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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1596 by zherczeg, Wed Aug 26 13:44:24 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  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    38. Fixed a corner case of range optimization in JIT.
147    
148    
149    Version 8.37 28-April-2015
150    --------------------------
151    
152    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
153        for those parentheses to be closed with whatever has been captured so far.
154        However, it was failing to mark any other groups between the hightest
155        capture so far and the currrent group as "unset". Thus, the ovector for
156        those groups contained whatever was previously there. An example is the
157        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
158    
159    2.  If an assertion condition was quantified with a minimum of zero (an odd
160        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
161    
162    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
163        unrecognized modifier, a crash could occur.
164    
165    4.  An attempt to do global matching in pcretest with a zero-length ovector
166        caused a crash.
167    
168    5.  Fixed a memory leak during matching that could occur for a subpattern
169        subroutine call (recursive or otherwise) if the number of captured groups
170        that had to be saved was greater than ten.
171    
172    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
173        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
174        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
175    
176    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
177        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
178    
179    8. When a pattern is compiled, it remembers the highest back reference so that
180       when matching, if the ovector is too small, extra memory can be obtained to
181       use instead. A conditional subpattern whose condition is a check on a
182       capture having happened, such as, for example in the pattern
183       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
184       setting the highest backreference number. This mattered only if pcre_exec()
185       was called with an ovector that was too small to hold the capture, and there
186       was no other kind of back reference (a situation which is probably quite
187       rare). The effect of the bug was that the condition was always treated as
188       FALSE when the capture could not be consulted, leading to a incorrect
189       behaviour by pcre_exec(). This bug has been fixed.
190    
191    9. A reference to a duplicated named group (either a back reference or a test
192       for being set in a conditional) that occurred in a part of the pattern where
193       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
194       to be incorrectly calculated, leading to overwriting.
195    
196    10. A mutually recursive set of back references such as (\2)(\1) caused a
197        segfault at study time (while trying to find the minimum matching length).
198        The infinite loop is now broken (with the minimum length unset, that is,
199        zero).
200    
201    11. If an assertion that was used as a condition was quantified with a minimum
202        of zero, matching went wrong. In particular, if the whole group had
203        unlimited repetition and could match an empty string, a segfault was
204        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
205        assertions to be quantified, but not if they are being used as conditions,
206        so the above pattern is faulted by Perl. PCRE has now been changed so that
207        it also rejects such patterns.
208    
209    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
210        failed to allow the zero-repeat case if pcre2_exec() was called with an
211        ovector too small to capture the group.
212    
213    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
214        Red Hat Product Security:
215    
216        (a) A crash if /K and /F were both set with the option to save the compiled
217        pattern.
218    
219        (b) Another crash if the option to print captured substrings in a callout
220        was combined with setting a null ovector, for example \O\C+ as a subject
221        string.
222    
223    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
224        forward reference repeated a large (but limited) number of times within a
225        repeated outer group that has a zero minimum quantifier, caused incorrect
226        code to be compiled, leading to the error "internal error:
227        previously-checked referenced subpattern not found" when an incorrect
228        memory address was read. This bug was reported as "heap overflow",
229        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
230        CVE-2015-2325.
231    
232    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
233        call within a group that also contained a recursive back reference caused
234        incorrect code to be compiled. This bug was reported as "heap overflow",
235        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
236        number CVE-2015-2326.
237    
238    24. Computing the size of the JIT read-only data in advance has been a source
239        of various issues, and new ones are still appear unfortunately. To fix
240        existing and future issues, size computation is eliminated from the code,
241        and replaced by on-demand memory allocation.
242    
243    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
244        adjacent to the end of the range, and the range contained characters with
245        more than one other case, caused incorrect behaviour when compiled in UTF
246        mode. In that example, the range a-j was left out of the class.
247    
248    26. Fix JIT compilation of conditional blocks, which assertion
249        is converted to (*FAIL). E.g: /(?(?!))/.
250    
251    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
252        discovered by the LLVM fuzzer.
253    
254    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
255        when this assertion was used as a condition, for example (?(?!)a|b). In
256        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
257        error about an unsupported item.
258    
259    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
260        possessification code could take exponential time to complete. A recursion
261        depth limit of 1000 has been imposed to limit the resources used by this
262        optimization.
263    
264    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
265        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
266        because \S ensures they are all in the class. The code for doing this was
267        interacting badly with the code for computing the amount of space needed to
268        compile the pattern, leading to a buffer overflow. This bug was discovered
269        by the LLVM fuzzer.
270    
271    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
272        other kinds of group caused stack overflow at compile time. This bug was
273        discovered by the LLVM fuzzer.
274    
275    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
276        between a subroutine call and its quantifier was incorrectly compiled,
277        leading to buffer overflow or other errors. This bug was discovered by the
278        LLVM fuzzer.
279    
280    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
281        assertion after (?(. The code was failing to check the character after
282        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
283        was discovered by the LLVM fuzzer.
284    
285    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
286        a fixed maximum following a group that contains a subroutine reference was
287        incorrectly compiled and could trigger buffer overflow. This bug was
288        discovered by the LLVM fuzzer.
289    
290    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
291        caused a stack overflow instead of the diagnosis of a non-fixed length
292        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
293    
294    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
295        (e.g. /(?<=\Ka)/) could make pcregrep loop.
296    
297    37. There was a similar problem to 36 in pcretest for global matches.
298    
299    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
300        and a subsequent item in the pattern caused a non-match, backtracking over
301        the repeated \X did not stop, but carried on past the start of the subject,
302        causing reference to random memory and/or a segfault. There were also some
303        other cases where backtracking after \C could crash. This set of bugs was
304        discovered by the LLVM fuzzer.
305    
306    39. The function for finding the minimum length of a matching string could take
307        a very long time if mutual recursion was present many times in a pattern,
308        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
309        method has been implemented. This infelicity was discovered by the LLVM
310        fuzzer.
311    
312    40. Static linking against the PCRE library using the pkg-config module was
313        failing on missing pthread symbols.
314    
315    
316    Version 8.36 26-September-2014
317    ------------------------------
318    
319    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
320        -Wmissing-field-initializers and -Wunused-parameter.
321    
322    2.  The tests for quantifiers being too big (greater than 65535) were being
323        applied after reading the number, and stupidly assuming that integer
324        overflow would give a negative number. The tests are now applied as the
325        numbers are read.
326    
327    3.  Tidy code in pcre_exec.c where two branches that used to be different are
328        now the same.
329    
330    4.  The JIT compiler did not generate match limit checks for certain
331        bracketed expressions with quantifiers. This may lead to exponential
332        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
333        issue should be resolved now.
334    
335    5.  Fixed an issue, which occures when nested alternatives are optimized
336        with table jumps.
337    
338    6.  Inserted two casts and changed some ints to size_t in the light of some
339        reported 64-bit compiler warnings (Bugzilla 1477).
340    
341    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
342        an empty string, which sometimes were behaving incorrectly in the
343        interpreter (though correctly in the JIT matcher). This pcretest input is
344        an example:
345    
346          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
347          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
348    
349        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
350        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
351        for an empty string was breaking the inner loop and carrying on at a lower
352        level, when possessive repeated groups should always return to a higher
353        level as they have no backtrack points in them. The empty string test now
354        occurs at the outer level.
355    
356    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
357        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
358    
359    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
360        doesn't).
361    
362    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
363        to hold a pattern to be incorrectly computed (too small) when there were
364        named back references to duplicated names. This could cause "internal
365        error: code overflow" or "double free or corruption" or other memory
366        handling errors.
367    
368    11. When named subpatterns had the same prefixes, back references could be
369        confused. For example, in this pattern:
370    
371          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
372    
373        the reference to 'Name' was incorrectly treated as a reference to a
374        duplicate name.
375    
376    12. A pattern such as /^s?c/mi8 where the optional character has more than
377        one "other case" was incorrectly compiled such that it would only try to
378        match starting at "c".
379    
380    13. When a pattern starting with \s was studied, VT was not included in the
381        list of possible starting characters; this should have been part of the
382        8.34/18 patch.
383    
384    14. If a character class started [\Qx]... where x is any character, the class
385        was incorrectly terminated at the ].
386    
387    15. If a pattern that started with a caseless match for a character with more
388        than one "other case" was studied, PCRE did not set up the starting code
389        unit bit map for the list of possible characters. Now it does. This is an
390        optimization improvement, not a bug fix.
