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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1595 by zherczeg, Mon Aug 24 09:49:06 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-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 xx-xxx-xxxx
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    
147    Version 8.37 28-April-2015
148    --------------------------
149    
150    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
151        for those parentheses to be closed with whatever has been captured so far.
152        However, it was failing to mark any other groups between the hightest
153        capture so far and the currrent group as "unset". Thus, the ovector for
154        those groups contained whatever was previously there. An example is the
155        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
156    
157    2.  If an assertion condition was quantified with a minimum of zero (an odd
158        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
159    
160    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
161        unrecognized modifier, a crash could occur.
162    
163    4.  An attempt to do global matching in pcretest with a zero-length ovector
164        caused a crash.
165    
166    5.  Fixed a memory leak during matching that could occur for a subpattern
167        subroutine call (recursive or otherwise) if the number of captured groups
168        that had to be saved was greater than ten.
169    
170    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
171        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
172        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
173    
174    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
175        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
176    
177    8. When a pattern is compiled, it remembers the highest back reference so that
178       when matching, if the ovector is too small, extra memory can be obtained to
179       use instead. A conditional subpattern whose condition is a check on a
180       capture having happened, such as, for example in the pattern
181       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
182       setting the highest backreference number. This mattered only if pcre_exec()
183       was called with an ovector that was too small to hold the capture, and there
184       was no other kind of back reference (a situation which is probably quite
185       rare). The effect of the bug was that the condition was always treated as
186       FALSE when the capture could not be consulted, leading to a incorrect
187       behaviour by pcre_exec(). This bug has been fixed.
188    
189    9. A reference to a duplicated named group (either a back reference or a test
190       for being set in a conditional) that occurred in a part of the pattern where
191       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
192       to be incorrectly calculated, leading to overwriting.
193    
194    10. A mutually recursive set of back references such as (\2)(\1) caused a
195        segfault at study time (while trying to find the minimum matching length).
196        The infinite loop is now broken (with the minimum length unset, that is,
197        zero).
198    
199    11. If an assertion that was used as a condition was quantified with a minimum
200        of zero, matching went wrong. In particular, if the whole group had
201        unlimited repetition and could match an empty string, a segfault was
202        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
203        assertions to be quantified, but not if they are being used as conditions,
204        so the above pattern is faulted by Perl. PCRE has now been changed so that
205        it also rejects such patterns.
206    
207    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
208        failed to allow the zero-repeat case if pcre2_exec() was called with an
209        ovector too small to capture the group.
210    
211    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
212        Red Hat Product Security:
213    
214        (a) A crash if /K and /F were both set with the option to save the compiled
215        pattern.
216    
217        (b) Another crash if the option to print captured substrings in a callout
218        was combined with setting a null ovector, for example \O\C+ as a subject
219        string.
220    
221    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
222        forward reference repeated a large (but limited) number of times within a
223        repeated outer group that has a zero minimum quantifier, caused incorrect
224        code to be compiled, leading to the error "internal error:
225        previously-checked referenced subpattern not found" when an incorrect
226        memory address was read. This bug was reported as "heap overflow",
227        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
228        CVE-2015-2325.
229    
230    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
231        call within a group that also contained a recursive back reference caused
232        incorrect code to be compiled. This bug was reported as "heap overflow",
233        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
234        number CVE-2015-2326.
235    
236    24. Computing the size of the JIT read-only data in advance has been a source
237        of various issues, and new ones are still appear unfortunately. To fix
238        existing and future issues, size computation is eliminated from the code,
239        and replaced by on-demand memory allocation.
240    
241    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
242        adjacent to the end of the range, and the range contained characters with
243        more than one other case, caused incorrect behaviour when compiled in UTF
244        mode. In that example, the range a-j was left out of the class.
245    
246    26. Fix JIT compilation of conditional blocks, which assertion
247        is converted to (*FAIL). E.g: /(?(?!))/.
248    
249    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
250        discovered by the LLVM fuzzer.
251    
252    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
253        when this assertion was used as a condition, for example (?(?!)a|b). In
254        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
255        error about an unsupported item.
256    
257    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
258        possessification code could take exponential time to complete. A recursion
259        depth limit of 1000 has been imposed to limit the resources used by this
260        optimization.
261    
262    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
263        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
264        because \S ensures they are all in the class. The code for doing this was
265        interacting badly with the code for computing the amount of space needed to
266        compile the pattern, leading to a buffer overflow. This bug was discovered
267        by the LLVM fuzzer.
268    
269    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
270        other kinds of group caused stack overflow at compile time. This bug was
271        discovered by the LLVM fuzzer.
272    
273    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
274        between a subroutine call and its quantifier was incorrectly compiled,
275        leading to buffer overflow or other errors. This bug was discovered by the
276        LLVM fuzzer.
277    
278    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
279        assertion after (?(. The code was failing to check the character after
280        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
281        was discovered by the LLVM fuzzer.
282    
283    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
284        a fixed maximum following a group that contains a subroutine reference was
285        incorrectly compiled and could trigger buffer overflow. This bug was
286        discovered by the LLVM fuzzer.
287    
288    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
289        caused a stack overflow instead of the diagnosis of a non-fixed length
290        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
291    
292    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
293        (e.g. /(?<=\Ka)/) could make pcregrep loop.
294    
295    37. There was a similar problem to 36 in pcretest for global matches.
296    
297    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
298        and a subsequent item in the pattern caused a non-match, backtracking over
299        the repeated \X did not stop, but carried on past the start of the subject,
300        causing reference to random memory and/or a segfault. There were also some
301        other cases where backtracking after \C could crash. This set of bugs was
302        discovered by the LLVM fuzzer.
303    
304    39. The function for finding the minimum length of a matching string could take
305        a very long time if mutual recursion was present many times in a pattern,
306        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
307        method has been implemented. This infelicity was discovered by the LLVM
308        fuzzer.
309    
310    40. Static linking against the PCRE library using the pkg-config module was
311        failing on missing pthread symbols.
312    
313    
314    Version 8.36 26-September-2014
315    ------------------------------
316    
317    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
318        -Wmissing-field-initializers and -Wunused-parameter.
319    
320    2.  The tests for quantifiers being too big (greater than 65535) were being
321        applied after reading the number, and stupidly assuming that integer
322        overflow would give a negative number. The tests are now applied as the
323        numbers are read.
324    
325    3.  Tidy code in pcre_exec.c where two branches that used to be different are
326        now the same.
327    
328    4.  The JIT compiler did not generate match limit checks for certain
329        bracketed expressions with quantifiers. This may lead to exponential
330        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
331        issue should be resolved now.
332    
333    5.  Fixed an issue, which occures when nested alternatives are optimized
334        with table jumps.
335    
336    6.  Inserted two casts and changed some ints to size_t in the light of some
337        reported 64-bit compiler warnings (Bugzilla 1477).
338    
339    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
340        an empty string, which sometimes were behaving incorrectly in the
341        interpreter (though correctly in the JIT matcher). This pcretest input is
342        an example:
343    
344          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
345          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
346    
347        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
348        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
349        for an empty string was breaking the inner loop and carrying on at a lower
350        level, when possessive repeated groups should always return to a higher
351        level as they have no backtrack points in them. The empty string test now
352        occurs at the outer level.
353    
354    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
355        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
356    
357    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
358        doesn't).
359    
360    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
361        to hold a pattern to be incorrectly computed (too small) when there were
362        named back references to duplicated names. This could cause "internal
363        error: code overflow" or "double free or corruption" or other memory
364        handling errors.
365    
366    11. When named subpatterns had the same prefixes, back references could be
367        confused. For example, in this pattern:
368    
369          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
370    
371        the reference to 'Name' was incorrectly treated as a reference to a
372        duplicate name.
373    
374    12. A pattern such as /^s?c/mi8 where the optional character has more than
375        one "other case" was incorrectly compiled such that it would only try to
376        match starting at "c".
377    
378    13. When a pattern starting with \s was studied, VT was not included in the
379        list of possible starting characters; this should have been part of the
380        8.34/18 patch.
381    
382    14. If a character class started [\Qx]... where x is any character, the class
383        was incorrectly terminated at the ].
384    
385    15. If a pattern that started with a caseless match for a character with more
386        than one "other case" was studied, PCRE did not set up the starting code
387        unit bit map for the list of possible characters. Now it does. This is an
388        optimization improvement, not a bug fix.
389    
390    16. The Unicode data tables have been updated to Unicode 7.0.0.
391    
392    17. Fixed a number of memory leaks in pcregrep.
393    
394    18. Avoid a compiler warning (from some compilers) for a function call with
395        a cast that removes "const" from an lvalue by using an intermediate
396        variable (to which the compiler does not object).
397    
398    19. Incorrect code was compiled if a group that contained an internal recursive
399        back reference was optional (had quantifier with a minimum of zero). This
400        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
401        caused segmentation faults because of stack overflows at compile time.
402    
403    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
404        group that is quantified with an indefinite repeat, caused a compile-time
405        loop which used up all the system stack and provoked a segmentation fault.
406        This was not the same bug as 19 above.
407    
408    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
409        Patch by Mike Frysinger.
410    
411    
412    Version 8.35 04-April-2014
413    --------------------------
414    
415    1.  A new flag is set, when property checks are present in an XCLASS.
416        When this flag is not set, PCRE can perform certain optimizations
417        such as studying these XCLASS-es.
418    
419    2.  The auto-possessification of character sets were improved: a normal
420        and an extended character set can be compared now. Furthermore
421        the JIT compiler optimizes more character set checks.
422    
423    3.  Got rid of some compiler warnings for potentially uninitialized variables
424        that show up only when compiled with -O2.
425    
426    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
427        of a match later then the end of the match. The pcretest program was not
428        handling the case sensibly - it was outputting from the start to the next
429        binary zero. It now reports this situation in a message, and outputs the
430        text from the end to the start.
431    
432    5.  Fast forward search is improved in JIT. Instead of the first three
433        characters, any three characters with fixed position can be searched.
434        Search order: first, last, middle.
435    
436    6.  Improve character range checks in JIT. Characters are read by an inprecise
437        function now, which returns with an unknown value if the character code is
438        above a certain threshold (e.g: 256). The only limitation is that the value
439        must be bigger than the threshold as well. This function is useful when
440        the characters above the threshold are handled in the same way.
441    
442    7.  The macros whose names start with RAWUCHAR are placeholders for a future
443        mode in which only the bottom 21 bits of 32-bit data items are used. To
444        make this more memorable for those maintaining the code, the names have
445        been changed to start with UCHAR21, and an extensive comment has been added
446        to their definition.
447    
448    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
449        to the export list in Makefile.am (they were accidentally omitted from the
450        8.34 tarball).
451    
452    9.  The informational output from pcretest used the phrase "starting byte set"
453        which is inappropriate for the 16-bit and 32-bit libraries. As the output
454        for "first char" and "need char" really means "non-UTF-char", I've changed
455        "byte" to "char", and slightly reworded the output. The documentation about
456        these values has also been (I hope) clarified.
457    
458    10. Another JIT related optimization: use table jumps for selecting the correct
459        backtracking path, when more than four alternatives are present inside a
460        bracket.
461    
462    11. Empty match is not possible, when the minimum length is greater than zero,
463        and there is no \K in the pattern. JIT should avoid empty match checks in
464        such cases.
465    
466    12. In a caseless character class with UCP support, when a character with more
467        than one alternative case was not the first character of a range, not all
468        the alternative cases were added to the class. For example, s and \x{17f}
469        are both alternative cases for S: the class [RST] was handled correctly,
470        but [R-T] was not.
471    
472    13. The configure.ac file always checked for pthread support when JIT was
473        enabled. This is not used in Windows, so I have put this test inside a
474        check for the presence of windows.h (which was already tested for).
475    
476    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
477        The algorithm provides a way to skip certain starting offsets, and usually
478        faster than linear prefix searches.
479    
480    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
481        as for 'fr_FR' and 'french'. For some reason, however, it then used the
482        Windows-specific input and output files, which have 'french' screwed in.
483        So this could never have worked. One of the problems with locales is that
484        they aren't always the same. I have now updated RunTest so that it checks
485        the output of the locale test (test 3) against three different output
486        files, and it allows the test to pass if any one of them matches. With luck
487        this should make the test pass on some versions of Solaris where it was
488        failing. Because of the uncertainty, the script did not used to stop if
489        test 3 failed; it now does. If further versions of a French locale ever
490        come to light, they can now easily be added.
491    
492    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
493        there was a message during ./configure, but it did not stop. This now
494        provokes an error. The invalid example in README has been corrected.
495        If a value less than the minimum is given, the minimum value has always
496        been used, but now a warning is given.
497    
498    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
499        was a bug in the test system, which is now fixed. Also, the list of various
500        configurations that are tested for each release did not have one with both
501        16/32 bits and --enable-bar-anycrlf. It now does.
502    
503    18. pcretest was missing "-C bsr" for displaying the \R default setting.
