - U+0060d
2834          U+0061e - U+00612
2835          U+0064b - U+0065e
2836          U+0074d - U+0076d
2837          U+01800 - U+01805
2838          U+01d00 - U+01d77
2839          U+01d9b - U+01dbf
2840          U+0200b - U+0200f
2841          U+030fc - U+030fe
2842          U+03260 - U+0327f
2843          U+0fb46 - U+0fbb1
2844          U+10450 - U+1049d
2845    
2846    12. The -o option (show only the matching part of a line) for pcregrep was not
2847        compatible with GNU grep in that, if there was more than one match in a
2848        line, it showed only the first of them. It now behaves in the same way as
2849        GNU grep.
2850    
2851    13. If the -o and -v options were combined for pcregrep, it printed a blank
2852        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2853        does the same. The return code can be used to tell if there were any
2854        non-matching lines.
2855    
2856    14. Added --file-offsets and --line-offsets to pcregrep.
2857    
2858    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2859        infinitely looping recursion. The bug was that positive lookaheads were not
2860        being skipped when checking for a possible empty match (negative lookaheads
2861        and both kinds of lookbehind were skipped).
2862    
2863    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2864        inclusion of <windows.h> to before rather than after the definition of
2865        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2866    
2867    17. Specifying a possessive quantifier with a specific limit for a Unicode
2868        character property caused pcre_compile() to compile bad code, which led at
2869        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2870        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2871        caused the error; without that there was no problem.
2872    
2873    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2874    
2875    19. Added --enable-pcretest-libreadline.
2876    
2877    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2878        RE::GlobalReplace(). As a result, the number of replacements returned was
2879        double what it should be. I removed one of the increments, but Craig sent a
2880        later patch that removed the other one (the right fix) and added unit tests
2881        that check the return values (which was not done before).
2882    
2883    21. Several CMake things:
2884    
2885        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2886            the names libpcre and libpcreposix, not just pcre and pcreposix.
2887    
2888        (2) The above change means that pcretest and pcregrep are now correctly
2889            linked with the newly-built libraries, not previously installed ones.
2890    
2891        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2892    
2893    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2894        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2895        UTF-8 newline character). The key issue is that the pattern starts .*;
2896        this means that the match must be either at the beginning, or after a
2897        newline. The bug was in the code for advancing after a failed match and
2898        checking that the new position followed a newline. It was not taking
2899        account of UTF-8 characters correctly.
2900    
2901    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2902        character classes. PCRE was not treating the sequence [:...:] as a
2903        character class unless the ... were all letters. Perl, however, seems to
2904        allow any characters between [: and :], though of course it rejects as
2905        unknown any "names" that contain non-letters, because all the known class
2906        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2907        for example, whereas PCRE did not - it did not recognize a POSIX character
2908        class. This seemed a bit dangerous, so the code has been changed to be
2909        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2910        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2911        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2912        Perl does, and where it didn't before.
2913    
2914    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2915        Windows environments %n is disabled by default.
2916    
2917    
2918    Version 7.4 21-Sep-07
2919    ---------------------
2920    
2921    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2922        means that a class such as [\s] counted as "explicit reference to CR or
2923        LF". That isn't really right - the whole point of the change was to try to
2924        help when there was an actual mention of one of the two characters. So now
2925        the change happens only if \r or \n (or a literal CR or LF) character is
2926        encountered.
2927    
2928    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2929        of both had grown to the point where there were only 3 bits left.
2930        Fortunately, there was spare space in the data structure, and so I have
2931        moved the internal flags into a new 16-bit field to free up more option
2932        bits.
2933    
2934    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2935        but did not set the internal JCHANGED flag - either of these is enough to
2936        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2937        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2938        start sets both bits.
2939    
2940    4.  Added options (at build time, compile time, exec time) to change \R from
2941        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2942    
2943    5.  doc/pcresyntax.html was missing from the distribution.
2944    
2945    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2946        compatibility, even though it is no longer used.
2947    
2948    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2949        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2950        windows.h file is present (where different names are used). [This was
2951        reversed later after testing - see 16 below.]
2952    
2953    8.  Changed all #include <config.h> to #include "config.h". There were also
2954        some further <pcre.h> cases that I changed to "pcre.h".
2955    
2956    9.  When pcregrep was used with the --colour option, it missed the line ending
2957        sequence off the lines that it output.
2958    
2959    10. It was pointed out to me that arrays of string pointers cause lots of
2960        relocations when a shared library is dynamically loaded. A technique of
2961        using a single long string with a table of offsets can drastically reduce
2962        these. I have refactored PCRE in four places to do this. The result is
2963        dramatic:
2964    
2965          Originally:                          290
2966          After changing UCP table:            187
2967          After changing error message table:   43
2968          After changing table of "verbs"       36
2969          After changing table of Posix names   22
2970    
2971        Thanks to the folks working on Gregex for glib for this insight.
2972    
2973    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2974        unicode-properties was also set.
2975    
2976    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2977    
2978    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2979        checked only for CRLF.