391    
392    16. The Unicode data tables have been updated to Unicode 7.0.0.
393    
394    17. Fixed a number of memory leaks in pcregrep.
395    
396    18. Avoid a compiler warning (from some compilers) for a function call with
397        a cast that removes "const" from an lvalue by using an intermediate
398        variable (to which the compiler does not object).
399    
400    19. Incorrect code was compiled if a group that contained an internal recursive
401        back reference was optional (had quantifier with a minimum of zero). This
402        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
403        caused segmentation faults because of stack overflows at compile time.
404    
405    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
406        group that is quantified with an indefinite repeat, caused a compile-time
407        loop which used up all the system stack and provoked a segmentation fault.
408        This was not the same bug as 19 above.
409    
410    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
411        Patch by Mike Frysinger.
412    
413    
414    Version 8.35 04-April-2014
415    --------------------------
416    
417    1.  A new flag is set, when property checks are present in an XCLASS.
418        When this flag is not set, PCRE can perform certain optimizations
419        such as studying these XCLASS-es.
420    
421    2.  The auto-possessification of character sets were improved: a normal
422        and an extended character set can be compared now. Furthermore
423        the JIT compiler optimizes more character set checks.
424    
425    3.  Got rid of some compiler warnings for potentially uninitialized variables
426        that show up only when compiled with -O2.
427    
428    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
429        of a match later then the end of the match. The pcretest program was not
430        handling the case sensibly - it was outputting from the start to the next
431        binary zero. It now reports this situation in a message, and outputs the
432        text from the end to the start.
433    
434    5.  Fast forward search is improved in JIT. Instead of the first three
435        characters, any three characters with fixed position can be searched.
436        Search order: first, last, middle.
437    
438    6.  Improve character range checks in JIT. Characters are read by an inprecise
439        function now, which returns with an unknown value if the character code is
440        above a certain threshold (e.g: 256). The only limitation is that the value
441        must be bigger than the threshold as well. This function is useful when
442        the characters above the threshold are handled in the same way.
443    
444    7.  The macros whose names start with RAWUCHAR are placeholders for a future
445        mode in which only the bottom 21 bits of 32-bit data items are used. To
446        make this more memorable for those maintaining the code, the names have
447        been changed to start with UCHAR21, and an extensive comment has been added
448        to their definition.
449    
450    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
451        to the export list in Makefile.am (they were accidentally omitted from the
452        8.34 tarball).
453    
454    9.  The informational output from pcretest used the phrase "starting byte set"
455        which is inappropriate for the 16-bit and 32-bit libraries. As the output
456        for "first char" and "need char" really means "non-UTF-char", I've changed
457        "byte" to "char", and slightly reworded the output. The documentation about
458        these values has also been (I hope) clarified.
459    
460    10. Another JIT related optimization: use table jumps for selecting the correct
461        backtracking path, when more than four alternatives are present inside a
462        bracket.
463    
464    11. Empty match is not possible, when the minimum length is greater than zero,
465        and there is no \K in the pattern. JIT should avoid empty match checks in
466        such cases.
467    
468    12. In a caseless character class with UCP support, when a character with more
469        than one alternative case was not the first character of a range, not all
470        the alternative cases were added to the class. For example, s and \x{17f}
471        are both alternative cases for S: the class [RST] was handled correctly,
472        but [R-T] was not.
473    
474    13. The configure.ac file always checked for pthread support when JIT was
475        enabled. This is not used in Windows, so I have put this test inside a
476        check for the presence of windows.h (which was already tested for).
477    
478    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
479        The algorithm provides a way to skip certain starting offsets, and usually
480        faster than linear prefix searches.
481    
482    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
483        as for 'fr_FR' and 'french'. For some reason, however, it then used the
484        Windows-specific input and output files, which have 'french' screwed in.
485        So this could never have worked. One of the problems with locales is that
486        they aren't always the same. I have now updated RunTest so that it checks
487        the output of the locale test (test 3) against three different output
488        files, and it allows the test to pass if any one of them matches. With luck
489        this should make the test pass on some versions of Solaris where it was
490        failing. Because of the uncertainty, the script did not used to stop if
491        test 3 failed; it now does. If further versions of a French locale ever
492        come to light, they can now easily be added.
493    
494    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
495        there was a message during ./configure, but it did not stop. This now
496        provokes an error. The invalid example in README has been corrected.
497        If a value less than the minimum is given, the minimum value has always
498        been used, but now a warning is given.
499    
500    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
501        was a bug in the test system, which is now fixed. Also, the list of various
502        configurations that are tested for each release did not have one with both
503        16/32 bits and --enable-bar-anycrlf. It now does.
504    
505    18. pcretest was missing "-C bsr" for displaying the \R default setting.
506    
507    19. Little endian PowerPC systems are supported now by the JIT compiler.
508    
509    20. The fast forward newline mechanism could enter to an infinite loop on
510        certain invalid UTF-8 input. Although we don't support these cases
511        this issue can be fixed by a performance optimization.
512    
513    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
514        not take account if existing stack usage. There is now a new global
515        variable called pcre_stack_guard that can be set to point to an external
516        function to check stack availability. It is called at the start of
517        processing every parenthesized group.
518    
519    22. A typo in the code meant that in ungreedy mode the max/min qualifier
520        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
521        not match "ab".
522    
523    23. When UTF was disabled, the JIT program reported some incorrect compile
524        errors. These messages are silenced now.
525    
526    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
527        compiler.
528    
529    25. Change all the temporary files used in RunGrepTest to be different to those
530        used by RunTest so that the tests can be run simultaneously, for example by
531        "make -j check".
532    
533    
534    Version 8.34 15-December-2013
535    -----------------------------
536    
537    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
538        executable memory. Patch inspired by Carsten Klein.
539    
540    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
541        this macro is never tested and has no effect, because the work to support
542        coverage involves only compiling and linking options and special targets in
543        the Makefile. The comment in config.h implied that defining the macro would
544        enable coverage support, which is totally false. There was also support for
545        setting this macro in the CMake files (my fault, I just copied it from
546        configure). SUPPORT_GCOV has now been removed.
547    
548    3.  Make a small performance improvement in strlen16() and strlen32() in
549        pcretest.
550    
551    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
552        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
553        these cases). There was also one in pcretest.c.
554    
555    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
556    
557    6.  In UTF mode, the code for checking whether a group could match an empty
558        string (which is used for indefinitely repeated groups to allow for
559        breaking an infinite loop) was broken when the group contained a repeated
560        negated single-character class with a character that occupied more than one
561        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
562        UTF-8 mode). The effect was undefined: the group might or might not be
563        deemed as matching an empty string, or the program might have crashed.
564    
565    7.  The code for checking whether a group could match an empty string was not
566        recognizing that \h, \H, \v, \V, and \R must match a character.
567    
568    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
569        an empty string. If it can, pcretest shows this in its information output.
570    
571    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
572        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
573        matched by pcre_exec() without using JIT:
574    
575        (a) If the rest of the pattern did not match after a maximal run of
576            grapheme clusters, the code for backing up to try with fewer of them
577            did not always back up over a full grapheme when characters that do not
578            have the modifier quality were involved, e.g. Hangul syllables.
579    
580        (b) If the match point in a subject started with modifier character, and
581            there was no match, the code could incorrectly back up beyond the match
582            point, and potentially beyond the first character in the subject,
583            leading to a segfault or an incorrect match result.
584    
585    10. A conditional group with an assertion condition could lead to PCRE
586        recording an incorrect first data item for a match if no other first data
587        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
588        first data item, and therefore matched "ca" after "c" instead of at the
589        start.
590    
591    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
592        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
593    
594    12. The source of pcregrep now includes z/OS-specific code so that it can be
595        compiled for z/OS as part of the special z/OS distribution.
596    
597    13. Added the -T and -TM options to pcretest.
598    
599    14. The code in pcre_compile.c for creating the table of named capturing groups
600        has been refactored. Instead of creating the table dynamically during the
601        actual compiling pass, the information is remembered during the pre-compile
602        pass (on the stack unless there are more than 20 named groups, in which
603        case malloc() is used) and the whole table is created before the actual
604        compile happens. This has simplified the code (it is now nearly 150 lines
605        shorter) and prepared the way for better handling of references to groups
606        with duplicate names.