504    
505    19. Little endian PowerPC systems are supported now by the JIT compiler.
506    
507    20. The fast forward newline mechanism could enter to an infinite loop on
508        certain invalid UTF-8 input. Although we don't support these cases
509        this issue can be fixed by a performance optimization.
510    
511    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
512        not take account if existing stack usage. There is now a new global
513        variable called pcre_stack_guard that can be set to point to an external
514        function to check stack availability. It is called at the start of
515        processing every parenthesized group.
516    
517    22. A typo in the code meant that in ungreedy mode the max/min qualifier
518        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
519        not match "ab".
520    
521    23. When UTF was disabled, the JIT program reported some incorrect compile
522        errors. These messages are silenced now.
523    
524    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
525        compiler.
526    
527    25. Change all the temporary files used in RunGrepTest to be different to those
528        used by RunTest so that the tests can be run simultaneously, for example by
529        "make -j check".
530    
531    
532    Version 8.34 15-December-2013
533    -----------------------------
534    
535    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
536        executable memory. Patch inspired by Carsten Klein.
537    
538    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
539        this macro is never tested and has no effect, because the work to support
540        coverage involves only compiling and linking options and special targets in
541        the Makefile. The comment in config.h implied that defining the macro would
542        enable coverage support, which is totally false. There was also support for
543        setting this macro in the CMake files (my fault, I just copied it from
544        configure). SUPPORT_GCOV has now been removed.
545    
546    3.  Make a small performance improvement in strlen16() and strlen32() in
547        pcretest.
548    
549    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
550        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
551        these cases). There was also one in pcretest.c.
552    
553    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
554    
555    6.  In UTF mode, the code for checking whether a group could match an empty
556        string (which is used for indefinitely repeated groups to allow for
557        breaking an infinite loop) was broken when the group contained a repeated
558        negated single-character class with a character that occupied more than one
559        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
560        UTF-8 mode). The effect was undefined: the group might or might not be
561        deemed as matching an empty string, or the program might have crashed.
562    
563    7.  The code for checking whether a group could match an empty string was not
564        recognizing that \h, \H, \v, \V, and \R must match a character.
565    
566    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
567        an empty string. If it can, pcretest shows this in its information output.
568    
569    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
570        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
571        matched by pcre_exec() without using JIT:
572    
573        (a) If the rest of the pattern did not match after a maximal run of
574            grapheme clusters, the code for backing up to try with fewer of them
575            did not always back up over a full grapheme when characters that do not
576            have the modifier quality were involved, e.g. Hangul syllables.
577    
578        (b) If the match point in a subject started with modifier character, and
579            there was no match, the code could incorrectly back up beyond the match
580            point, and potentially beyond the first character in the subject,
581            leading to a segfault or an incorrect match result.
582    
583    10. A conditional group with an assertion condition could lead to PCRE
584        recording an incorrect first data item for a match if no other first data
585        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
586        first data item, and therefore matched "ca" after "c" instead of at the
587        start.
588    
589    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
590        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
591    
592    12. The source of pcregrep now includes z/OS-specific code so that it can be
593        compiled for z/OS as part of the special z/OS distribution.
594    
595    13. Added the -T and -TM options to pcretest.
596    
597    14. The code in pcre_compile.c for creating the table of named capturing groups
598        has been refactored. Instead of creating the table dynamically during the
599        actual compiling pass, the information is remembered during the pre-compile
600        pass (on the stack unless there are more than 20 named groups, in which
601        case malloc() is used) and the whole table is created before the actual
602        compile happens. This has simplified the code (it is now nearly 150 lines
603        shorter) and prepared the way for better handling of references to groups
604        with duplicate names.
605    
606    15. A back reference to a named subpattern when there is more than one of the
607        same name now checks them in the order in which they appear in the pattern.
608        The first one that is set is used for the reference. Previously only the
609        first one was inspected. This change makes PCRE more compatible with Perl.
610    
611    16. Unicode character properties were updated from Unicode 6.3.0.
612    
613    17. The compile-time code for auto-possessification has been refactored, based
614        on a patch by Zoltan Herczeg. It now happens after instead of during
615        compilation. The code is cleaner, and more cases are handled. The option
616        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
617        options in pcretest are provided to set it. It can also be set by
618        (*NO_AUTO_POSSESS) at the start of a pattern.
619    
620    18. The character VT has been added to the default ("C" locale) set of
621        characters that match \s and are generally treated as white space,
622        following this same change in Perl 5.18. There is now no difference between
623        "Perl space" and "POSIX space". Whether VT is treated as white space in
624        other locales depends on the locale.
625    
626    19. The code for checking named groups as conditions, either for being set or
627        for being recursed, has been refactored (this is related to 14 and 15
628        above). Processing unduplicated named groups should now be as fast at
629        numerical groups, and processing duplicated groups should be faster than
630        before.
631    
632    20. Two patches to the CMake build system, by Alexander Barkov:
633    
634          (1) Replace the "source" command by "." in CMakeLists.txt because
635              "source" is a bash-ism.
636    
637          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
638              without these the CMake build does not work on Solaris.
639    
640    21. Perl has changed its handling of \8 and \9. If there is no previously
641        encountered capturing group of those numbers, they are treated as the
642        literal characters 8 and 9 instead of a binary zero followed by the
643        literals. PCRE now does the same.
644    
645    22. Following Perl, added \o{} to specify codepoints in octal, making it
646        possible to specify values greater than 0777 and also making them
647        unambiguous.
648    
649    23. Perl now gives an error for missing closing braces after \x{... instead of
650        treating the string as literal. PCRE now does the same.
651    
652    24. RunTest used to grumble if an inappropriate test was selected explicitly,
653        but just skip it when running all tests. This make it awkward to run ranges
654        of tests when one of them was inappropriate. Now it just skips any
655        inappropriate tests, as it always did when running all tests.
656    
657    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
658        character types such as \d or \w, too many callouts were inserted, and the
659        data that they returned was rubbish.
660    
661    26. In UCP mode, \s was not matching two of the characters that Perl matches,
662        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
663        were matched by \h. The code has now been refactored so that the lists of
664        the horizontal and vertical whitespace characters used for \h and \v (which
665        are defined only in one place) are now also used for \s.
666    
667    27. Add JIT support for the 64 bit TileGX architecture.
668        Patch by Jiong Wang (Tilera Corporation).
669    
670    28. Possessive quantifiers for classes (both explicit and automatically
671        generated) now use special opcodes instead of wrapping in ONCE brackets.
672    
673    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
674        (because it's meaningless), this was not happening when PCRE_CASELESS was
675        set. Not wrong, but inefficient.
676    
677    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
678        properties for \w, \d, etc) is present in a test regex. Otherwise if the
679        test contains no characters greater than 255, Perl doesn't realise it
680        should be using Unicode semantics.
681    
682    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
683        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
684        does in Unicode mode.
685    
686    32. Added the "forbid" facility to pcretest so that putting tests into the
687        wrong test files can sometimes be quickly detected.
688    
689    33. There is now a limit (default 250) on the depth of nesting of parentheses.
690        This limit is imposed to control the amount of system stack used at compile
691        time. It can be changed at build time by --with-parens-nest-limit=xxx or
692        the equivalent in CMake.
693    
694    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
695        errors. Perl warns for these when in warning mode, but PCRE has no facility
696        for giving warnings.
697    
698    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
699        However, this was not working for (?!) because it is optimized to (*FAIL),
700        for which PCRE does not allow quantifiers. The optimization is now disabled
701        when a quantifier follows (?!). I can't see any use for this, but it makes
702        things uniform.
703    
704    36. Perl no longer allows group names to start with digits, so I have made this
705        change also in PCRE. It simplifies the code a bit.
706    
707    37. In extended mode, Perl ignores spaces before a + that indicates a
708        possessive quantifier. PCRE allowed a space before the quantifier, but not
709        before the possessive +. It now does.
710    
711    38. The use of \K (reset reported match start) within a repeated possessive
712        group such as (a\Kb)*+ was not working.
713    
714    40. Document that the same character tables must be used at compile time and
715        run time, and that the facility to pass tables to pcre_exec() and
716        pcre_dfa_exec() is for use only with saved/restored patterns.
717    
718    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
719        features for Builds with MSVC:
720    
721        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
722           builds). The .rc files can be used to set FileDescription and many other
723           attributes.
724    
725        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
726           This allows higher-level build scripts which want .pdb files to avoid
727           hard-coding the exact files needed."
728    
729    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
730        mean "start of word" and "end of word", respectively, as a transition aid.
731    
732    43. A minimizing repeat of a class containing codepoints greater than 255 in
733        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
734        compiled to use the heap for recursion.
735    
736    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
737        is configured.
738    
739    
740    Version 8.33 28-May-2013
741    ------------------------
742    
743    1.  Added 'U' to some constants that are compared to unsigned integers, to
744        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
745        variables that are added to signed variables, to ensure the result is
746        signed and can be negated.
747    
748    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
749        CMake config files.
750    
751    3.  Revise the creation of config.h.generic so that all boolean macros are
752        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
753        overriding via -D on the command line possible.
754    
755    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
756        to unsigned int is reported to make a quite noticeable speed difference in
757        a specific Windows environment. Testing on Linux did also appear to show
758        some benefit (and it is clearly not harmful). Also fixed the definition of
759        Xop which should be unsigned.
760    
761    5.  Related to (4), changing the definition of the intermediate variable cc
762        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
763        improvements.
764    
765    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
766        unnecessary spaces.
767    
768    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
769        and later.
770    
771    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
772    
773    9.  Optimizing fast_forward_start_bits in JIT.
774    
775    10. Adding support for callouts in JIT, and fixing some issues revealed
776        during this work. Namely:
777    
778        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
779    
780        (b) Minimum length was not checked before the matching is started.
781    
782    11. The value of capture_last that is passed to callouts was incorrect in some
783        cases when there was a capture on one path that was subsequently abandoned
784        after a backtrack. Also, the capture_last value is now reset after a
785        recursion, since all captures are also reset in this case.
786    
787    12. The interpreter no longer returns the "too many substrings" error in the
788        case when an overflowing capture is in a branch that is subsequently
789        abandoned after a backtrack.
790    
791    13. In the pathological case when an offset vector of size 2 is used, pcretest
792        now prints out the matched string after a yield of 0 or 1.
793    
794    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
795        Only supported by the JIT compiler at the moment.
796    
797    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
798    
799    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
800        offset of the starting point of the matching process, provided the offsets
801        vector is large enough.
802    
803    17. The \A escape now records a lookbehind value of 1, though its execution
804        does not actually inspect the previous character. This is to ensure that,
805        in partial multi-segment matching, at least one character from the old
806        segment is retained when a new segment is processed. Otherwise, if there
807        are no lookbehinds in the pattern, \A might match incorrectly at the start
808        of a new segment.
809    
810    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
811    
812    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
813        gives some modest performance improvement in 8-bit mode.
814    
815    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
816        be expressed in certain programming languages using Universal Character
817        Names.
818    
819    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
820        which points out that "non characters" are not "characters that may not
821        appear in Unicode strings" but rather "characters that are reserved for
822        internal use and have only local meaning".
823    
824    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
825        there was a conditional group that depended on an assertion, if the
826        assertion was false, the callout that immediately followed the alternation
827        in the condition was skipped when pcre_exec() was used for matching.
828    
829    23. Allow an explicit callout to be inserted before an assertion that is the
830        condition for a conditional group, for compatibility with automatic
831        callouts, which always insert a callout at this point.
832    
833    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
834        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
835    
836    25. (*PRUNE) is now supported by the JIT compiler.
837    
838    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
839    
840    27. Fix the case where there are two or more SKIPs with arguments that may be
841        ignored.
842    
843    28. (*SKIP) is now supported by the JIT compiler.
844    
845    29. (*THEN) is now supported by the JIT compiler.
846    
847    30. Update RunTest with additional test selector options.
848    
849    31. The way PCRE handles backtracking verbs has been changed in two ways.
850    
851        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
852        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
853        backtracking. In some cases this makes it more Perl-compatible, but Perl's
854        rather obscure rules do not always do the same thing.
855    
856        (2) Previously, backtracking verbs were confined within assertions. This is
857        no longer the case for positive assertions, except for (*ACCEPT). Again,
858        this sometimes improves Perl compatibility, and sometimes does not.
859    
860    32. A number of tests that were in test 2 because Perl did things differently
861        have been moved to test 1, because either Perl or PCRE has changed, and
862        these tests are now compatible.
863    
864    32. Backtracking control verbs are now handled in the same way in JIT and
865        interpreter.
866    
867    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
868        contained a forward subroutine reference caused a compile error.
869    
870    34. Auto-detect and optimize limited repetitions in JIT.
871    
872    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
873        blocking (*UTF) etc.
874    
875    36. In the interpreter, maximizing pattern repetitions for characters and
876        character types now use tail recursion, which reduces stack usage.
877    
878    37. The value of the max lookbehind was not correctly preserved if a compiled
879        and saved regex was reloaded on a host of different endianness.
