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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.272  
changed lines
  Added in v.1604

  ViewVC Help
Powered by ViewVC 1.1.5