607    
608    15. A back reference to a named subpattern when there is more than one of the
609        same name now checks them in the order in which they appear in the pattern.
610        The first one that is set is used for the reference. Previously only the
611        first one was inspected. This change makes PCRE more compatible with Perl.
612    
613    16. Unicode character properties were updated from Unicode 6.3.0.
614    
615    17. The compile-time code for auto-possessification has been refactored, based
616        on a patch by Zoltan Herczeg. It now happens after instead of during
617        compilation. The code is cleaner, and more cases are handled. The option
618        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
619        options in pcretest are provided to set it. It can also be set by
620        (*NO_AUTO_POSSESS) at the start of a pattern.
621    
622    18. The character VT has been added to the default ("C" locale) set of
623        characters that match \s and are generally treated as white space,
624        following this same change in Perl 5.18. There is now no difference between
625        "Perl space" and "POSIX space". Whether VT is treated as white space in
626        other locales depends on the locale.
627    
628    19. The code for checking named groups as conditions, either for being set or
629        for being recursed, has been refactored (this is related to 14 and 15
630        above). Processing unduplicated named groups should now be as fast at
631        numerical groups, and processing duplicated groups should be faster than
632        before.
633    
634    20. Two patches to the CMake build system, by Alexander Barkov:
635    
636          (1) Replace the "source" command by "." in CMakeLists.txt because
637              "source" is a bash-ism.
638    
639          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
640              without these the CMake build does not work on Solaris.
641    
642    21. Perl has changed its handling of \8 and \9. If there is no previously
643        encountered capturing group of those numbers, they are treated as the
644        literal characters 8 and 9 instead of a binary zero followed by the
645        literals. PCRE now does the same.
646    
647    22. Following Perl, added \o{} to specify codepoints in octal, making it
648        possible to specify values greater than 0777 and also making them
649        unambiguous.
650    
651    23. Perl now gives an error for missing closing braces after \x{... instead of
652        treating the string as literal. PCRE now does the same.
653    
654    24. RunTest used to grumble if an inappropriate test was selected explicitly,
655        but just skip it when running all tests. This make it awkward to run ranges
656        of tests when one of them was inappropriate. Now it just skips any
657        inappropriate tests, as it always did when running all tests.
658    
659    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
660        character types such as \d or \w, too many callouts were inserted, and the
661        data that they returned was rubbish.
662    
663    26. In UCP mode, \s was not matching two of the characters that Perl matches,
664        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
665        were matched by \h. The code has now been refactored so that the lists of
666        the horizontal and vertical whitespace characters used for \h and \v (which
667        are defined only in one place) are now also used for \s.
668    
669    27. Add JIT support for the 64 bit TileGX architecture.
670        Patch by Jiong Wang (Tilera Corporation).
671    
672    28. Possessive quantifiers for classes (both explicit and automatically
673        generated) now use special opcodes instead of wrapping in ONCE brackets.
674    
675    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
676        (because it's meaningless), this was not happening when PCRE_CASELESS was
677        set. Not wrong, but inefficient.
678    
679    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
680        properties for \w, \d, etc) is present in a test regex. Otherwise if the
681        test contains no characters greater than 255, Perl doesn't realise it
682        should be using Unicode semantics.
683    
684    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
685        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
686        does in Unicode mode.
687    
688    32. Added the "forbid" facility to pcretest so that putting tests into the
689        wrong test files can sometimes be quickly detected.
690    
691    33. There is now a limit (default 250) on the depth of nesting of parentheses.
692        This limit is imposed to control the amount of system stack used at compile
693        time. It can be changed at build time by --with-parens-nest-limit=xxx or
694        the equivalent in CMake.
695    
696    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
697        errors. Perl warns for these when in warning mode, but PCRE has no facility
698        for giving warnings.
699    
700    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
701        However, this was not working for (?!) because it is optimized to (*FAIL),
702        for which PCRE does not allow quantifiers. The optimization is now disabled
703        when a quantifier follows (?!). I can't see any use for this, but it makes
704        things uniform.
705    
706    36. Perl no longer allows group names to start with digits, so I have made this
707        change also in PCRE. It simplifies the code a bit.
708    
709    37. In extended mode, Perl ignores spaces before a + that indicates a
710        possessive quantifier. PCRE allowed a space before the quantifier, but not
711        before the possessive +. It now does.
712    
713    38. The use of \K (reset reported match start) within a repeated possessive
714        group such as (a\Kb)*+ was not working.
715    
716    40. Document that the same character tables must be used at compile time and
717        run time, and that the facility to pass tables to pcre_exec() and
718        pcre_dfa_exec() is for use only with saved/restored patterns.
719    
720    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
721        features for Builds with MSVC:
722    
723        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
724           builds). The .rc files can be used to set FileDescription and many other
725           attributes.
726    
727        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
728           This allows higher-level build scripts which want .pdb files to avoid
729           hard-coding the exact files needed."
730    
731    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
732        mean "start of word" and "end of word", respectively, as a transition aid.
733    
734    43. A minimizing repeat of a class containing codepoints greater than 255 in
735        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
736        compiled to use the heap for recursion.
737    
738    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
739        is configured.
740    
741    
742    Version 8.33 28-May-2013
743    ------------------------
744    
745    1.  Added 'U' to some constants that are compared to unsigned integers, to
746        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
747        variables that are added to signed variables, to ensure the result is
748        signed and can be negated.
749    
750    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
751        CMake config files.
752    
753    3.  Revise the creation of config.h.generic so that all boolean macros are
754        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
755        overriding via -D on the command line possible.
756    
757    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
758        to unsigned int is reported to make a quite noticeable speed difference in
759        a specific Windows environment. Testing on Linux did also appear to show
760        some benefit (and it is clearly not harmful). Also fixed the definition of
761        Xop which should be unsigned.
762    
763    5.  Related to (4), changing the definition of the intermediate variable cc
764        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
765        improvements.
766    
767    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
768        unnecessary spaces.
769    
770    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
771        and later.
772    
773    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
774    
775    9.  Optimizing fast_forward_start_bits in JIT.
776    
777    10. Adding support for callouts in JIT, and fixing some issues revealed
778        during this work. Namely:
779    
780        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
781    
782        (b) Minimum length was not checked before the matching is started.
783    
784    11. The value of capture_last that is passed to callouts was incorrect in some
785        cases when there was a capture on one path that was subsequently abandoned
786        after a backtrack. Also, the capture_last value is now reset after a
787        recursion, since all captures are also reset in this case.
788    
789    12. The interpreter no longer returns the "too many substrings" error in the
790        case when an overflowing capture is in a branch that is subsequently
791        abandoned after a backtrack.
792    
793    13. In the pathological case when an offset vector of size 2 is used, pcretest
794        now prints out the matched string after a yield of 0 or 1.
795    
796    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
797        Only supported by the JIT compiler at the moment.
798    
799    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
800    
801    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
802        offset of the starting point of the matching process, provided the offsets
803        vector is large enough.
804    
805    17. The \A escape now records a lookbehind value of 1, though its execution
806        does not actually inspect the previous character. This is to ensure that,
807        in partial multi-segment matching, at least one character from the old
808        segment is retained when a new segment is processed. Otherwise, if there
809        are no lookbehinds in the pattern, \A might match incorrectly at the start
810        of a new segment.
811    
812    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
813    
814    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
815        gives some modest performance improvement in 8-bit mode.
816    
817    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
818        be expressed in certain programming languages using Universal Character
819        Names.
820    
821    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
822        which points out that "non characters" are not "characters that may not
823        appear in Unicode strings" but rather "characters that are reserved for
824        internal use and have only local meaning".
825    
826    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
827        there was a conditional group that depended on an assertion, if the
828        assertion was false, the callout that immediately followed the alternation
829        in the condition was skipped when pcre_exec() was used for matching.
830    
831    23. Allow an explicit callout to be inserted before an assertion that is the
832        condition for a conditional group, for compatibility with automatic
833        callouts, which always insert a callout at this point.
834    
835    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
836        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
837    
838    25. (*PRUNE) is now supported by the JIT compiler.
839    
840    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
841    
842    27. Fix the case where there are two or more SKIPs with arguments that may be
843        ignored.
844    
845    28. (*SKIP) is now supported by the JIT compiler.
846    
847    29. (*THEN) is now supported by the JIT compiler.
848    
849    30. Update RunTest with additional test selector options.