880    
881    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
882        of the compiled pattern block, expand the flags field from 16 to 32 bits
883        because it was almost full.
884    
885    39. Try madvise first before posix_madvise.
886    
887    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
888        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
889        original change locked out all matches of empty strings. This has been
890        changed so that one match of an empty string per line is recognized.
891        Subsequent searches on the same line (for colouring or for --only-matching,
892        for example) do not recognize empty strings.
893    
894    41. Applied a user patch to fix a number of spelling mistakes in comments.
895    
896    42. Data lines longer than 65536 caused pcretest to crash.
897    
898    43. Clarified the data type for length and startoffset arguments for pcre_exec
899        and pcre_dfa_exec in the function-specific man pages, where they were
900        explicitly stated to be in bytes, never having been updated. I also added
901        some clarification to the pcreapi man page.
902    
903    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
904        a segmentation fault.
905    
906    
907    Version 8.32 30-November-2012
908    -----------------------------
909    
910    1.  Improved JIT compiler optimizations for first character search and single
911        character iterators.
912    
913    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
914        Patch by Daniel Richard G.
915    
916    3.  Single character iterator optimizations in the JIT compiler.
917    
918    4.  Improved JIT compiler optimizations for character ranges.
919    
920    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
921        Reported by Giuseppe D'Angelo.
922    
923    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
924        of a line, was being set incorrectly in cases where .* appeared inside
925        atomic brackets at the start of a pattern, or where there was a subsequent
926        *PRUNE or *SKIP.
927    
928    7.  Improved instruction cache flush for POWER/PowerPC.
929        Patch by Daniel Richard G.
930    
931    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
932        grep:
933    
934        (a) There is now no limit to the number of patterns to be matched.
935    
936        (b) An error is given if a pattern is too long.
937    
938        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
939            are now supported.
940    
941        (d) --exclude-from and --include-from (multiple use) have been added.
942    
943        (e) Exclusions and inclusions now apply to all files and directories, not
944            just to those obtained from scanning a directory recursively.
945    
946        (f) Multiple uses of -f and --file-list are now supported.
947    
948        (g) In a Windows environment, the default for -d has been changed from
949            "read" (the GNU grep default) to "skip", because otherwise the presence
950            of a directory in the file list provokes an error.
951    
952        (h) The documentation has been revised and clarified in places.
953    
954    9.  Improve the matching speed of capturing brackets.
955    
956    10. Changed the meaning of \X so that it now matches a Unicode extended
957        grapheme cluster.
958    
959    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
960        out POSIX threads when JIT support is configured.
961    
962    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
963    
964    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
965        structure could go wrong in environments where size_t is not the same size
966        as int.
967    
968    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
969        set.
970    
971    15. The EBCDIC support had decayed; later updates to the code had included
972        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
973        general tidy up of EBCDIC-related issues, and the documentation was also
974        not quite right. There is now a test that can be run on ASCII systems to
975        check some of the EBCDIC-related things (but is it not a full test).
976    
977    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
978        in a small tidy to the code.
979    
980    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
981    
982    18. If the --only-matching (-o) option in pcregrep is specified multiple
983        times, each one causes appropriate output. For example, -o1 -o2 outputs the
984        substrings matched by the 1st and 2nd capturing parentheses. A separating
985        string can be specified by --om-separator (default empty).
986    
987    19. Improving the first n character searches.
988    
989    20. Turn case lists for horizontal and vertical white space into macros so that
990        they are defined only once.
991    
992    21. This set of changes together give more compatible Unicode case-folding
993        behaviour for characters that have more than one other case when UCP
994        support is available.
995    
996        (a) The Unicode property table now has offsets into a new table of sets of
997            three or more characters that are case-equivalent. The MultiStage2.py
998            script that generates these tables (the pcre_ucd.c file) now scans
999            CaseFolding.txt instead of UnicodeData.txt for character case
1000            information.
1001    
1002        (b) The code for adding characters or ranges of characters to a character
1003            class has been abstracted into a generalized function that also handles
1004            case-independence. In UTF-mode with UCP support, this uses the new data
1005            to handle characters with more than one other case.
1006    
1007        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1008            whose other case is greater than 256 are now correctly matched
1009            caselessly. Previously, the high codepoint matched the low one, but not
1010            vice versa.
1011    
1012        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1013            of the new class addition function, using character lists defined as
1014            macros alongside the case definitions of 20 above.
1015    
1016        (e) Caseless back references now work with characters that have more than
1017            one other case.
1018    
1019        (f) General caseless matching of characters with more than one other case
1020            is supported.
1021    
1022    22. Unicode character properties were updated from Unicode 6.2.0
1023    
1024    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1025    
1026    24. Add support for 32-bit character strings, and UTF-32
1027    
1028    25. Major JIT compiler update (code refactoring and bugfixing).
1029        Experimental Sparc 32 support is added.
1030    
1031    26. Applied a modified version of Daniel Richard G's patch to create
1032        pcre.h.generic and config.h.generic by "make" instead of in the
1033        PrepareRelease script.
1034    
1035    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1036        pcre_compile.c when checking for a zero character.
1037    
1038    28. Introducing a native interface for JIT. Through this interface, the compiled
1039        machine code can be directly executed. The purpose of this interface is to
1040        provide fast pattern matching, so several sanity checks are not performed.
1041        However, feature tests are still performed. The new interface provides
1042        1.4x speedup compared to the old one.
1043    
1044    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1045        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1046        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1047    
1048    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1049        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1050        the "old" RFC 2279). Instead, it ended up passing a negative length to
1051        pcre_exec().
1052    
1053    31. Add support for GCC's visibility feature to hide internal functions.
1054    
1055    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1056        "unknown -C option" after outputting 0 or 1.
1057    
1058    33. There is now support for generating a code coverage report for the test
1059        suite in environments where gcc is the compiler and lcov is installed. This
1060        is mainly for the benefit of the developers.
1061    
1062    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1063        unaddressable using valgrind annotations, allowing valgrind to detect
1064        invalid memory accesses. This is mainly for the benefit of the developers.
1065    
1066    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1067    
1068    26. Give configure error if --enable-cpp but no C++ compiler found.
1069    
1070    
1071    Version 8.31 06-July-2012
1072    -------------------------
1073    
1074    1.  Fixing a wrong JIT test case and some compiler warnings.
1075    
1076    2.  Removed a bashism from the RunTest script.
1077    
1078    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1079        to unsigned type, result still unsigned" that was given by an MS compiler
1080        on encountering the code "-sizeof(xxx)".
1081    
1082    4.  Partial matching support is added to the JIT compiler.
1083    
1084    5.  Fixed several bugs concerned with partial matching of items that consist
1085        of more than one character:
1086    
1087        (a) /^(..)\1/ did not partially match "aba" because checking references was
1088            done on an "all or nothing" basis. This also applied to repeated
1089            references.
1090    
1091        (b) \R did not give a hard partial match if \r was found at the end of the
1092            subject.
1093    
1094        (c) \X did not give a hard partial match after matching one or more
1095            characters at the end of the subject.
1096    
1097        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1098            a partial match for the string "\r".
1099    
1100        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1101            a partial match for a CR character at the end of the subject string.
1102    
1103    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1104        running pcretest, the text "(JIT)" added to the output whenever JIT is
1105        actually used to run the match.
1106    
1107    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1108        or /S+[+] with a digit between 1 and 7.
1109    
1110    8.  OP_NOT now supports any UTF character not just single-byte ones.
1111    
1112    9.  (*MARK) control verb is now supported by the JIT compiler.
1113    
1114    10. The command "./RunTest list" lists the available tests without actually
1115        running any of them. (Because I keep forgetting what they all are.)
1116    
1117    11. Add PCRE_INFO_MAXLOOKBEHIND.
1118    
1119    12. Applied a (slightly modified) user-supplied patch that improves performance
1120        when the heap is used for recursion (compiled with --disable-stack-for-
1121        recursion). Instead of malloc and free for each heap frame each time a
1122        logical recursion happens, frames are retained on a chain and re-used where
1123        possible. This sometimes gives as much as 30% improvement.
1124    
1125    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1126        call.
1127    
1128    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1129    
1130    15. It is now possible to link pcretest with libedit as an alternative to
1131        libreadline.
1132    
1133    16. (*COMMIT) control verb is now supported by the JIT compiler.
1134    
1135    17. The Unicode data tables have been updated to Unicode 6.1.0.
1136    
1137    18. Added --file-list option to pcregrep.
1138    
1139    19. Added binary file support to pcregrep, including the -a, --binary-files,
1140        -I, and --text options.
1141    
1142    20. The madvise function is renamed for posix_madvise for QNX compatibility
1143        reasons. Fixed by Giuseppe D'Angelo.
1144    
1145    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1146    
1147    22. Changed the option for creating symbolic links for 16-bit man pages from
1148        -s to -sf so that re-installing does not cause issues.
1149    
1150    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1151    
1152    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1153        matches in certain environments (the workspace was not being correctly
1154        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1155        some of the workspace data at the beginning of a restart.
1156    
1157    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1158        was not doing so when it should - probably a typo introduced by SVN 528
1159        (change 8.10/14).
1160    
1161    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1162        \w+ when the character tables indicated that \x{c4} was a word character.
1163        There were several related cases, all because the tests for doing a table
1164        lookup were testing for characters less than 127 instead of 255.
1165    
1166    27. If a pattern contains capturing parentheses that are not used in a match,
1167        their slots in the ovector are set to -1. For those that are higher than
1168        any matched groups, this happens at the end of processing. In the case when
1169        there were back references that the ovector was too small to contain
1170        (causing temporary malloc'd memory to be used during matching), and the
1171        highest capturing number was not used, memory off the end of the ovector
1172        was incorrectly being set to -1. (It was using the size of the temporary
1173        memory instead of the true size.)
1174    
1175    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1176        ovector size, it uses memory at the end of the block that it has got.
1177    
1178    29. Check for an overlong MARK name and give an error at compile time. The
1179        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1180    
1181    30. JIT compiler update.
1182    
1183    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1184        Rill for the patch.
1185    
1186    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1187    
1188    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1189    
1190    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1191        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1192        was enabled.
1193    
1194    35. Improve JIT code generation for greedy plus quantifier.
1195    
1196    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1197        1 to "aa" instead of to an empty string. The bug affected repeated groups
1198        that could potentially match an empty string.
1199    
1200    37. Optimizing single character iterators in JIT.
1201    
1202    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1203        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1204        codepoints that are too big for the mode are faulted, and in a UTF mode,
1205        disallowed codepoints are also faulted.
1206    
1207    39. If PCRE was compiled with UTF support, in three places in the DFA
1208        matcher there was code that should only have been obeyed in UTF mode, but
1209        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1210        processing when bytes with values greater than 127 were present. In 16-bit
1211        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1212        both cases the values are those that cannot be the first data item in a UTF
1213        character. The three items that might have provoked this were recursions,
1214        possessively repeated groups, and atomic groups.
1215    
1216    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1217        and pcregrep, because some OS require shared objects to be explicitly
1218        passed to ld, causing the link step to fail if they are not.
1219    
1220    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1221        mode, patterns that started with \h* or \R* might be incorrectly matched.
1222    
1223    
1224    Version 8.30 04-February-2012
1225    -----------------------------
1226    
1227    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1228        name is defined in ctype.h.
1229    
1230    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1231        only in quite long subpatterns.
1232    
1233    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1234        since it was replaced by pcre_fullinfo() in February 2000.
1235    
1236    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1237        match a (*MARK), and the match failed at the start of the subject, a
1238        reference to memory before the start of the subject could occur. This bug
1239        was introduced by fix 17 of release 8.21.
1240    
1241    5.  A reference to an unset group with zero minimum repetition was giving
1242        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1243        /(another)?(\1?)test/ matched against "hello world test". This bug was
1244        introduced in release 8.13.
1245    
1246    6.  Add support for 16-bit character strings (a large amount of work involving
1247        many changes and refactorings).
1248    
1249    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1250        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1251        from a file.
1252    
1253    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1254        rounding is not applied in this particular case).
1255    
1256    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1257        if they appear, or are escaped, in patterns.
1258    
1259    10. Get rid of a number of -Wunused-but-set-variable warnings.
1260    
1261    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1262        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1263        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1264        also returns the mark "x". This bug applied to capturing parentheses,
1265        non-capturing parentheses, and atomic parentheses. It also applied to some
1266        assertions.
1267    
1268    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1269        information out of configure.ac instead of relying on pcre.h.generic, which
1270        is not stored in the repository.
1271    
1272    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1273        -lreadline.
1274    
1275    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1276    
1277    15. Applied Graycode's patch to put the top-level frame on the stack rather
1278        than the heap when not using the stack for recursion. This gives a
1279        performance improvement in many cases when recursion is not deep.
1280    
1281    16. Experimental code added to "pcretest -C" to output the stack frame size.