2980    
2981    14. Added casts to pcretest.c to avoid compiler warnings.
2982    
2983    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2984    
2985    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2986        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2987        entirely. This removes changes made in 7 above.
2988    
2989    17. The CMake files have been updated, and there is now more information about
2990        building with CMake in the NON-UNIX-USE document.
2991    
2992    
2993    Version 7.3 28-Aug-07
2994  ---------------------  ---------------------
2995    
2996   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 3018  Version 7.3 05-Jul-07
3018   4. Debugging: adjusted the names of opcodes for different kinds of parentheses   4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3019      in debug output.      in debug output.
3020    
3021   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
3022      unittest when running under MinGW.      long printing in the pcrecpp unittest when running under MinGW.
3023    
3024   6. ESC_K was left out of the EBCDIC table.   6. ESC_K was left out of the EBCDIC table.
3025    
# Line 53  Version 7.3 05-Jul-07 Line 3042  Version 7.3 05-Jul-07
3042      pcreposix and pcrecpp parts of the library. These were overlooked when this      pcreposix and pcrecpp parts of the library. These were overlooked when this
3043      problem was solved for the main library.      problem was solved for the main library.
3044    
3045     9. There were some crude static tests to avoid integer overflow when computing
3046        the size of patterns that contain repeated groups with explicit upper
3047        limits. As the maximum quantifier is 65535, the maximum group length was
3048        set at 30,000 so that the product of these two numbers did not overflow a
3049        32-bit integer. However, it turns out that people want to use groups that
3050        are longer than 30,000 bytes (though not repeat them that many times).
3051        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3052        made it possible to implement the integer overflow checks in a much more
3053        dynamic way, which I have now done. The artificial limitation on group
3054        length has been removed - we now have only the limit on the total length of
3055        the compiled pattern, which depends on the LINK_SIZE setting.
3056    
3057    10. Fixed a bug in the documentation for get/copy named substring when
3058        duplicate names are permitted. If none of the named substrings are set, the
3059        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3060        empty string.
3061    
3062    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3063        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3064        because the ] is interpreted as the first data character and the
3065        terminating ] is not found. PCRE has been made compatible with Perl in this
3066        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3067        cause memory overwriting.
3068    
3069    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3070        string has been matched (to stop an infinite loop). It was not recognizing
3071        a conditional subpattern that could match an empty string if that
3072        subpattern was within another subpattern. For example, it looped when
3073        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3074        condition was not nested. This bug has been fixed.
3075    
3076    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3077        past the start of the subject in the presence of bytes with the top bit
3078        set, for example "\x8aBCD".
3079    
3080    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3081        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3082    
3083    14. Optimized (?!) to (*FAIL).
3084    
3085    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3086        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3087        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3088        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3089        does: it's just the validity check that is more restrictive.
3090    
3091    16. Inserted checks for integer overflows during escape sequence (backslash)
3092        processing, and also fixed erroneous offset values for syntax errors during
3093        backslash processing.
3094    
3095    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3096        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3097    
3098    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3099        caused an overrun.
3100    
3101    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3102        something other than just ASCII characters) inside a group that had an
3103        unlimited repeat caused a loop at compile time (while checking to see
3104        whether the group could match an empty string).
3105    
3106    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3107        [\P{Any}] did so. (Error in the code for printing property names.)
3108    
3109    21. An orphan \E inside a character class could cause a crash.
3110    
3111    22. A repeated capturing bracket such as (A)? could cause a wild memory
3112        reference during compilation.
3113    
3114    23. There are several functions in pcre_compile() that scan along a compiled
3115        expression for various reasons (e.g. to see if it's fixed length for look
3116        behind). There were bugs in these functions when a repeated \p or \P was
3117        present in the pattern. These operators have additional parameters compared
3118        with \d, etc, and these were not being taken into account when moving along
3119        the compiled data. Specifically:
3120    
3121        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3122            length.
3123    
3124        (b) An item such as \pL+ within a repeated group could cause crashes or
3125            loops.
3126    
3127        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3128            "reference to non-existent subpattern" error.
3129    
3130        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3131    
3132    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3133        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3134    
3135    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3136    
3137    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3138        character were causing crashes (broken optimization).
3139    
3140    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3141        \p or \P) caused a compile-time loop.
3142    
3143    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3144        break. For example, the unstudied pattern [\r\n]A does not match the string
3145        "\r\nA" because change 7.0/46 below moves the current point on by two
3146        characters after failing to match at the start. However, the pattern \nA
3147        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3148        the same is true. There doesn't seem any very clean way out of this, but
3149        what I have chosen to do makes the common cases work: PCRE now takes note
3150        of whether there can be an explicit match for \r or \n anywhere in the
3151        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3152        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3153        pattern has explicit CR or LF references.
3154    
3155    29. Added (*CR) etc for changing newline setting at start of pattern.
3156    
3157    
3158  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07
3159  ---------------------  ---------------------

Legend:
Removed from v.200  
changed lines
  Added in v.1612

  ViewVC Help
Powered by ViewVC 1.1.5