850    
851    31. The way PCRE handles backtracking verbs has been changed in two ways.
852    
853        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
854        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
855        backtracking. In some cases this makes it more Perl-compatible, but Perl's
856        rather obscure rules do not always do the same thing.
857    
858        (2) Previously, backtracking verbs were confined within assertions. This is
859        no longer the case for positive assertions, except for (*ACCEPT). Again,
860        this sometimes improves Perl compatibility, and sometimes does not.
861    
862    32. A number of tests that were in test 2 because Perl did things differently
863        have been moved to test 1, because either Perl or PCRE has changed, and
864        these tests are now compatible.
865    
866    32. Backtracking control verbs are now handled in the same way in JIT and
867        interpreter.
868    
869    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
870        contained a forward subroutine reference caused a compile error.
871    
872    34. Auto-detect and optimize limited repetitions in JIT.
873    
874    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
875        blocking (*UTF) etc.
876    
877    36. In the interpreter, maximizing pattern repetitions for characters and
878        character types now use tail recursion, which reduces stack usage.
879    
880    37. The value of the max lookbehind was not correctly preserved if a compiled
881        and saved regex was reloaded on a host of different endianness.
882    
883    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
884        of the compiled pattern block, expand the flags field from 16 to 32 bits
885        because it was almost full.
886    
887    39. Try madvise first before posix_madvise.
888    
889    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
890        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
891        original change locked out all matches of empty strings. This has been
892        changed so that one match of an empty string per line is recognized.
893        Subsequent searches on the same line (for colouring or for --only-matching,
894        for example) do not recognize empty strings.
895    
896    41. Applied a user patch to fix a number of spelling mistakes in comments.
897    
898    42. Data lines longer than 65536 caused pcretest to crash.
899    
900    43. Clarified the data type for length and startoffset arguments for pcre_exec
901        and pcre_dfa_exec in the function-specific man pages, where they were
902        explicitly stated to be in bytes, never having been updated. I also added
903        some clarification to the pcreapi man page.
904    
905    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
906        a segmentation fault.
907    
908    
909    Version 8.32 30-November-2012
910    -----------------------------
911    
912    1.  Improved JIT compiler optimizations for first character search and single
913        character iterators.
914    
915    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
916        Patch by Daniel Richard G.
917    
918    3.  Single character iterator optimizations in the JIT compiler.
919    
920    4.  Improved JIT compiler optimizations for character ranges.
921    
922    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
923        Reported by Giuseppe D'Angelo.
924    
925    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
926        of a line, was being set incorrectly in cases where .* appeared inside
927        atomic brackets at the start of a pattern, or where there was a subsequent
928        *PRUNE or *SKIP.
929    
930    7.  Improved instruction cache flush for POWER/PowerPC.
931        Patch by Daniel Richard G.
932    
933    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
934        grep:
935    
936        (a) There is now no limit to the number of patterns to be matched.
937    
938        (b) An error is given if a pattern is too long.
939    
940        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
941            are now supported.
942    
943        (d) --exclude-from and --include-from (multiple use) have been added.
944    
945        (e) Exclusions and inclusions now apply to all files and directories, not
946            just to those obtained from scanning a directory recursively.
947    
948        (f) Multiple uses of -f and --file-list are now supported.
949    
950        (g) In a Windows environment, the default for -d has been changed from
951            "read" (the GNU grep default) to "skip", because otherwise the presence
952            of a directory in the file list provokes an error.
953    
954        (h) The documentation has been revised and clarified in places.
955    
956    9.  Improve the matching speed of capturing brackets.
957    
958    10. Changed the meaning of \X so that it now matches a Unicode extended
959        grapheme cluster.
960    
961    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
962        out POSIX threads when JIT support is configured.
963    
964    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
965    
966    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
967        structure could go wrong in environments where size_t is not the same size
968        as int.
969    
970    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
971        set.
972    
973    15. The EBCDIC support had decayed; later updates to the code had included
974        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
975        general tidy up of EBCDIC-related issues, and the documentation was also
976        not quite right. There is now a test that can be run on ASCII systems to
977        check some of the EBCDIC-related things (but is it not a full test).
978    
979    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
980        in a small tidy to the code.
981    
982    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
983    
984    18. If the --only-matching (-o) option in pcregrep is specified multiple
985        times, each one causes appropriate output. For example, -o1 -o2 outputs the
986        substrings matched by the 1st and 2nd capturing parentheses. A separating
987        string can be specified by --om-separator (default empty).
988    
989    19. Improving the first n character searches.
990    
991    20. Turn case lists for horizontal and vertical white space into macros so that
992        they are defined only once.
993    
994    21. This set of changes together give more compatible Unicode case-folding
995        behaviour for characters that have more than one other case when UCP
996        support is available.
997    
998        (a) The Unicode property table now has offsets into a new table of sets of
999            three or more characters that are case-equivalent. The MultiStage2.py
1000            script that generates these tables (the pcre_ucd.c file) now scans
1001            CaseFolding.txt instead of UnicodeData.txt for character case
1002            information.
1003    
1004        (b) The code for adding characters or ranges of characters to a character
1005            class has been abstracted into a generalized function that also handles
1006            case-independence. In UTF-mode with UCP support, this uses the new data
1007            to handle characters with more than one other case.
1008    
1009        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1010            whose other case is greater than 256 are now correctly matched
1011            caselessly. Previously, the high codepoint matched the low one, but not
1012            vice versa.
1013    
1014        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1015            of the new class addition function, using character lists defined as
1016            macros alongside the case definitions of 20 above.
1017    
1018        (e) Caseless back references now work with characters that have more than
1019            one other case.
1020    
1021        (f) General caseless matching of characters with more than one other case
1022            is supported.
1023    
1024    22. Unicode character properties were updated from Unicode 6.2.0
1025    
1026    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1027    
1028    24. Add support for 32-bit character strings, and UTF-32
1029    
1030    25. Major JIT compiler update (code refactoring and bugfixing).
1031        Experimental Sparc 32 support is added.
1032    
1033    26. Applied a modified version of Daniel Richard G's patch to create
1034        pcre.h.generic and config.h.generic by "make" instead of in the
1035        PrepareRelease script.
1036    
1037    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1038        pcre_compile.c when checking for a zero character.
1039    
1040    28. Introducing a native interface for JIT. Through this interface, the compiled
1041        machine code can be directly executed. The purpose of this interface is to
1042        provide fast pattern matching, so several sanity checks are not performed.
1043        However, feature tests are still performed. The new interface provides
1044        1.4x speedup compared to the old one.
1045    
1046    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1047        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1048        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1049    
1050    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1051        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1052        the "old" RFC 2279). Instead, it ended up passing a negative length to
1053        pcre_exec().
1054    
1055    31. Add support for GCC's visibility feature to hide internal functions.
1056    
1057    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1058        "unknown -C option" after outputting 0 or 1.
1059    
1060    33. There is now support for generating a code coverage report for the test
1061        suite in environments where gcc is the compiler and lcov is installed. This
1062        is mainly for the benefit of the developers.
1063    
1064    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1065        unaddressable using valgrind annotations, allowing valgrind to detect
1066        invalid memory accesses. This is mainly for the benefit of the developers.
1067    
1068    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1069    
1070    26. Give configure error if --enable-cpp but no C++ compiler found.
1071    
1072    
1073    Version 8.31 06-July-2012
1074    -------------------------
1075    
1076    1.  Fixing a wrong JIT test case and some compiler warnings.
1077    
1078    2.  Removed a bashism from the RunTest script.
1079    
1080    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1081        to unsigned type, result still unsigned" that was given by an MS compiler
1082        on encountering the code "-sizeof(xxx)".
1083    
1084    4.  Partial matching support is added to the JIT compiler.
1085    
1086    5.  Fixed several bugs concerned with partial matching of items that consist
1087        of more than one character:
1088    
1089        (a) /^(..)\1/ did not partially match "aba" because checking references was
1090            done on an "all or nothing" basis. This also applied to repeated
1091            references.
1092    
1093        (b) \R did not give a hard partial match if \r was found at the end of the
1094            subject.
1095    
1096        (c) \X did not give a hard partial match after matching one or more
1097            characters at the end of the subject.
1098    
1099        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1100            a partial match for the string "\r".
1101    
1102        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1103            a partial match for a CR character at the end of the subject string.