1282    
1283    
1284    Version 8.21 12-Dec-2011
1285    ------------------------
1286    
1287    1.  Updating the JIT compiler.
1288    
1289    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1290        are added as well.
1291    
1292    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1293        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1294        calling _pcre_jit_exec. Some extra comments are added.
1295    
1296    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1297        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1298        was introduced by change 18 for 8.20.
1299    
1300    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1301        ECMA-262 standard.
1302    
1303    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1304        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1305        This bug was probably introduced by change 9 of 8.13.
1306    
1307    7.  While fixing 6 above, I noticed that a number of other items were being
1308        incorrectly rejected as "not fixed length". This arose partly because newer
1309        opcodes had not been added to the fixed-length checking code. I have (a)
1310        corrected the bug and added tests for these items, and (b) arranged for an
1311        error to occur if an unknown opcode is encountered while checking for fixed
1312        length instead of just assuming "not fixed length". The items that were
1313        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1314        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1315        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1316    
1317    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1318        being incorrectly compiled and would have given unpredicatble results.
1319    
1320    9.  A possessively repeated subpattern with minimum repeat count greater than
1321        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1322        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1323        the first (A) could occur when it should not.
1324    
1325    10. Add a cast and remove a redundant test from the code.
1326    
1327    11. JIT should use pcre_malloc/pcre_free for allocation.
1328    
1329    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1330        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1331        is anything other than /usr, -L is still shown).
1332    
1333    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1334    
1335    14. Perl does not support \N without a following name in a [] class; PCRE now
1336        also gives an error.
1337    
1338    15. If a forward reference was repeated with an upper limit of around 2000,
1339        it caused the error "internal error: overran compiling workspace". The
1340        maximum number of forward references (including repeats) was limited by the
1341        internal workspace, and dependent on the LINK_SIZE. The code has been
1342        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1343        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1344        of around 200,000 forward references. While doing this, I also speeded up
1345        the filling in of repeated forward references.
1346    
1347    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1348        incorrectly expecting the subject to contain another "a" after the start.
1349    
1350    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1351        in the match, the SKIP should be ignored. This was not happening; instead
1352        the SKIP was being treated as NOMATCH. For patterns such as
1353        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1354        tested.
1355    
1356    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1357        now much more compatible with Perl, in particular in cases where the result
1358        is a non-match for a non-anchored pattern. For example, if
1359        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1360        "m", where previously it did not return a name. A side effect of this
1361        change is that for partial matches, the last encountered mark name is
1362        returned, as for non matches. A number of tests that were previously not
1363        Perl-compatible have been moved into the Perl-compatible test files. The
1364        refactoring has had the pleasing side effect of removing one argument from
1365        the match() function, thus reducing its stack requirements.
1366    
1367    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1368        subsequent uses of /S (without +) incorrectly behaved like /S+.
1369    
1370    21. Retrieve executable code size support for the JIT compiler and fixing
1371        some warnings.
1372    
1373    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1374        not work when the shorter character appeared right at the end of the
1375        subject string.
1376    
1377    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1378        systems.
1379    
1380    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1381        output it when the /M option is used in pcretest.
1382    
1383    25. The CheckMan script was not being included in the distribution. Also, added
1384        an explicit "perl" to run Perl scripts from the PrepareRelease script
1385        because this is reportedly needed in Windows.
1386    
1387    26. If study data was being save in a file and studying had not found a set of
1388        "starts with" bytes for the pattern, the data written to the file (though
1389        never used) was taken from uninitialized memory and so caused valgrind to
1390        complain.
1391    
1392    27. Updated RunTest.bat as provided by Sheri Pierce.
1393    
1394    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1395    
1396    29. Computation of memory usage for the table of capturing group names was
1397        giving an unnecessarily large value.
1398    
1399    
1400    Version 8.20 21-Oct-2011
1401    ------------------------
1402    
1403    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1404        a POSIX class. After further experiments with Perl, which convinced me that
1405        Perl has bugs and confusions, a closing square bracket is no longer allowed
1406        in a POSIX name. This bug also affected patterns with classes that started
1407        with full stops.
1408    
1409    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1410        captured substring, but while checking the failing first alternative,
1411        substring 1 is temporarily captured. If the output vector supplied to
1412        pcre_exec() was not big enough for this capture, the yield of the function
1413        was still zero ("insufficient space for captured substrings"). This cannot
1414        be totally fixed without adding another stack variable, which seems a lot
1415        of expense for a edge case. However, I have improved the situation in cases
1416        such as /(a)(b)x|abc/ matched against "abc", where the return code
1417        indicates that fewer than the maximum number of slots in the ovector have
1418        been set.
1419    
1420    3.  Related to (2) above: when there are more back references in a pattern than
1421        slots in the output vector, pcre_exec() uses temporary memory during
1422        matching, and copies in the captures as far as possible afterwards. It was
1423        using the entire output vector, but this conflicts with the specification
1424        that only 2/3 is used for passing back captured substrings. Now it uses
1425        only the first 2/3, for compatibility. This is, of course, another edge
1426        case.
1427    
1428    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1429        main code base, and can be used by building with --enable-jit. When this is
1430        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1431        runtime --no-jit option is given.
1432    
1433    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1434        ovector, the return from the function was zero, implying that there were
1435        other matches that did not fit. The correct "exactly full" value is now
1436        returned.
1437    
1438    6.  If a subpattern that was called recursively or as a subroutine contained
1439        (*PRUNE) or any other control that caused it to give a non-standard return,
1440        invalid errors such as "Error -26 (nested recursion at the same subject
1441        position)" or even infinite loops could occur.
1442    
1443    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1444        computing the minimum length on reaching *ACCEPT, and so ended up with the
1445        wrong value of 1 rather than 0. Further investigation indicates that
1446        computing a minimum subject length in the presence of *ACCEPT is difficult
1447        (think back references, subroutine calls), and so I have changed the code
1448        so that no minimum is registered for a pattern that contains *ACCEPT.
1449    
1450    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1451        it was not handled as intended. [But see 16 below.]
1452    
1453    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1454        Sheri Pierce.
1455    
1456    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1457        the first byte in a match must be "a".
1458    
1459    11. Change 17 for 8.13 increased the recursion depth for patterns like
1460        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1461        pattern contains any instances of (*THEN). If it does not, the old
1462        optimizations are restored. It would be nice to do this on a per-group
1463        basis, but at the moment that is not feasible.
1464    
1465    12. In some environments, the output of pcretest -C is CRLF terminated. This
1466        broke RunTest's code that checks for the link size. A single white space
1467        character after the value is now allowed for.
1468    
1469    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1470        For "fr", it uses the Windows-specific input and output files.
1471    
1472    14. If (*THEN) appeared in a group that was called recursively or as a
1473        subroutine, it did not work as intended. [But see next item.]
1474    
1475    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1476        pattern fragments (but not containing any | characters). If A and B are
1477        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1478        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1479        D. In other words, Perl considers parentheses that do not contain any |
1480        characters to be part of a surrounding alternative, whereas PCRE was
1481        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1482        differently. PCRE now behaves in the same way as Perl, except in the case
1483        of subroutine/recursion calls such as (?1) which have in any case always
1484        been different (but PCRE had them first :-).
1485    
1486    16. Related to 15 above: Perl does not treat the | in a conditional group as
1487        creating alternatives. Such a group is treated in the same way as an
1488        ordinary group without any | characters when processing (*THEN). PCRE has
1489        been changed to match Perl's behaviour.
1490    
1491    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1492        RunGrepTest script failed.
1493    
1494    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1495        inevitable for groups that contain captures, but it can lead to a lot of
1496        stack use in large patterns. The old behaviour has been restored for atomic
1497        groups that do not contain any capturing parentheses.
1498    
1499    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1500        suppress the check for a minimum subject length at run time. (If it was
1501        given to pcre_exec() or pcre_dfa_exec() it did work.)
1502    
1503    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1504        fail to work when decoding hex characters in data strings in EBCDIC
1505        environments.
1506    
1507    21. It appears that in at least one Mac OS environment, the isxdigit() function
1508        is implemented as a macro that evaluates to its argument more than once,
1509        contravening the C 90 Standard (I haven't checked a later standard). There
1510        was an instance in pcretest which caused it to go wrong when processing
1511        \x{...} escapes in subject strings. The has been rewritten to avoid using
1512        things like p++ in the argument of isxdigit().
1513    
1514    
1515    Version 8.13 16-Aug-2011
1516    ------------------------
1517    
1518    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1519    
1520    2.  Two minor typos in pcre_internal.h have been fixed.
1521    
1522    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1523        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1524        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1525    
1526    4.  There were a number of related bugs in the code for matching backrefences
1527        caselessly in UTF-8 mode when codes for the characters concerned were
1528        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1529        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1530        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1531        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1532        2-byte code at the end of the subject (it thought there wasn't enough data
1533        left).
1534    
1535    5.  Comprehensive information about what went wrong is now returned by
1536        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1537        as the output vector has at least 2 elements. The offset of the start of
1538        the failing character and a reason code are placed in the vector.
1539    
1540    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1541        now returned is for the first byte of the failing character, instead of the
1542        last byte inspected. This is an incompatible change, but I hope it is small
1543        enough not to be a problem. It makes the returned offset consistent with
1544        pcre_exec() and pcre_dfa_exec().
1545    
1546    7.  pcretest now gives a text phrase as well as the error number when
1547        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1548        failure, the offset and reason code are output.
1549    
1550    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1551        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1552        back over a single character (\n). This seems wrong (because it treated the
1553        two characters as a single entity when going forwards), conflicts with the
1554        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1555        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1556        has been changed.
1557    
1558    9.  Some internal refactoring has changed the processing so that the handling
1559        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1560        time (the PCRE_DOTALL option was changed this way some time ago: version
1561        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1562        which was always a bit of a fudge. It also means that there is one less
1563        argument for the match() function, which reduces its stack requirements
1564        slightly. This change also fixes an incompatibility with Perl: the pattern
1565        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1566    
1567    10. More internal refactoring has drastically reduced the number of recursive
1568        calls to match() for possessively repeated groups such as (abc)++ when
1569        using pcre_exec().
1570    
1571    11. While implementing 10, a number of bugs in the handling of groups were
1572        discovered and fixed:
1573    
1574        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1575        (a|)*(?1) gave a compile-time internal error.
1576        ((a|)+)+  did not notice that the outer group could match an empty string.
1577        (^a|^)+   was not marked as anchored.
1578        (.*a|.*)+ was not marked as matching at start or after a newline.
1579    
1580    12. Yet more internal refactoring has removed another argument from the match()
1581        function. Special calls to this function are now indicated by setting a
1582        value in a variable in the "match data" data block.
1583    
1584    13. Be more explicit in pcre_study() instead of relying on "default" for
1585        opcodes that mean there is no starting character; this means that when new
1586        ones are added and accidentally left out of pcre_study(), testing should
1587        pick them up.
1588    
1589    14. The -s option of pcretest has been documented for ages as being an old
1590        synonym of -m (show memory usage). I have changed it to mean "force study
1591        for every regex", that is, assume /S for every regex. This is similar to -i
1592        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1593        using it. It makes it easier to run collections of tests with and without
1594        study enabled, and thereby test pcre_study() more easily. All the standard
1595        tests are now run with and without -s (but some patterns can be marked as
1596        "never study" - see 20 below).
1597    
1598    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1599        restoration of the capturing data to the outer values was not happening
1600        correctly.
1601    
1602    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1603        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1604        pattern had matched an empty string, and so incorrectly returned a no
1605        match.
1606    
1607    17. There was optimizing code for the last branch of non-capturing parentheses,
1608        and also for the obeyed branch of a conditional subexpression, which used
1609        tail recursion to cut down on stack usage. Unfortunately, now that there is
1610        the possibility of (*THEN) occurring in these branches, tail recursion is
1611        no longer possible because the return has to be checked for (*THEN). These
1612        two optimizations have therefore been removed. [But see 8.20/11 above.]
1613    
1614    18. If a pattern containing \R was studied, it was assumed that \R always
1615        matched two bytes, thus causing the minimum subject length to be
1616        incorrectly computed because \R can also match just one byte.
1617    
1618    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1619        was incorrectly computed.
1620    
1621    20. If /S is present twice on a test pattern in pcretest input, it now
1622        *disables* studying, thereby overriding the use of -s on the command line
1623        (see 14 above). This is necessary for one or two tests to keep the output
1624        identical in both cases.
1625    
1626    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1627        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1628    
1629    22. When an atomic group that contained a capturing parenthesis was
1630        successfully matched, but the branch in which it appeared failed, the
1631        capturing was not being forgotten if a higher numbered group was later
1632        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1633        group 1 to "a", when in fact it should be unset. This applied to multi-
1634        branched capturing and non-capturing groups, repeated or not, and also to
1635        positive assertions (capturing in negative assertions does not happen
1636        in PCRE) and also to nested atomic groups.
1637    
1638    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1639        subject after a captured substring, to make it easier to tell which of a
1640        number of identical substrings has been captured.
1641    
1642    24. The way atomic groups are processed by pcre_exec() has been changed so that
1643        if they are repeated, backtracking one repetition now resets captured
1644        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1645        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1646        "aaa". Previously, it would have been "a". As part of this code
1647        refactoring, the way recursive calls are handled has also been changed.
1648    
1649    25. If an assertion condition captured any substrings, they were not passed
1650        back unless some other capturing happened later. For example, if
1651        (?(?=(a))a) was matched against "a", no capturing was returned.
1652    
1653    26. When studying a pattern that contained subroutine calls or assertions,
1654        the code for finding the minimum length of a possible match was handling
1655        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1656        group 1 called group 2 while simultaneously a separate group 2 called group
1657        1). A stack overflow occurred in this case. I have fixed this by limiting
1658        the recursion depth to 10.
1659    
1660    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1661        Fortmann. This supports explicit test numbers on the command line, and has
1662        argument validation and error reporting.
1663    
1664    28. An instance of \X with an unlimited repeat could fail if at any point the
1665        first character it looked at was a mark character.
1666    
1667    29. Some minor code refactoring concerning Unicode properties and scripts
1668        should reduce the stack requirement of match() slightly.
1669    
1670    30. Added the '=' option to pcretest to check the setting of unused capturing
1671        slots at the end of the pattern, which are documented as being -1, but are
1672        not included in the return count.
1673    
1674    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1675        compiled something random. Now it gives a compile-time error (as does
1676        Perl).
1677    
1678    32. A *MARK encountered during the processing of a positive assertion is now
1679        recorded and passed back (compatible with Perl).
1680    
1681    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1682        had alternative anchored branches, the search for a second match in a line
1683        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1684        matched the line "0102" twice. The same bug affected patterns that started
1685        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1686        twice.
1687    
1688    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1689        does, and because of capturing effects, quantifying parenthesized
1690        assertions may at times be useful. Quantifiers are now allowed for
1691        parenthesized assertions.
1692    
1693    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1694    
1695    36. \g was being checked for fancy things in a character class, when it should
1696        just be a literal "g".
1697    
1698    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1699        appearance of a nested POSIX class supersedes an apparent external class.
1700        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1701        unescaped square brackets may also appear as part of class names. For
1702        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1703        more like Perl. (But see 8.20/1 above.)
1704    
1705    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1706        was because it thought it was \N{name}, which is not supported).
1707    
1708    39. Add minix to OS list not supporting the -S option in pcretest.
1709    
1710    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1711        cannot analyze patterns in sufficient detail to catch mutual recursions
1712        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1713        subgroup is called recursively as a subpattern for a second time at the
1714        same position in the subject string. In previous releases this might have
1715        been caught by the recursion limit, or it might have run out of stack.
1716    
1717    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1718        happen only once. PCRE was, however incorrectly giving a compile time error
1719        "recursive call could loop indefinitely" because it cannot analyze the
1720        pattern in sufficient detail. The compile time test no longer happens when
1721        PCRE is compiling a conditional subpattern, but actual runaway loops are
1722        now caught at runtime (see 40 above).
1723    
1724    42. It seems that Perl allows any characters other than a closing parenthesis
1725        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1726        has been changed to be the same.
1727    
1728    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1729        as not to get warnings when autogen.sh is called. Also changed
1730        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1731    
1732    44. To help people who use pcregrep to scan files containing exceedingly long
1733        lines, the following changes have been made:
1734    
1735        (a) The default value of the buffer size parameter has been increased from
1736            8K to 20K. (The actual buffer used is three times this size.)
1737    
1738        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1739            PCRE is built.
1740    
1741        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1742            to be set at run time.
1743    
1744        (d) Numerical values in pcregrep options can be followed by K or M, for
1745            example --buffer-size=50K.
1746    
1747        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1748            given and the return code is set to 2.
1749    
1750    45. Add a pointer to the latest mark to the callout data block.
1751    
1752    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1753        partial match of an empty string instead of no match. This was specific to
1754        the use of ".".
1755    
1756    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1757        complete match instead of a partial match. This bug was dependent on both
1758        the PCRE_UTF8 and PCRE_DOTALL options being set.
1759    
1760    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1761        starting byte set, because \b was not being ignored.
1762    
1763    
1764    Version 8.12 15-Jan-2011
1765    ------------------------
1766    
1767    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1768        checks for such things as part of the documentation building process.
1769    
1770    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1771        --match-limit and --recursion-limit options (added for 8.11). In
1772        particular, this made one of the standard tests fail. (The integer value
1773        went into the wrong half of a long int.)
1774    
1775    3.  If the --colour option was given to pcregrep with -v (invert match), it
1776        did strange things, either producing crazy output, or crashing. It should,
1777        of course, ignore a request for colour when reporting lines that do not
1778        match.
1779    
1780    4.  Another pcregrep bug caused similar problems if --colour was specified with
1781        -M (multiline) and the pattern match finished with a line ending.
1782    
1783    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1784        matched in multiline mode, the following line was shown as part of the
1785        match. This seems wrong, so I have changed it.
1786    
1787    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1788        the check for further matches in the same line (so they could be coloured)
1789        to overrun the end of the current line. If another match was found, it was
1790        incorrectly shown (and then shown again when found in the next line).
1791    
1792    7.  If pcregrep was compiled under Windows, there was a reference to the
1793        function pcregrep_exit() before it was defined. I am assuming this was
1794        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1795        reported by a user. I've moved the definition above the reference.
1796    
1797    
1798    Version 8.11 10-Dec-2010
1799    ------------------------
1800    
1801    1.  (*THEN) was not working properly if there were untried alternatives prior
1802        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1803        backtracked to try for "b" instead of moving to the next alternative branch
1804        at the same level (in this case, to look for "c"). The Perl documentation
1805        is clear that when (*THEN) is backtracked onto, it goes to the "next
1806        alternative in the innermost enclosing group".
1807    
1808    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1809        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1810        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1811        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1812        (*THEN).
1813    
1814    3.  If \s appeared in a character class, it removed the VT character from
1815        the class, even if it had been included by some previous item, for example
1816        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1817        of \s, but is part of the POSIX "space" class.)
1818    
1819    4.  A partial match never returns an empty string (because you can always
1820        match an empty string at the end of the subject); however the checking for
1821        an empty string was starting at the "start of match" point. This has been
1822        changed to the "earliest inspected character" point, because the returned
1823        data for a partial match starts at this character. This means that, for
1824        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1825        (previously it gave "no match").
1826    
1827    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1828        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1829        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1830        has an implication that the given string is incomplete (because a partial
1831        match is preferred over a full match). For this reason, these items now
1832        give a partial match in this situation. [Aside: previously, the one case
1833        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1834        match rather than a full match, which was wrong by the old rules, but is
1835        now correct.]
1836    
1837    6.  There was a bug in the handling of #-introduced comments, recognized when
1838        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1839        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1840        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1841        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1842        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1843        places in pcre_compile().
1844    
1845    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1846        comments when looking ahead for named forward references to subpatterns,
1847        the only newline sequence it recognized was NL. It now handles newlines
1848        according to the set newline convention.
1849    
1850    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1851        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1852        cater for a lack of strerror(). These oversights have been fixed.
1853    
1854    9.  Added --match-limit and --recursion-limit to pcregrep.
1855    
1856    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1857    
1858    11. When the -o option was used, pcregrep was setting a return code of 1, even
1859        when matches were found, and --line-buffered was not being honoured.
1860    
1861    12. Added an optional parentheses number to the -o and --only-matching options
1862        of pcregrep.
1863    
1864    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1865        can match an empty string. The code to do it in pcretest and pcredemo
1866        needed fixing:
1867    
1868        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1869            only one byte after an empty string match just before CRLF (this case
1870            just got forgotten; "any" and "anycrlf" were OK).
1871    
1872        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1873            mode when an empty string match preceded an ASCII character followed by
1874            a non-ASCII character. (The code for advancing by one character rather
1875            than one byte was nonsense.)
1876    
1877        (c) The pcredemo.c sample program did not have any code at all to handle
1878            the cases when CRLF is a valid newline sequence.
1879    
1880    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1881        as a starting offset was within the subject string. There is now a new
1882        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1883        negative or greater than the length of the string. In order to test this,
1884        pcretest is extended to allow the setting of negative starting offsets.
1885    
1886    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1887        starting offset points to the beginning of a UTF-8 character was
1888        unnecessarily clumsy. I tidied it up.
1889    
1890    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1891        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1892    
1893    17. Nobody had reported that the --include_dir option, which was added in
1894        release 7.7 should have been called --include-dir (hyphen, not underscore)
1895        for compatibility with GNU grep. I have changed it to --include-dir, but
1896        left --include_dir as an undocumented synonym, and the same for
1897        --exclude-dir, though that is not available in GNU grep, at least as of
1898        release 2.5.4.
1899    
1900    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1901        characters from a string of bytes) have been redefined so as not to use
1902        loops, in order to improve performance in some environments. At the same
1903        time, I abstracted some of the common code into auxiliary macros to save
1904        repetition (this should not affect the compiled code).
1905    
1906    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1907        compile-time error is now given if \c is not followed by an ASCII
1908        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1909        different, and any byte value is allowed.)
1910    
1911    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1912        START_OPTIMIZE option, which is now allowed at compile time - but just
1913        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1914        to pcregrep and other applications that have no direct access to PCRE
1915        options. The new /Y option in pcretest sets this option when calling
1916        pcre_compile().
1917    
1918    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1919        back references. Groups containing recursive back references were forced to
1920        be atomic by that change, but in the case of named groups, the amount of
1921        memory required was incorrectly computed, leading to "Failed: internal
1922        error: code overflow". This has been fixed.
1923    
1924    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1925        pcretest.c, to avoid build problems in some Borland environments.
1926    
1927    
1928    Version 8.10 25-Jun-2010
1929    ------------------------
1930    
1931    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1932        THEN.
1933    
1934    2.  (*ACCEPT) was not working when inside an atomic group.
1935    
1936    3.  Inside a character class, \B is treated as a literal by default, but
1937        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1938        causes the error). The code is unchanged, but I tidied the documentation.
1939    
1940    4.  Inside a character class, PCRE always treated \R and \X as literals,
1941        whereas Perl faults them if its -w option is set. I have changed PCRE so
1942        that it faults them when PCRE_EXTRA is set.
1943    
1944    5.  Added support for \N, which always matches any character other than
1945        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1946    
1947    6.  When compiling pcregrep with newer versions of gcc which may have
1948        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1949        declared with attribute warn_unused_result" were given. Just casting the
1950        result to (void) does not stop the warnings; a more elaborate fudge is
1951        needed. I've used a macro to implement this.
1952    
1953    7.  Minor change to pcretest.c to avoid a compiler warning.
1954    
1955    8.  Added four artifical Unicode properties to help with an option to make
1956        \s etc use properties (see next item). The new properties are: Xan
1957        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1958    
1959    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1960        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1961        this option. Modified pcretest to add /W to test this facility. Added
1962        REG_UCP to make it available via the POSIX interface.
1963    
1964    10. Added --line-buffered to pcregrep.
1965    
1966    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1967        studied, and the match started with a letter with a code point greater than
1968        127 whose first byte was different to the first byte of the other case of
1969        the letter, the other case of this starting letter was not recognized
1970        (#976).
1971    
1972    12. If a pattern that was studied started with a repeated Unicode property
1973        test, for example, \p{Nd}+, there was the theoretical possibility of
1974        setting up an incorrect bitmap of starting bytes, but fortunately it could
1975        not have actually happened in practice until change 8 above was made (it
1976        added property types that matched character-matching opcodes).
1977    
1978    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1979        possible starting bytes for non-anchored patterns.
1980    
1981    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1982        \R, and also a number of cases that involve Unicode properties, both
1983        explicit and implicit when PCRE_UCP is set.
1984    
1985    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1986        input, it could crash or give wrong results if characters with values
1987        greater than 0xc0 were present in the subject string. (Detail: it assumed
1988        UTF-8 input when processing these items.)
1989    
1990    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1991        size_t is 64-bit (#991).
1992    
1993    17. Added a check for running out of memory when PCRE is compiled with
1994        --disable-stack-for-recursion (#990).
1995    
1996    18. If the last data line in a file for pcretest does not have a newline on
1997        the end, a newline was missing in the output.
1998    
1999    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2000        less than 128) in its various bitmaps. However, there is a facility for
2001        generating tables according to the current locale when PCRE is compiled. It
2002        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2003        characters, are recognized by isspace() and therefore were getting set in
2004        these tables, and indeed these tables seem to approximate to ISO 8859. This
2005        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2006        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2007        which of course cannot start UTF-8 characters. I have changed the code so
2008        that only real ASCII characters (less than 128) and the correct starting
2009        bytes for UTF-8 encodings are set for characters greater than 127 when in
2010        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2011        altogether.)