1104    
1105    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1106        running pcretest, the text "(JIT)" added to the output whenever JIT is
1107        actually used to run the match.
1108    
1109    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1110        or /S+[+] with a digit between 1 and 7.
1111    
1112    8.  OP_NOT now supports any UTF character not just single-byte ones.
1113    
1114    9.  (*MARK) control verb is now supported by the JIT compiler.
1115    
1116    10. The command "./RunTest list" lists the available tests without actually
1117        running any of them. (Because I keep forgetting what they all are.)
1118    
1119    11. Add PCRE_INFO_MAXLOOKBEHIND.
1120    
1121    12. Applied a (slightly modified) user-supplied patch that improves performance
1122        when the heap is used for recursion (compiled with --disable-stack-for-
1123        recursion). Instead of malloc and free for each heap frame each time a
1124        logical recursion happens, frames are retained on a chain and re-used where
1125        possible. This sometimes gives as much as 30% improvement.
1126    
1127    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1128        call.
1129    
1130    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1131    
1132    15. It is now possible to link pcretest with libedit as an alternative to
1133        libreadline.
1134    
1135    16. (*COMMIT) control verb is now supported by the JIT compiler.
1136    
1137    17. The Unicode data tables have been updated to Unicode 6.1.0.
1138    
1139    18. Added --file-list option to pcregrep.
1140    
1141    19. Added binary file support to pcregrep, including the -a, --binary-files,
1142        -I, and --text options.
1143    
1144    20. The madvise function is renamed for posix_madvise for QNX compatibility
1145        reasons. Fixed by Giuseppe D'Angelo.
1146    
1147    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1148    
1149    22. Changed the option for creating symbolic links for 16-bit man pages from
1150        -s to -sf so that re-installing does not cause issues.
1151    
1152    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1153    
1154    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1155        matches in certain environments (the workspace was not being correctly
1156        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1157        some of the workspace data at the beginning of a restart.
1158    
1159    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1160        was not doing so when it should - probably a typo introduced by SVN 528
1161        (change 8.10/14).
1162    
1163    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1164        \w+ when the character tables indicated that \x{c4} was a word character.
1165        There were several related cases, all because the tests for doing a table
1166        lookup were testing for characters less than 127 instead of 255.
1167    
1168    27. If a pattern contains capturing parentheses that are not used in a match,
1169        their slots in the ovector are set to -1. For those that are higher than
1170        any matched groups, this happens at the end of processing. In the case when
1171        there were back references that the ovector was too small to contain
1172        (causing temporary malloc'd memory to be used during matching), and the
1173        highest capturing number was not used, memory off the end of the ovector
1174        was incorrectly being set to -1. (It was using the size of the temporary
1175        memory instead of the true size.)
1176    
1177    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1178        ovector size, it uses memory at the end of the block that it has got.
1179    
1180    29. Check for an overlong MARK name and give an error at compile time. The
1181        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1182    
1183    30. JIT compiler update.
1184    
1185    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1186        Rill for the patch.
1187    
1188    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1189    
1190    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1191    
1192    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1193        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1194        was enabled.
1195    
1196    35. Improve JIT code generation for greedy plus quantifier.
1197    
1198    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1199        1 to "aa" instead of to an empty string. The bug affected repeated groups
1200        that could potentially match an empty string.
1201    
1202    37. Optimizing single character iterators in JIT.
1203    
1204    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1205        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1206        codepoints that are too big for the mode are faulted, and in a UTF mode,
1207        disallowed codepoints are also faulted.
1208    
1209    39. If PCRE was compiled with UTF support, in three places in the DFA
1210        matcher there was code that should only have been obeyed in UTF mode, but
1211        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1212        processing when bytes with values greater than 127 were present. In 16-bit
1213        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1214        both cases the values are those that cannot be the first data item in a UTF
1215        character. The three items that might have provoked this were recursions,
1216        possessively repeated groups, and atomic groups.
1217    
1218    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1219        and pcregrep, because some OS require shared objects to be explicitly
1220        passed to ld, causing the link step to fail if they are not.
1221    
1222    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1223        mode, patterns that started with \h* or \R* might be incorrectly matched.
1224    
1225    
1226    Version 8.30 04-February-2012
1227    -----------------------------
1228    
1229    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1230        name is defined in ctype.h.
1231    
1232    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1233        only in quite long subpatterns.
1234    
1235    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1236        since it was replaced by pcre_fullinfo() in February 2000.
1237    
1238    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1239        match a (*MARK), and the match failed at the start of the subject, a
1240        reference to memory before the start of the subject could occur. This bug
1241        was introduced by fix 17 of release 8.21.
1242    
1243    5.  A reference to an unset group with zero minimum repetition was giving
1244        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1245        /(another)?(\1?)test/ matched against "hello world test". This bug was
1246        introduced in release 8.13.
1247    
1248    6.  Add support for 16-bit character strings (a large amount of work involving
1249        many changes and refactorings).
1250    
1251    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1252        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1253        from a file.
1254    
1255    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1256        rounding is not applied in this particular case).
1257    
1258    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1259        if they appear, or are escaped, in patterns.
1260    
1261    10. Get rid of a number of -Wunused-but-set-variable warnings.
1262    
1263    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1264        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1265        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1266        also returns the mark "x". This bug applied to capturing parentheses,
1267        non-capturing parentheses, and atomic parentheses. It also applied to some
1268        assertions.
1269    
1270    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1271        information out of configure.ac instead of relying on pcre.h.generic, which
1272        is not stored in the repository.
1273    
1274    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1275        -lreadline.
1276    
1277    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1278    
1279    15. Applied Graycode's patch to put the top-level frame on the stack rather
1280        than the heap when not using the stack for recursion. This gives a
1281        performance improvement in many cases when recursion is not deep.
1282    
1283    16. Experimental code added to "pcretest -C" to output the stack frame size.
1284    
1285    
1286    Version 8.21 12-Dec-2011
1287    ------------------------
1288    
1289    1.  Updating the JIT compiler.
1290    
1291    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1292        are added as well.
1293    
1294    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1295        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1296        calling _pcre_jit_exec. Some extra comments are added.
1297    
1298    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1299        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1300        was introduced by change 18 for 8.20.
1301    
1302    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1303        ECMA-262 standard.
1304    
1305    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1306        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1307        This bug was probably introduced by change 9 of 8.13.
1308    
1309    7.  While fixing 6 above, I noticed that a number of other items were being
1310        incorrectly rejected as "not fixed length". This arose partly because newer
1311        opcodes had not been added to the fixed-length checking code. I have (a)
1312        corrected the bug and added tests for these items, and (b) arranged for an
1313        error to occur if an unknown opcode is encountered while checking for fixed
1314        length instead of just assuming "not fixed length". The items that were
1315        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1316        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1317        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1318    
1319    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1320        being incorrectly compiled and would have given unpredicatble results.
1321    
1322    9.  A possessively repeated subpattern with minimum repeat count greater than
1323        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1324        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1325        the first (A) could occur when it should not.
1326    
1327    10. Add a cast and remove a redundant test from the code.
1328    
1329    11. JIT should use pcre_malloc/pcre_free for allocation.
1330    
1331    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1332        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1333        is anything other than /usr, -L is still shown).
1334    
1335    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1336    
1337    14. Perl does not support \N without a following name in a [] class; PCRE now
1338        also gives an error.
1339    
1340    15. If a forward reference was repeated with an upper limit of around 2000,
1341        it caused the error "internal error: overran compiling workspace". The
1342        maximum number of forward references (including repeats) was limited by the
1343        internal workspace, and dependent on the LINK_SIZE. The code has been
1344        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1345        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1346        of around 200,000 forward references. While doing this, I also speeded up
1347        the filling in of repeated forward references.
1348    
1349    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1350        incorrectly expecting the subject to contain another "a" after the start.
1351    
1352    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1353        in the match, the SKIP should be ignored. This was not happening; instead
1354        the SKIP was being treated as NOMATCH. For patterns such as
1355        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1356        tested.
1357    
1358    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1359        now much more compatible with Perl, in particular in cases where the result
1360        is a non-match for a non-anchored pattern. For example, if
1361        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1362        "m", where previously it did not return a name. A side effect of this
1363        change is that for partial matches, the last encountered mark name is
1364        returned, as for non matches. A number of tests that were previously not
1365        Perl-compatible have been moved into the Perl-compatible test files. The
1366        refactoring has had the pleasing side effect of removing one argument from
1367        the match() function, thus reducing its stack requirements.
1368    
1369    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1370        subsequent uses of /S (without +) incorrectly behaved like /S+.
1371    
1372    21. Retrieve executable code size support for the JIT compiler and fixing
1373        some warnings.
1374    
1375    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1376        not work when the shorter character appeared right at the end of the
1377        subject string.
1378    
1379    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1380        systems.
1381    
1382    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1383        output it when the /M option is used in pcretest.
1384    
1385    25. The CheckMan script was not being included in the distribution. Also, added
1386        an explicit "perl" to run Perl scripts from the PrepareRelease script
1387        because this is reportedly needed in Windows.
1388    
1389    26. If study data was being save in a file and studying had not found a set of
1390        "starts with" bytes for the pattern, the data written to the file (though
1391        never used) was taken from uninitialized memory and so caused valgrind to
1392        complain.
1393    
1394    27. Updated RunTest.bat as provided by Sheri Pierce.
1395    
1396    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1397    
1398    29. Computation of memory usage for the table of capturing group names was
1399        giving an unnecessarily large value.
1400    
1401    
1402    Version 8.20 21-Oct-2011
1403    ------------------------
1404    
1405    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1406        a POSIX class. After further experiments with Perl, which convinced me that
1407        Perl has bugs and confusions, a closing square bracket is no longer allowed
1408        in a POSIX name. This bug also affected patterns with classes that started
1409        with full stops.
1410    
1411    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1412        captured substring, but while checking the failing first alternative,
1413        substring 1 is temporarily captured. If the output vector supplied to
1414        pcre_exec() was not big enough for this capture, the yield of the function
1415        was still zero ("insufficient space for captured substrings"). This cannot
1416        be totally fixed without adding another stack variable, which seems a lot
1417        of expense for a edge case. However, I have improved the situation in cases
1418        such as /(a)(b)x|abc/ matched against "abc", where the return code
1419        indicates that fewer than the maximum number of slots in the ovector have
1420        been set.
1421    
1422    3.  Related to (2) above: when there are more back references in a pattern than
1423        slots in the output vector, pcre_exec() uses temporary memory during
1424        matching, and copies in the captures as far as possible afterwards. It was
1425        using the entire output vector, but this conflicts with the specification
1426        that only 2/3 is used for passing back captured substrings. Now it uses
1427        only the first 2/3, for compatibility. This is, of course, another edge
1428        case.
1429    
1430    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1431        main code base, and can be used by building with --enable-jit. When this is
1432        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1433        runtime --no-jit option is given.
1434    
1435    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1436        ovector, the return from the function was zero, implying that there were
1437        other matches that did not fit. The correct "exactly full" value is now
1438        returned.
1439    
1440    6.  If a subpattern that was called recursively or as a subroutine contained
1441        (*PRUNE) or any other control that caused it to give a non-standard return,
1442        invalid errors such as "Error -26 (nested recursion at the same subject
1443        position)" or even infinite loops could occur.
1444    
1445    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1446        computing the minimum length on reaching *ACCEPT, and so ended up with the
1447        wrong value of 1 rather than 0. Further investigation indicates that
1448        computing a minimum subject length in the presence of *ACCEPT is difficult
1449        (think back references, subroutine calls), and so I have changed the code
1450        so that no minimum is registered for a pattern that contains *ACCEPT.
1451    
1452    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1453        it was not handled as intended. [But see 16 below.]
1454    
1455    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1456        Sheri Pierce.
1457    
1458    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1459        the first byte in a match must be "a".
1460    
1461    11. Change 17 for 8.13 increased the recursion depth for patterns like
1462        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1463        pattern contains any instances of (*THEN). If it does not, the old
1464        optimizations are restored. It would be nice to do this on a per-group
1465        basis, but at the moment that is not feasible.
1466    
1467    12. In some environments, the output of pcretest -C is CRLF terminated. This
1468        broke RunTest's code that checks for the link size. A single white space
1469        character after the value is now allowed for.
1470    
1471    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1472        For "fr", it uses the Windows-specific input and output files.
1473    
1474    14. If (*THEN) appeared in a group that was called recursively or as a
1475        subroutine, it did not work as intended. [But see next item.]
1476    
1477    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1478        pattern fragments (but not containing any | characters). If A and B are
1479        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1480        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1481        D. In other words, Perl considers parentheses that do not contain any |
1482        characters to be part of a surrounding alternative, whereas PCRE was
1483        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1484        differently. PCRE now behaves in the same way as Perl, except in the case
1485        of subroutine/recursion calls such as (?1) which have in any case always
1486        been different (but PCRE had them first :-).
1487    
1488    16. Related to 15 above: Perl does not treat the | in a conditional group as
1489        creating alternatives. Such a group is treated in the same way as an
1490        ordinary group without any | characters when processing (*THEN). PCRE has
1491        been changed to match Perl's behaviour.
1492    
1493    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1494        RunGrepTest script failed.
1495    
1496    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1497        inevitable for groups that contain captures, but it can lead to a lot of
1498        stack use in large patterns. The old behaviour has been restored for atomic
1499        groups that do not contain any capturing parentheses.
1500    
1501    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1502        suppress the check for a minimum subject length at run time. (If it was
1503        given to pcre_exec() or pcre_dfa_exec() it did work.)
1504    
1505    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1506        fail to work when decoding hex characters in data strings in EBCDIC
1507        environments.
1508    
1509    21. It appears that in at least one Mac OS environment, the isxdigit() function
1510        is implemented as a macro that evaluates to its argument more than once,
1511        contravening the C 90 Standard (I haven't checked a later standard). There
1512        was an instance in pcretest which caused it to go wrong when processing
1513        \x{...} escapes in subject strings. The has been rewritten to avoid using
1514        things like p++ in the argument of isxdigit().
1515    
1516    
1517    Version 8.13 16-Aug-2011
1518    ------------------------
1519    
1520    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1521    
1522    2.  Two minor typos in pcre_internal.h have been fixed.
1523    
1524    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1525        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1526        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1527    
1528    4.  There were a number of related bugs in the code for matching backrefences
1529        caselessly in UTF-8 mode when codes for the characters concerned were
1530        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1531        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1532        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1533        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1534        2-byte code at the end of the subject (it thought there wasn't enough data
1535        left).
1536    
1537    5.  Comprehensive information about what went wrong is now returned by
1538        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1539        as the output vector has at least 2 elements. The offset of the start of
1540        the failing character and a reason code are placed in the vector.
1541    
1542    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1543        now returned is for the first byte of the failing character, instead of the
1544        last byte inspected. This is an incompatible change, but I hope it is small
1545        enough not to be a problem. It makes the returned offset consistent with
1546        pcre_exec() and pcre_dfa_exec().
1547    
1548    7.  pcretest now gives a text phrase as well as the error number when
1549        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1550        failure, the offset and reason code are output.
1551    
1552    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1553        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1554        back over a single character (\n). This seems wrong (because it treated the
1555        two characters as a single entity when going forwards), conflicts with the
1556        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1557        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1558        has been changed.
1559    
1560    9.  Some internal refactoring has changed the processing so that the handling
1561        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1562        time (the PCRE_DOTALL option was changed this way some time ago: version
1563        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1564        which was always a bit of a fudge. It also means that there is one less
1565        argument for the match() function, which reduces its stack requirements
1566        slightly. This change also fixes an incompatibility with Perl: the pattern
1567        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1568    
1569    10. More internal refactoring has drastically reduced the number of recursive
1570        calls to match() for possessively repeated groups such as (abc)++ when
1571        using pcre_exec().
1572    
1573    11. While implementing 10, a number of bugs in the handling of groups were
1574        discovered and fixed:
1575    
1576        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1577        (a|)*(?1) gave a compile-time internal error.
1578        ((a|)+)+  did not notice that the outer group could match an empty string.
1579        (^a|^)+   was not marked as anchored.
1580        (.*a|.*)+ was not marked as matching at start or after a newline.
1581    
1582    12. Yet more internal refactoring has removed another argument from the match()
1583        function. Special calls to this function are now indicated by setting a
1584        value in a variable in the "match data" data block.
1585    
1586    13. Be more explicit in pcre_study() instead of relying on "default" for
1587        opcodes that mean there is no starting character; this means that when new
1588        ones are added and accidentally left out of pcre_study(), testing should
1589        pick them up.