2012    
2013    20. Added the /T option to pcretest so as to be able to run tests with non-
2014        standard character tables, thus making it possible to include the tests
2015        used for 19 above in the standard set of tests.
2016    
2017    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2018        reference to a subpattern the other side of a comment that contains an
2019        opening parenthesis caused either an internal compiling error, or a
2020        reference to the wrong subpattern.
2021    
2022    
2023    Version 8.02 19-Mar-2010
2024    ------------------------
2025    
2026    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2027    
2028    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2029        configured.
2030    
2031    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2032        original author of that file, following a query about its status.
2033    
2034    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2035        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2036    
2037    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2038        quantifier applied to a forward-referencing subroutine call, could compile
2039        incorrect code or give the error "internal error: previously-checked
2040        referenced subpattern not found".
2041    
2042    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2043        variables to point to external functions. For these systems, therefore,
2044        pcre_malloc etc. are now initialized to local functions that call the
2045        relevant global functions.
2046    
2047    7.  There were two entries missing in the vectors called coptable and poptable
2048        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2049        I've fixed the data, and added a kludgy way of testing at compile time that
2050        the lengths are correct (equal to the number of opcodes).
2051    
2052    8.  Following on from 7, I added a similar kludge to check the length of the
2053        eint vector in pcreposix.c.
2054    
2055    9.  Error texts for pcre_compile() are held as one long string to avoid too
2056        much relocation at load time. To find a text, the string is searched,
2057        counting zeros. There was no check for running off the end of the string,
2058        which could happen if a new error number was added without updating the
2059        string.
2060    
2061    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2062    
2063    11. \K was not working if it appeared in an atomic group or in a group that
2064        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2065        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2066        the assertion is positive, but not if it is negative.
2067    
2068    12. Change 11 fortuitously reduced the size of the stack frame used in the
2069        "match()" function of pcre_exec.c by one pointer. Forthcoming
2070        implementation of support for (*MARK) will need an extra pointer on the
2071        stack; I have reserved it now, so that the stack frame size does not
2072        decrease.
2073    
2074    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2075        item in branch that calls a recursion is a subroutine call - as in the
2076        second branch in the above example - was incorrectly given the compile-
2077        time error "recursive call could loop indefinitely" because pcre_compile()
2078        was not correctly checking the subroutine for matching a non-empty string.
2079    
2080    14. The checks for overrunning compiling workspace could trigger after an
2081        overrun had occurred. This is a "should never occur" error, but it can be
2082        triggered by pathological patterns such as hundreds of nested parentheses.
2083        The checks now trigger 100 bytes before the end of the workspace.
2084    
2085    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2086    
2087    
2088    Version 8.01 19-Jan-2010
2089    ------------------------
2090    
2091    1.  If a pattern contained a conditional subpattern with only one branch (in
2092        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2093        computed the wrong minimum data length (which is of course zero for such
2094        subpatterns). This could cause incorrect "no match" results.
2095    
2096    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2097        the pattern is reset in the first branch, pcre_compile() failed with
2098        "internal error: code overflow at offset...". This happened only when
2099        the reset was to the original external option setting. (An optimization
2100        abstracts leading options settings into an external setting, which was the
2101        cause of this.)
2102    
2103    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2104        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2105        assertion pattern did not match (meaning that the assertion was true), it
2106        was incorrectly treated as false if the SKIP had been reached during the
2107        matching. This also applied to assertions used as conditions.
2108    
2109    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2110        assertion subpattern, including such a pattern used as a condition,
2111        unpredictable results occurred, instead of the error return
2112        PCRE_ERROR_DFA_UITEM.
2113    
2114    5.  The C++ GlobalReplace function was not working like Perl for the special
2115        situation when an empty string is matched. It now does the fancy magic
2116        stuff that is necessary.
2117    
2118    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2119        removed. (These were left over from very, very early versions of PCRE.)
2120    
2121    7.  Some cosmetic changes to the code to make life easier when compiling it
2122        as part of something else:
2123    
2124        (a) Change DEBUG to PCRE_DEBUG.
2125    
2126        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2127            called "current" as "current_branch", to prevent a collision with the
2128            Linux macro when compiled as a kernel module.
2129    
2130        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2131            prevent a collision with the Linux macro when compiled as a kernel
2132            module.
2133    
2134    8.  In pcre_compile() there are some checks for integer overflows that used to
2135        cast potentially large values to (double). This has been changed to that
2136        when building, a check for int64_t is made, and if it is found, it is used
2137        instead, thus avoiding the use of floating point arithmetic. (There is no
2138        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2139        double.
2140    
2141    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2142        2005 (difference between two addresses compared to an unsigned value).
2143    
2144    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2145        custom one, because of the following reported problem in Windows:
2146    
2147          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2148              under Win32.
2149          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2150              therefore missing the function definition.
2151          - The compiler thus generates a "C" signature for the test function.
2152          - The linker fails to find the "C" function.
2153          - PCRE fails to configure if asked to do so against libbz2.
2154    
2155    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2156        messages were output:
2157    
2158          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2159          rerunning libtoolize, to keep the correct libtool macros in-tree.
2160          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2161    
2162        I have done both of these things.
2163    
2164    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2165        most of the time, it *can* run out if it is given a pattern that contains a
2166        runaway infinite recursion. I updated the discussion in the pcrestack man
2167        page.
2168    
2169    13. Now that we have gone to the x.xx style of version numbers, the minor
2170        version may start with zero. Using 08 or 09 is a bad idea because users
2171        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2172        interpreted as invalid octal numbers. I've updated the previous comment in
2173        configure.ac, and also added a check that gives an error if 08 or 09 are
2174        used.
2175    
2176    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2177        causing partial matching to fail when the end of the subject matched \W
2178        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2179    
2180    15. There were some discrepancies between the declarations in pcre_internal.h
2181        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2182        their definitions. The declarations used "const uschar *" and the
2183        definitions used USPTR. Even though USPTR is normally defined as "const
2184        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2185        reported that: "This difference in casting confuses some C++ compilers, for
2186        example, SunCC recognizes above declarations as different functions and
2187        generates broken code for hbpcre." I have changed the declarations to use
2188        USPTR.
2189    
2190    16. GNU libtool is named differently on some systems. The autogen.sh script now
2191        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2192        (FreeBSD).
2193    
2194    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2195        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2196        comment: "Figure out how to create a longlong from a string: strtoll and
2197        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2198        instance, but it only takes 2 args instead of 3!"
2199    
2200    18. A subtle bug concerned with back references has been fixed by a change of
2201        specification, with a corresponding code fix. A pattern such as
2202        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2203        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2204        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2205        same bug. Such groups have to be quantified to be useful, or contained
2206        inside another quantified group. (If there's no repetition, the reference
2207        can never match.) The problem arises because, having left the group and
2208        moved on to the rest of the pattern, a later failure that backtracks into
2209        the group uses the captured value from the final iteration of the group
2210        rather than the correct earlier one. I have fixed this in PCRE by forcing
2211        any group that contains a reference to itself to be an atomic group; that
2212        is, there cannot be any backtracking into it once it has completed. This is
2213        similar to recursive and subroutine calls.
2214    
2215    
2216    Version 8.00 19-Oct-09
2217    ----------------------
2218    
2219    1.  The table for translating pcre_compile() error codes into POSIX error codes
2220        was out-of-date, and there was no check on the pcre_compile() error code
2221        being within the table. This could lead to an OK return being given in
2222        error.
2223    
2224    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2225        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2226        in a Windows environment.
2227    
2228    3.  The pcregrep --count option prints the count for each file even when it is
2229        zero, as does GNU grep. However, pcregrep was also printing all files when
2230        --files-with-matches was added. Now, when both options are given, it prints
2231        counts only for those files that have at least one match. (GNU grep just
2232        prints the file name in this circumstance, but including the count seems
2233        more useful - otherwise, why use --count?) Also ensured that the
2234        combination -clh just lists non-zero counts, with no names.
2235    
2236    4.  The long form of the pcregrep -F option was incorrectly implemented as
2237        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2238        but it seems right to fix it, and I didn't think it was worth preserving
2239        the old behaviour.
2240    
2241    5.  The command line items --regex=pattern and --regexp=pattern were not
2242        recognized by pcregrep, which required --regex pattern or --regexp pattern
2243        (with a space rather than an '='). The man page documented the '=' forms,
2244        which are compatible with GNU grep; these now work.
2245    
2246    6.  No libpcreposix.pc file was created for pkg-config; there was just
2247        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2248    
2249    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2250        when UCP support is not needed, by modifying the Python script that
2251        generates it from Unicode data files. This should not matter if the module
2252        is correctly used as a library, but I received one complaint about 50K of
2253        unwanted data. My guess is that the person linked everything into his
2254        program rather than using a library. Anyway, it does no harm.
2255    
2256    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2257        was a minimum greater than 1 for a wide character in a possessive
2258        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2259        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2260        character. Chaos in the form of incorrect output or a compiling loop could
2261        result.
2262    
2263    9.  The restrictions on what a pattern can contain when partial matching is
2264        requested for pcre_exec() have been removed. All patterns can now be
2265        partially matched by this function. In addition, if there are at least two
2266        slots in the offset vector, the offset of the earliest inspected character
2267        for the match and the offset of the end of the subject are set in them when
2268        PCRE_ERROR_PARTIAL is returned.
2269    
2270    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2271        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2272        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2273        and may be more useful for multi-segment matching.
2274    
2275    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2276        used to be given if ever the end of the subject was reached; now it is
2277        given only if matching could not proceed because another character was
2278        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2279        string "Z", which now yields "no match" instead of "partial match". In the
2280        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2281        final character ended with (*FAIL).
2282    
2283    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2284        if the pattern had a "must contain" character that was already found in the
2285        earlier partial match, unless partial matching was again requested. For
2286        example, with the pattern /dog.(body)?/, the "must contain" character is
2287        "g". If the first part-match was for the string "dog", restarting with
2288        "sbody" failed. This bug has been fixed.
2289    
2290    13. The string returned by pcre_dfa_exec() after a partial match has been
2291        changed so that it starts at the first inspected character rather than the
2292        first character of the match. This makes a difference only if the pattern
2293        starts with a lookbehind assertion or \b or \B (\K is not supported by
2294        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2295        matching functions compatible, and I think it's the right thing to do.
2296    
2297    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2298        so that the demonstration program is easily available in environments where
2299        PCRE has not been installed from source.
2300    
2301    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2302        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2303        library.
2304    
2305    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2306        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2307        is not the first non-POSIX option to be added. Clearly some people find
2308        these options useful.
2309    
2310    17. If a caller to the POSIX matching function regexec() passes a non-zero
2311        value for nmatch with a NULL value for pmatch, the value of
2312        nmatch is forced to zero.
2313    
2314    18. RunGrepTest did not have a test for the availability of the -u option of
2315        the diff command, as RunTest does. It now checks in the same way as
2316        RunTest, and also checks for the -b option.
2317    
2318    19. If an odd number of negated classes containing just a single character
2319        interposed, within parentheses, between a forward reference to a named
2320        subpattern and the definition of the subpattern, compilation crashed with
2321        an internal error, complaining that it could not find the referenced
2322        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2323        [The bug was that it was starting one character too far in when skipping
2324        over the character class, thus treating the ] as data rather than
2325        terminating the class. This meant it could skip too much.]
2326    
2327    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2328        /g option in pcretest when the pattern contains \K, which makes it possible
2329        to have an empty string match not at the start, even when the pattern is
2330        anchored. Updated pcretest and pcredemo to use this option.
2331    
2332    21. If the maximum number of capturing subpatterns in a recursion was greater
2333        than the maximum at the outer level, the higher number was returned, but
2334        with unset values at the outer level. The correct (outer level) value is
2335        now given.
2336    
2337    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2338        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2339        make it do so. The string so far is captured, making this feature
2340        compatible with Perl.
2341    
2342    23. The tests have been re-organized, adding tests 11 and 12, to make it
2343        possible to check the Perl 5.10 features against Perl 5.10.
2344    
2345    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2346        pattern matches a fixed length string. PCRE did not allow this; now it
2347        does. Neither allows recursion.
2348    
2349    25. I finally figured out how to implement a request to provide the minimum
2350        length of subject string that was needed in order to match a given pattern.
2351        (It was back references and recursion that I had previously got hung up
2352        on.) This code has now been added to pcre_study(); it finds a lower bound
2353        to the length of subject needed. It is not necessarily the greatest lower
2354        bound, but using it to avoid searching strings that are too short does give
2355        some useful speed-ups. The value is available to calling programs via
2356        pcre_fullinfo().
2357    
2358    26. While implementing 25, I discovered to my embarrassment that pcretest had
2359        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2360        study optimizations had never been tested with that matching function.
2361        Oops. What is worse, even when it was passed study data, there was a bug in
2362        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2363        were also very few tests of studied patterns with pcre_dfa_exec().
2364    
2365    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2366        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2367        on the other side of the coin, they are no longer allowed to have different
2368        names, because these cannot be distinguished in PCRE, and this has caused
2369        confusion. (This is a difference from Perl.)