1590    
1591    14. The -s option of pcretest has been documented for ages as being an old
1592        synonym of -m (show memory usage). I have changed it to mean "force study
1593        for every regex", that is, assume /S for every regex. This is similar to -i
1594        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1595        using it. It makes it easier to run collections of tests with and without
1596        study enabled, and thereby test pcre_study() more easily. All the standard
1597        tests are now run with and without -s (but some patterns can be marked as
1598        "never study" - see 20 below).
1599    
1600    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1601        restoration of the capturing data to the outer values was not happening
1602        correctly.
1603    
1604    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1605        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1606        pattern had matched an empty string, and so incorrectly returned a no
1607        match.
1608    
1609    17. There was optimizing code for the last branch of non-capturing parentheses,
1610        and also for the obeyed branch of a conditional subexpression, which used
1611        tail recursion to cut down on stack usage. Unfortunately, now that there is
1612        the possibility of (*THEN) occurring in these branches, tail recursion is
1613        no longer possible because the return has to be checked for (*THEN). These
1614        two optimizations have therefore been removed. [But see 8.20/11 above.]
1615    
1616    18. If a pattern containing \R was studied, it was assumed that \R always
1617        matched two bytes, thus causing the minimum subject length to be
1618        incorrectly computed because \R can also match just one byte.
1619    
1620    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1621        was incorrectly computed.
1622    
1623    20. If /S is present twice on a test pattern in pcretest input, it now
1624        *disables* studying, thereby overriding the use of -s on the command line
1625        (see 14 above). This is necessary for one or two tests to keep the output
1626        identical in both cases.
1627    
1628    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1629        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1630    
1631    22. When an atomic group that contained a capturing parenthesis was
1632        successfully matched, but the branch in which it appeared failed, the
1633        capturing was not being forgotten if a higher numbered group was later
1634        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1635        group 1 to "a", when in fact it should be unset. This applied to multi-
1636        branched capturing and non-capturing groups, repeated or not, and also to
1637        positive assertions (capturing in negative assertions does not happen
1638        in PCRE) and also to nested atomic groups.
1639    
1640    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1641        subject after a captured substring, to make it easier to tell which of a
1642        number of identical substrings has been captured.
1643    
1644    24. The way atomic groups are processed by pcre_exec() has been changed so that
1645        if they are repeated, backtracking one repetition now resets captured
1646        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1647        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1648        "aaa". Previously, it would have been "a". As part of this code
1649        refactoring, the way recursive calls are handled has also been changed.
1650    
1651    25. If an assertion condition captured any substrings, they were not passed
1652        back unless some other capturing happened later. For example, if
1653        (?(?=(a))a) was matched against "a", no capturing was returned.
1654    
1655    26. When studying a pattern that contained subroutine calls or assertions,
1656        the code for finding the minimum length of a possible match was handling
1657        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1658        group 1 called group 2 while simultaneously a separate group 2 called group
1659        1). A stack overflow occurred in this case. I have fixed this by limiting
1660        the recursion depth to 10.
1661    
1662    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1663        Fortmann. This supports explicit test numbers on the command line, and has
1664        argument validation and error reporting.
1665    
1666    28. An instance of \X with an unlimited repeat could fail if at any point the
1667        first character it looked at was a mark character.
1668    
1669    29. Some minor code refactoring concerning Unicode properties and scripts
1670        should reduce the stack requirement of match() slightly.
1671    
1672    30. Added the '=' option to pcretest to check the setting of unused capturing
1673        slots at the end of the pattern, which are documented as being -1, but are
1674        not included in the return count.
1675    
1676    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1677        compiled something random. Now it gives a compile-time error (as does
1678        Perl).
1679    
1680    32. A *MARK encountered during the processing of a positive assertion is now
1681        recorded and passed back (compatible with Perl).
1682    
1683    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1684        had alternative anchored branches, the search for a second match in a line
1685        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1686        matched the line "0102" twice. The same bug affected patterns that started
1687        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1688        twice.
1689    
1690    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1691        does, and because of capturing effects, quantifying parenthesized
1692        assertions may at times be useful. Quantifiers are now allowed for
1693        parenthesized assertions.
1694    
1695    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1696    
1697    36. \g was being checked for fancy things in a character class, when it should
1698        just be a literal "g".
1699    
1700    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1701        appearance of a nested POSIX class supersedes an apparent external class.
1702        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1703        unescaped square brackets may also appear as part of class names. For
1704        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1705        more like Perl. (But see 8.20/1 above.)
1706    
1707    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1708        was because it thought it was \N{name}, which is not supported).
1709    
1710    39. Add minix to OS list not supporting the -S option in pcretest.
1711    
1712    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1713        cannot analyze patterns in sufficient detail to catch mutual recursions
1714        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1715        subgroup is called recursively as a subpattern for a second time at the
1716        same position in the subject string. In previous releases this might have
1717        been caught by the recursion limit, or it might have run out of stack.
1718    
1719    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1720        happen only once. PCRE was, however incorrectly giving a compile time error
1721        "recursive call could loop indefinitely" because it cannot analyze the
1722        pattern in sufficient detail. The compile time test no longer happens when
1723        PCRE is compiling a conditional subpattern, but actual runaway loops are
1724        now caught at runtime (see 40 above).
1725    
1726    42. It seems that Perl allows any characters other than a closing parenthesis
1727        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1728        has been changed to be the same.
1729    
1730    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1731        as not to get warnings when autogen.sh is called. Also changed
1732        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1733    
1734    44. To help people who use pcregrep to scan files containing exceedingly long
1735        lines, the following changes have been made:
1736    
1737        (a) The default value of the buffer size parameter has been increased from
1738            8K to 20K. (The actual buffer used is three times this size.)
1739    
1740        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1741            PCRE is built.
1742    
1743        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1744            to be set at run time.
1745    
1746        (d) Numerical values in pcregrep options can be followed by K or M, for
1747            example --buffer-size=50K.
1748    
1749        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1750            given and the return code is set to 2.
1751    
1752    45. Add a pointer to the latest mark to the callout data block.
1753    
1754    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1755        partial match of an empty string instead of no match. This was specific to
1756        the use of ".".
1757    
1758    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1759        complete match instead of a partial match. This bug was dependent on both
1760        the PCRE_UTF8 and PCRE_DOTALL options being set.
1761    
1762    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1763        starting byte set, because \b was not being ignored.
1764    
1765    
1766    Version 8.12 15-Jan-2011
1767    ------------------------
1768    
1769    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1770        checks for such things as part of the documentation building process.
1771    
1772    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1773        --match-limit and --recursion-limit options (added for 8.11). In
1774        particular, this made one of the standard tests fail. (The integer value
1775        went into the wrong half of a long int.)
1776    
1777    3.  If the --colour option was given to pcregrep with -v (invert match), it
1778        did strange things, either producing crazy output, or crashing. It should,
1779        of course, ignore a request for colour when reporting lines that do not
1780        match.
1781    
1782    4.  Another pcregrep bug caused similar problems if --colour was specified with
1783        -M (multiline) and the pattern match finished with a line ending.
1784    
1785    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1786        matched in multiline mode, the following line was shown as part of the
1787        match. This seems wrong, so I have changed it.
1788    
1789    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1790        the check for further matches in the same line (so they could be coloured)
1791        to overrun the end of the current line. If another match was found, it was
1792        incorrectly shown (and then shown again when found in the next line).
1793    
1794    7.  If pcregrep was compiled under Windows, there was a reference to the
1795        function pcregrep_exit() before it was defined. I am assuming this was
1796        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1797        reported by a user. I've moved the definition above the reference.
1798    
1799    
1800    Version 8.11 10-Dec-2010
1801    ------------------------
1802    
1803    1.  (*THEN) was not working properly if there were untried alternatives prior
1804        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1805        backtracked to try for "b" instead of moving to the next alternative branch
1806        at the same level (in this case, to look for "c"). The Perl documentation
1807        is clear that when (*THEN) is backtracked onto, it goes to the "next
1808        alternative in the innermost enclosing group".
1809    
1810    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1811        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1812        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1813        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1814        (*THEN).
1815    
1816    3.  If \s appeared in a character class, it removed the VT character from
1817        the class, even if it had been included by some previous item, for example
1818        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1819        of \s, but is part of the POSIX "space" class.)