2370    
2371    28. When duplicate subpattern names are present (necessarily with different
2372        numbers, as required by 27 above), and a test is made by name in a
2373        conditional pattern, either for a subpattern having been matched, or for
2374        recursion in such a pattern, all the associated numbered subpatterns are
2375        tested, and the overall condition is true if the condition is true for any
2376        one of them. This is the way Perl works, and is also more like the way
2377        testing by number works.
2378    
2379    
2380    Version 7.9 11-Apr-09
2381    ---------------------
2382    
2383    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2384        (pcretest), all targets were linked against these libraries. This included
2385        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2386        libraries. This caused unwanted dependencies to be created. This problem
2387        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2388        pcretest is linked with readline.
2389    
2390    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2391        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2392        moved outside it again, because FALSE and TRUE are already defined in AIX,
2393        but BOOL is not.
2394    
2395    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2396        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2397    
2398    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2399        hyphens) following file names and line numbers when outputting matching
2400        lines. This is not true; no spaces are inserted. I have also clarified the
2401        wording for the --colour (or --color) option.
2402    
2403    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2404        was not coloured; this is different to GNU grep, so I have changed it to be
2405        the same.
2406    
2407    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2408        each matching line was coloured. Now it goes on to look for further matches
2409        of any of the test patterns, which is the same behaviour as GNU grep.
2410    
2411    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2412        doesn't make sense to accept an empty string match in pcregrep, so I have
2413        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2414        seems to be how GNU grep behaves. [But see later change 40 for release
2415        8.33.]
2416    
2417    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2418        start or after a newline", because the conditional assertion was not being
2419        correctly handled. The rule now is that both the assertion and what follows
2420        in the first alternative must satisfy the test.
2421    
2422    9.  If auto-callout was enabled in a pattern with a conditional group whose
2423        condition was an assertion, PCRE could crash during matching, both with
2424        pcre_exec() and pcre_dfa_exec().
2425    
2426    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2427        used for matching.
2428    
2429    11. Unicode property support in character classes was not working for
2430        characters (bytes) greater than 127 when not in UTF-8 mode.
2431    
2432    12. Added the -M command line option to pcretest.
2433    
2434    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2435    
2436    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2437    
2438    16. Added comments and documentation about mis-use of no_arg in the C++
2439        wrapper.
2440    
2441    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2442        from Martin Jerabek that uses macro names for all relevant character and
2443        string constants.
2444    
2445    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2446        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2447        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2448        these, but not everybody uses configure.
2449    
2450    19. A conditional group that had only one branch was not being correctly
2451        recognized as an item that could match an empty string. This meant that an
2452        enclosing group might also not be so recognized, causing infinite looping
2453        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2454        with the subject "ab", where knowledge that the repeated group can match
2455        nothing is needed in order to break the loop.
2456    
2457    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2458        exec(), but without supplying a callout function, matching went wrong.
2459    
2460    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2461        leak if the size of the offset vector was greater than 30. When the vector
2462        is smaller, the saved offsets during recursion go onto a local stack
2463        vector, but for larger vectors malloc() is used. It was failing to free
2464        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2465        error, in fact).
2466    
2467    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2468        heapframe that is used only when UTF-8 support is enabled. This caused no
2469        problem, but was untidy.
2470    
2471    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2472        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2473        included within another project.
2474    
2475    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2476        slightly modified by me:
2477    
2478          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2479              not building pcregrep.
2480    
2481          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2482              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2483    
2484    25. Forward references, both numeric and by name, in patterns that made use of
2485        duplicate group numbers, could behave incorrectly or give incorrect errors,
2486        because when scanning forward to find the reference group, PCRE was not
2487        taking into account the duplicate group numbers. A pattern such as
2488        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2489    
2490    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2491        the feature of a custom pointer more persuasive (as requested by a user).
2492    
2493    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2494        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2495        pre-defined.
2496    
2497    28. Added support for (*UTF8) at the start of a pattern.
2498    
2499    29. Arrange for flags added by the "release type" setting in CMake to be shown
2500        in the configuration summary.
2501    
2502    
2503    Version 7.8 05-Sep-08
2504    ---------------------
2505    
2506    1.  Replaced UCP searching code with optimized version as implemented for Ad
2507        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2508        stage table and inline lookup instead of a function, giving speed ups of 2
2509        to 5 times on some simple patterns that I tested. Permission was given to
2510        distribute the MultiStage2.py script that generates the tables (it's not in
2511        the tarball, but is in the Subversion repository).
2512    
2513    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2514        scripts.
2515    
2516    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2517        a group with a zero qualifier. The result of the study could be incorrect,
2518        or the function might crash, depending on the pattern.
2519    
2520    4.  Caseless matching was not working for non-ASCII characters in back
2521        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2522        It now works when Unicode Property Support is available.
2523    
2524    5.  In pcretest, an escape such as \x{de} in the data was always generating
2525        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2526        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2527        truncation.
2528    
2529    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2530    
2531    7.  Added two (int) casts to pcregrep when printing the difference of two
2532        pointers, in case they are 64-bit values.
2533    
2534    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2535        test 2 if it fails.
2536    
2537    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2538        and a #define of that name to empty if it is not externally set. This is to
2539        allow users of MSVC to set it if necessary.
2540    
2541    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2542        the convenience functions in the pcre_get.c source file.
2543    
2544    11. An option change at the start of a pattern that had top-level alternatives
2545        could cause overwriting and/or a crash. This command provoked a crash in
2546        some environments:
2547    
2548          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2549    
2550        This potential security problem was recorded as CVE-2008-2371.
2551    
2552    12. For a pattern where the match had to start at the beginning or immediately
2553        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2554        pcre_dfa_exec() could read past the end of the passed subject if there was
2555        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2556        pcretest so that it places the subject at the end of its malloc-ed buffer.
2557    
2558    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2559        exec() might read past the end of the data buffer in UTF-8 mode.
2560    
2561    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2562        the data contained the byte 0x85 as part of a UTF-8 character within its
2563        first line. This applied both to normal and DFA matching.
2564    
2565    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2566        /^[^d]*?$/8 failed to match "abc".
2567    
2568    16. Added a missing copyright notice to pcrecpp_internal.h.
2569    
2570    17. Make it more clear in the documentation that values returned from
2571        pcre_exec() in ovector are byte offsets, not character counts.
2572    
2573    18. Tidied a few places to stop certain compilers from issuing warnings.
2574    
2575    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2576        supplied by Stefan Weber. I made a further small update for 7.8 because
2577        there is a change of source arrangements: the pcre_searchfuncs.c module is
2578        replaced by pcre_ucd.c.
2579    
2580    
2581    Version 7.7 07-May-08
2582    ---------------------
2583    
2584    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2585        a string to a long long, pretend we don't even have a long long." This is
2586        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2587    
2588    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2589        pre-7.6 versions, which defined a global no_arg variable instead of putting
2590        it in the RE class. (See also #8 below.)
2591    
2592    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2593        Lopes.
2594    
2595    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2596    
2597        (1) The include/exclude patterns were being applied to the whole pathnames
2598            of files, instead of just to the final components.
2599    
2600        (2) If there was more than one level of directory, the subdirectories were
2601            skipped unless they satisfied the include/exclude conditions. This is
2602            inconsistent with GNU grep (and could even be seen as contrary to the
2603            pcregrep specification - which I improved to make it absolutely clear).
2604            The action now is always to scan all levels of directory, and just
2605            apply the include/exclude patterns to regular files.
2606    
2607    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2608        --exclude_dir in the tests to avoid scanning .svn directories.
2609    
2610    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2611        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2612        doesn't support NULs in patterns.
2613    
2614    7.  Added some missing "const"s to declarations of static tables in
2615        pcre_compile.c and pcre_dfa_exec.c.
2616    
2617    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2618        caused by fix #2  above. (Subsequently also a second patch to fix the
2619        first patch. And a third patch - this was a messy problem.)
2620    
2621    9.  Applied Craig's patch to remove the use of push_back().
2622    
2623    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2624        matching function regexec().
2625    
2626    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2627        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2628        references. PCRE supports relative numbers with this syntax (I don't think
2629        Oniguruma does).
2630    
2631    12. Previously, a group with a zero repeat such as (...){0} was completely
2632        omitted from the compiled regex. However, this means that if the group
2633        was called as a subroutine from elsewhere in the pattern, things went wrong
2634        (an internal error was given). Such groups are now left in the compiled
2635        pattern, with a new opcode that causes them to be skipped at execution
2636        time.
2637    
2638    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2639        to the way PCRE behaves:
2640    
2641        (a) A lone ] character is dis-allowed (Perl treats it as data).
2642    
2643        (b) A back reference to an unmatched subpattern matches an empty string
2644            (Perl fails the current match path).
2645    
2646        (c) A data ] in a character class must be notated as \] because if the
2647            first data character in a class is ], it defines an empty class. (In
2648            Perl it is not possible to have an empty class.) The empty class []
2649            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2650            The negative empty class [^] matches any one character, independently
2651            of the DOTALL setting.
2652    
2653    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2654        non-existent subpattern following a character class starting with ']' and
2655        containing () gave an internal compiling error instead of "reference to
2656        non-existent subpattern". Fortunately, when the pattern did exist, the
2657        compiled code was correct. (When scanning forwards to check for the
2658        existence of the subpattern, it was treating the data ']' as terminating
2659        the class, so got the count wrong. When actually compiling, the reference
2660        was subsequently set up correctly.)
2661    
2662    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2663        it was being rejected as not supported by pcre_dfa_exec(), even though
2664        other assertions are supported. I have made pcre_dfa_exec() support
2665        (*FAIL).
2666    
2667    16. The implementation of 13c above involved the invention of a new opcode,
2668        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2669        cannot be changed at match time, I realized I could make a small
2670        improvement to matching performance by compiling OP_ALLANY instead of
2671        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2672        on the OP_ANY path.
2673    
2674    17. Compiling pcretest on Windows with readline support failed without the
2675        following two fixes: (1) Make the unistd.h include conditional on
2676        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2677    
2678    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2679        ncurses library to be included for pcretest when ReadLine support is
2680        requested, but also to allow for it to be overridden. This patch came from
2681        Daniel Bergström.
2682    
2683    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2684        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2685        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2686        spotting this.
2687    
2688    
2689    Version 7.6 28-Jan-08
2690    ---------------------
2691    
2692    1.  A character class containing a very large number of characters with
2693        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2694        overflow.
2695    
2696    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2697        HAVE_LONG_LONG is not defined.
2698    
2699    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2700        bring them up to date and include new features. This patch includes:
2701    
2702        - Fixed PH's badly added libz and libbz2 support.
2703        - Fixed a problem with static linking.
2704        - Added pcredemo. [But later removed - see 7 below.]
2705        - Fixed dftables problem and added an option.
2706        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2707            HAVE_LONG_LONG.
2708        - Added readline support for pcretest.
2709        - Added an listing of the option settings after cmake has run.
2710    
2711    4.  A user submitted a patch to Makefile that makes it easy to create
2712        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2713        Makefile.am that cause it to include this special target, without
2714        affecting anything else. Note that the same mingw target plus all
2715        the other distribution libraries and programs are now supported
2716        when configuring with CMake (see 6 below) instead of with
2717        Configure/Make.
2718    
2719    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2720        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2721        exported in the Windows port". It has not yet been confirmed that the patch
2722        solves the problem, but it does no harm.
2723    
2724    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2725        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2726        with CMake, and also correct the comment about stack recursion.
2727    
2728    7.  Remove the automatic building of pcredemo from the ./configure system and
2729        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2730        of a program that users should build themselves after PCRE is installed, so
2731        building it automatically is not really right. What is more, it gave
2732        trouble in some build environments.
2733    
2734    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2735    
2736    
2737    Version 7.5 10-Jan-08
2738    ---------------------
2739    
2740    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2741        values in parens when parsing an RE using the C++ wrapper."
2742    
2743    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2744        Characters greater than 255 were excluded from the class instead of being
2745        included.
2746    
2747    3.  The same bug as (2) above applied to negated POSIX classes such as
2748        [:^space:].
2749    
2750    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2751        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2752        I have changed it.
2753    
2754    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2755        first named subpattern) and a construct such as (?&a) would reference the
2756        first named subpattern whose name started with "a" (in other words, the
2757        length check was missing). Both these problems are fixed. "Subpattern name
2758        expected" is now given for (?&) (a zero-length name), and this patch also
2759        makes it give the same error for \k'' (previously it complained that that
2760        was a reference to a non-existent subpattern).
2761    
2762    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2763        this is right because (?- can be followed by option settings as well as by
2764        digits. I have, however, made the messages clearer.
2765    
2766    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2767        than the number used in the conditional) now cause a compile-time error.
2768        This is actually not compatible with Perl, which accepts such patterns, but
2769        treats the conditional as always being FALSE (as PCRE used to), but it
2770        seems to me that giving a diagnostic is better.
2771    
2772    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2773        and messages.
2774    
2775    9.  Fix two occurrences of "backslash" in comments that should have been
2776        "backspace".
2777    
2778    10. Remove two redundant lines of code that can never be obeyed (their function
2779        was moved elsewhere).
2780    
2781    11. The program that makes PCRE's Unicode character property table had a bug
2782        which caused it to generate incorrect table entries for sequences of
2783        characters that have the same character type, but are in different scripts.
2784        It amalgamated them into a single range, with the script of the first of
2785        them. In other words, some characters were in the wrong script. There were
2786        thirteen such cases, affecting characters in the following ranges:
2787    
2788          U+002b0 - U+002c1
2789          U+0060c - U+0060d
2790          U+0061e - U+00612
2791          U+0064b - U+0065e
2792          U+0074d - U+0076d
2793          U+01800 - U+01805
2794          U+01d00 - U+01d77
2795          U+01d9b - U+01dbf
2796          U+0200b - U+0200f
2797          U+030fc - U+030fe
2798          U+03260 - U+0327f
2799          U+0fb46 - U+0fbb1
2800          U+10450 - U+1049d
2801    
2802    12. The -o option (show only the matching part of a line) for pcregrep was not
2803        compatible with GNU grep in that, if there was more than one match in a
2804        line, it showed only the first of them. It now behaves in the same way as
2805        GNU grep.
2806    
2807    13. If the -o and -v options were combined for pcregrep, it printed a blank
2808        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2809        does the same. The return code can be used to tell if there were any
2810        non-matching lines.
2811    
2812    14. Added --file-offsets and --line-offsets to pcregrep.
2813    
2814    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2815        infinitely looping recursion. The bug was that positive lookaheads were not
2816        being skipped when checking for a possible empty match (negative lookaheads
2817        and both kinds of lookbehind were skipped).
2818    
2819    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2820        inclusion of <windows.h> to before rather than after the definition of
2821        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2822    
2823    17. Specifying a possessive quantifier with a specific limit for a Unicode
2824        character property caused pcre_compile() to compile bad code, which led at
2825        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2826        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2827        caused the error; without that there was no problem.
2828    
2829    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2830    
2831    19. Added --enable-pcretest-libreadline.
2832    
2833    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2834        RE::GlobalReplace(). As a result, the number of replacements returned was
2835        double what it should be. I removed one of the increments, but Craig sent a
2836        later patch that removed the other one (the right fix) and added unit tests
2837        that check the return values (which was not done before).
2838    
2839    21. Several CMake things:
2840    
2841        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2842            the names libpcre and libpcreposix, not just pcre and pcreposix.
2843    
2844        (2) The above change means that pcretest and pcregrep are now correctly
2845            linked with the newly-built libraries, not previously installed ones.
2846    
2847        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2848    
2849    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2850        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2851        UTF-8 newline character). The key issue is that the pattern starts .*;
2852        this means that the match must be either at the beginning, or after a
2853        newline. The bug was in the code for advancing after a failed match and
2854        checking that the new position followed a newline. It was not taking
2855        account of UTF-8 characters correctly.
2856    
2857    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2858        character classes. PCRE was not treating the sequence [:...:] as a
2859        character class unless the ... were all letters. Perl, however, seems to
2860        allow any characters between [: and :], though of course it rejects as
2861        unknown any "names" that contain non-letters, because all the known class
2862        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2863        for example, whereas PCRE did not - it did not recognize a POSIX character
2864        class. This seemed a bit dangerous, so the code has been changed to be
2865        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2866        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2867        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2868        Perl does, and where it didn't before.
2869    
2870    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2871        Windows environments %n is disabled by default.
2872    
2873    
2874    Version 7.4 21-Sep-07
2875    ---------------------
2876    
2877    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2878        means that a class such as [\s] counted as "explicit reference to CR or
2879        LF". That isn't really right - the whole point of the change was to try to
2880        help when there was an actual mention of one of the two characters. So now
2881        the change happens only if \r or \n (or a literal CR or LF) character is
2882        encountered.
2883    
2884    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2885        of both had grown to the point where there were only 3 bits left.
2886        Fortunately, there was spare space in the data structure, and so I have
2887        moved the internal flags into a new 16-bit field to free up more option
2888        bits.
2889    
2890    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2891        but did not set the internal JCHANGED flag - either of these is enough to
2892        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2893        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2894        start sets both bits.
2895    
2896    4.  Added options (at build time, compile time, exec time) to change \R from
2897        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2898    
2899    5.  doc/pcresyntax.html was missing from the distribution.
2900    
2901    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2902        compatibility, even though it is no longer used.
2903    
2904    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2905        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2906        windows.h file is present (where different names are used). [This was
2907        reversed later after testing - see 16 below.]
2908    
2909    8.  Changed all #include <config.h> to #include "config.h". There were also
2910        some further <pcre.h> cases that I changed to "pcre.h".
2911    
2912    9.  When pcregrep was used with the --colour option, it missed the line ending
2913        sequence off the lines that it output.
2914    
2915    10. It was pointed out to me that arrays of string pointers cause lots of
2916        relocations when a shared library is dynamically loaded. A technique of
2917        using a single long string with a table of offsets can drastically reduce
2918        these. I have refactored PCRE in four places to do this. The result is
2919        dramatic:
2920    
2921          Originally:                          290
2922          After changing UCP table:            187
2923          After changing error message table:   43
2924          After changing table of "verbs"       36
2925          After changing table of Posix names   22
2926    
2927        Thanks to the folks working on Gregex for glib for this insight.
2928    
2929    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2930        unicode-properties was also set.
2931    
2932    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2933    
2934    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2935        checked only for CRLF.
2936    
2937    14. Added casts to pcretest.c to avoid compiler warnings.
2938    
2939    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2940    
2941    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2942        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2943        entirely. This removes changes made in 7 above.
2944    
2945    17. The CMake files have been updated, and there is now more information about
2946        building with CMake in the NON-UNIX-USE document.
2947    
2948    
2949    Version 7.3 28-Aug-07
2950    ---------------------
2951    
2952     1. In the rejigging of the build system that eventually resulted in 7.1, the
2953        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2954        brackets there is not right, since it causes compilers to look for an
2955        installed pcre.h, not the version that is in the source that is being
2956        compiled (which of course may be different). I have changed it back to:
2957    
2958          #include "pcre.h"
2959    
2960        I have a vague recollection that the change was concerned with compiling in
2961        different directories, but in the new build system, that is taken care of
2962        by the VPATH setting the Makefile.
2963    
2964     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2965        when the subject happened to end in the byte 0x85 (e.g. if the last
2966        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2967        characters but of course it shouldn't be taken as a newline when it is part
2968        of another character. The bug was that, for an unlimited repeat of . in
2969        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2970        characters when looking for a newline.
2971    
2972     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2973    
2974     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2975        in debug output.
2976    
2977     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2978        long printing in the pcrecpp unittest when running under MinGW.
2979    
2980     6. ESC_K was left out of the EBCDIC table.
2981    
2982     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2983        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2984        limit also applies to "virtual nesting" when a pattern is recursive, and in
2985        this case 1000 isn't so big. I have been able to remove this limit at the
2986        expense of backing off one optimization in certain circumstances. Normally,
2987        when pcre_exec() would call its internal match() function recursively and
2988        immediately return the result unconditionally, it uses a "tail recursion"
2989        feature to save stack. However, when a subpattern that can match an empty
2990        string has an unlimited repetition quantifier, it no longer makes this
2991        optimization. That gives it a stack frame in which to save the data for
2992        checking that an empty string has been matched. Previously this was taken
2993        from the 1000-entry workspace that had been reserved. So now there is no
2994        explicit limit, but more stack is used.
2995    
2996     8. Applied Daniel's patches to solve problems with the import/export magic
2997        syntax that is required for Windows, and which was going wrong for the
2998        pcreposix and pcrecpp parts of the library. These were overlooked when this
2999        problem was solved for the main library.
3000    
3001     9. There were some crude static tests to avoid integer overflow when computing
3002        the size of patterns that contain repeated groups with explicit upper
3003        limits. As the maximum quantifier is 65535, the maximum group length was
3004        set at 30,000 so that the product of these two numbers did not overflow a
3005        32-bit integer. However, it turns out that people want to use groups that
3006        are longer than 30,000 bytes (though not repeat them that many times).
3007        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3008        made it possible to implement the integer overflow checks in a much more
3009        dynamic way, which I have now done. The artificial limitation on group
3010        length has been removed - we now have only the limit on the total length of
3011        the compiled pattern, which depends on the LINK_SIZE setting.
3012    
3013    10. Fixed a bug in the documentation for get/copy named substring when
3014        duplicate names are permitted. If none of the named substrings are set, the
3015        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3016        empty string.
3017    
3018    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3019        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3020        because the ] is interpreted as the first data character and the
3021        terminating ] is not found. PCRE has been made compatible with Perl in this
3022        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3023        cause memory overwriting.
3024    
3025    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3026        string has been matched (to stop an infinite loop). It was not recognizing
3027        a conditional subpattern that could match an empty string if that
3028        subpattern was within another subpattern. For example, it looped when
3029        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3030        condition was not nested. This bug has been fixed.
3031    
3032    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3033        past the start of the subject in the presence of bytes with the top bit
3034        set, for example "\x8aBCD".
3035    
3036    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3037        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3038    
3039    14. Optimized (?!) to (*FAIL).
3040    
3041    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3042        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3043        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3044        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3045        does: it's just the validity check that is more restrictive.
3046    
3047    16. Inserted checks for integer overflows during escape sequence (backslash)
3048        processing, and also fixed erroneous offset values for syntax errors during
3049        backslash processing.
3050    
3051    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3052        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3053    
3054    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3055        caused an overrun.
3056    
3057    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3058        something other than just ASCII characters) inside a group that had an
3059        unlimited repeat caused a loop at compile time (while checking to see
3060        whether the group could match an empty string).
3061    
3062    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3063        [\P{Any}] did so. (Error in the code for printing property names.)
3064    
3065    21. An orphan \E inside a character class could cause a crash.
3066    
3067    22. A repeated capturing bracket such as (A)? could cause a wild memory
3068        reference during compilation.
3069    
3070    23. There are several functions in pcre_compile() that scan along a compiled
3071        expression for various reasons (e.g. to see if it's fixed length for look
3072        behind). There were bugs in these functions when a repeated \p or \P was
3073        present in the pattern. These operators have additional parameters compared
3074        with \d, etc, and these were not being taken into account when moving along
3075        the compiled data. Specifically:
3076    
3077        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3078            length.
3079    
3080        (b) An item such as \pL+ within a repeated group could cause crashes or
3081            loops.
3082    
3083        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3084            "reference to non-existent subpattern" error.
3085    
3086        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3087    
3088    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3089        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3090    
3091    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3092    
3093    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3094        character were causing crashes (broken optimization).
3095    
3096    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3097        \p or \P) caused a compile-time loop.
3098    
3099    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3100        break. For example, the unstudied pattern [\r\n]A does not match the string
3101        "\r\nA" because change 7.0/46 below moves the current point on by two
3102        characters after failing to match at the start. However, the pattern \nA
3103        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3104        the same is true. There doesn't seem any very clean way out of this, but
3105        what I have chosen to do makes the common cases work: PCRE now takes note
3106        of whether there can be an explicit match for \r or \n anywhere in the
3107        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3108        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3109        pattern has explicit CR or LF references.
3110    
3111    29. Added (*CR) etc for changing newline setting at start of pattern.
3112    
3113    
3114    Version 7.2 19-Jun-07
3115  ---------------------  ---------------------
3116    
3117   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3131  Version 7.2 01-May-07
3131      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3132      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3133      in all cases.      in all cases.
3134    
3135   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3136    
3137      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3138          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3139    
3140      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3141          to be opened parentheses.          to be opened parentheses.
3142    
3143      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3144          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3145    
3146        (d) \K resets the start of the current match so that everything before
3147            is not part of it.
3148    
3149        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3150    
3151        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3152            reference syntax.
3153    
3154        (g) (?| introduces a group in which the numbering of parentheses in each
3155            alternative starts with the same number.
3156    
3157        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3158    
3159     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3160        PCRE_INFO_JCHANGED.
3161    
3162     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3163        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3164        for detecting groups that can match an empty string.
3165    
3166     9. A pattern with a very large number of alternatives (more than several
3167        hundred) was running out of internal workspace during the pre-compile
3168        phase, where pcre_compile() figures out how much memory will be needed. A
3169        bit of new cunning has reduced the workspace needed for groups with
3170        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3171        workspace instead of running out of the 4096 that are available.
3172    
3173    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3174    
3175    11. Applied patch from Google to remove an optimization that didn't quite work.
3176        The report of the bug said:
3177    
3178          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3179          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3180          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3181    
3182    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3183        it matched the wrong number of bytes.
3184    
3185    
3186  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3187  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1595

  ViewVC Help
Powered by ViewVC 1.1.5