1820    
1821    4.  A partial match never returns an empty string (because you can always
1822        match an empty string at the end of the subject); however the checking for
1823        an empty string was starting at the "start of match" point. This has been
1824        changed to the "earliest inspected character" point, because the returned
1825        data for a partial match starts at this character. This means that, for
1826        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1827        (previously it gave "no match").
1828    
1829    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1830        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1831        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1832        has an implication that the given string is incomplete (because a partial
1833        match is preferred over a full match). For this reason, these items now
1834        give a partial match in this situation. [Aside: previously, the one case
1835        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1836        match rather than a full match, which was wrong by the old rules, but is
1837        now correct.]
1838    
1839    6.  There was a bug in the handling of #-introduced comments, recognized when
1840        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1841        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1842        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1843        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1844        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1845        places in pcre_compile().
1846    
1847    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1848        comments when looking ahead for named forward references to subpatterns,
1849        the only newline sequence it recognized was NL. It now handles newlines
1850        according to the set newline convention.
1851    
1852    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1853        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1854        cater for a lack of strerror(). These oversights have been fixed.
1855    
1856    9.  Added --match-limit and --recursion-limit to pcregrep.
1857    
1858    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1859    
1860    11. When the -o option was used, pcregrep was setting a return code of 1, even
1861        when matches were found, and --line-buffered was not being honoured.
1862    
1863    12. Added an optional parentheses number to the -o and --only-matching options
1864        of pcregrep.
1865    
1866    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1867        can match an empty string. The code to do it in pcretest and pcredemo
1868        needed fixing:
1869    
1870        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1871            only one byte after an empty string match just before CRLF (this case
1872            just got forgotten; "any" and "anycrlf" were OK).
1873    
1874        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1875            mode when an empty string match preceded an ASCII character followed by
1876            a non-ASCII character. (The code for advancing by one character rather
1877            than one byte was nonsense.)
1878    
1879        (c) The pcredemo.c sample program did not have any code at all to handle
1880            the cases when CRLF is a valid newline sequence.
1881    
1882    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1883        as a starting offset was within the subject string. There is now a new
1884        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1885        negative or greater than the length of the string. In order to test this,
1886        pcretest is extended to allow the setting of negative starting offsets.
1887    
1888    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1889        starting offset points to the beginning of a UTF-8 character was
1890        unnecessarily clumsy. I tidied it up.
1891    
1892    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1893        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1894    
1895    17. Nobody had reported that the --include_dir option, which was added in
1896        release 7.7 should have been called --include-dir (hyphen, not underscore)
1897        for compatibility with GNU grep. I have changed it to --include-dir, but
1898        left --include_dir as an undocumented synonym, and the same for
1899        --exclude-dir, though that is not available in GNU grep, at least as of
1900        release 2.5.4.
1901    
1902    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1903        characters from a string of bytes) have been redefined so as not to use
1904        loops, in order to improve performance in some environments. At the same
1905        time, I abstracted some of the common code into auxiliary macros to save
1906        repetition (this should not affect the compiled code).
1907    
1908    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1909        compile-time error is now given if \c is not followed by an ASCII
1910        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1911        different, and any byte value is allowed.)
1912    
1913    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1914        START_OPTIMIZE option, which is now allowed at compile time - but just
1915        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1916        to pcregrep and other applications that have no direct access to PCRE
1917        options. The new /Y option in pcretest sets this option when calling
1918        pcre_compile().
1919    
1920    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1921        back references. Groups containing recursive back references were forced to
1922        be atomic by that change, but in the case of named groups, the amount of
1923        memory required was incorrectly computed, leading to "Failed: internal
1924        error: code overflow". This has been fixed.
1925    
1926    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1927        pcretest.c, to avoid build problems in some Borland environments.
1928    
1929    
1930    Version 8.10 25-Jun-2010
1931    ------------------------
1932    
1933    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1934        THEN.
1935    
1936    2.  (*ACCEPT) was not working when inside an atomic group.
1937    
1938    3.  Inside a character class, \B is treated as a literal by default, but
1939        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1940        causes the error). The code is unchanged, but I tidied the documentation.
1941    
1942    4.  Inside a character class, PCRE always treated \R and \X as literals,
1943        whereas Perl faults them if its -w option is set. I have changed PCRE so
1944        that it faults them when PCRE_EXTRA is set.
1945    
1946    5.  Added support for \N, which always matches any character other than
1947        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1948    
1949    6.  When compiling pcregrep with newer versions of gcc which may have
1950        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1951        declared with attribute warn_unused_result" were given. Just casting the
1952        result to (void) does not stop the warnings; a more elaborate fudge is
1953        needed. I've used a macro to implement this.
1954    
1955    7.  Minor change to pcretest.c to avoid a compiler warning.
1956    
1957    8.  Added four artifical Unicode properties to help with an option to make
1958        \s etc use properties (see next item). The new properties are: Xan
1959        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1960    
1961    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1962        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1963        this option. Modified pcretest to add /W to test this facility. Added
1964        REG_UCP to make it available via the POSIX interface.
1965    
1966    10. Added --line-buffered to pcregrep.
1967    
1968    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1969        studied, and the match started with a letter with a code point greater than
1970        127 whose first byte was different to the first byte of the other case of
1971        the letter, the other case of this starting letter was not recognized
1972        (#976).
1973    
1974    12. If a pattern that was studied started with a repeated Unicode property
1975        test, for example, \p{Nd}+, there was the theoretical possibility of
1976        setting up an incorrect bitmap of starting bytes, but fortunately it could
1977        not have actually happened in practice until change 8 above was made (it
1978        added property types that matched character-matching opcodes).
1979    
1980    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1981        possible starting bytes for non-anchored patterns.
1982    
1983    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1984        \R, and also a number of cases that involve Unicode properties, both
1985        explicit and implicit when PCRE_UCP is set.
1986    
1987    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1988        input, it could crash or give wrong results if characters with values
1989        greater than 0xc0 were present in the subject string. (Detail: it assumed
1990        UTF-8 input when processing these items.)
1991    
1992    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1993        size_t is 64-bit (#991).
1994    
1995    17. Added a check for running out of memory when PCRE is compiled with
1996        --disable-stack-for-recursion (#990).
1997    
1998    18. If the last data line in a file for pcretest does not have a newline on
1999        the end, a newline was missing in the output.
2000    
2001    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2002        less than 128) in its various bitmaps. However, there is a facility for
2003        generating tables according to the current locale when PCRE is compiled. It
2004        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2005        characters, are recognized by isspace() and therefore were getting set in
2006        these tables, and indeed these tables seem to approximate to ISO 8859. This
2007        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2008        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2009        which of course cannot start UTF-8 characters. I have changed the code so
2010        that only real ASCII characters (less than 128) and the correct starting
2011        bytes for UTF-8 encodings are set for characters greater than 127 when in
2012        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2013        altogether.)
2014    
2015    20. Added the /T option to pcretest so as to be able to run tests with non-
2016        standard character tables, thus making it possible to include the tests
2017        used for 19 above in the standard set of tests.
2018    
2019    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2020        reference to a subpattern the other side of a comment that contains an
2021        opening parenthesis caused either an internal compiling error, or a
2022        reference to the wrong subpattern.
2023    
2024    
2025    Version 8.02 19-Mar-2010
2026    ------------------------
2027    
2028    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2029    
2030    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2031        configured.
2032    
2033    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2034        original author of that file, following a query about its status.
2035    
2036    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2037        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2038    
2039    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2040        quantifier applied to a forward-referencing subroutine call, could compile
2041        incorrect code or give the error "internal error: previously-checked
2042        referenced subpattern not found".
2043    
2044    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2045        variables to point to external functions. For these systems, therefore,
2046        pcre_malloc etc. are now initialized to local functions that call the
2047        relevant global functions.
2048    
2049    7.  There were two entries missing in the vectors called coptable and poptable
2050        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2051        I've fixed the data, and added a kludgy way of testing at compile time that
2052        the lengths are correct (equal to the number of opcodes).
2053    
2054    8.  Following on from 7, I added a similar kludge to check the length of the
2055        eint vector in pcreposix.c.
2056    
2057    9.  Error texts for pcre_compile() are held as one long string to avoid too
2058        much relocation at load time. To find a text, the string is searched,
2059        counting zeros. There was no check for running off the end of the string,
2060        which could happen if a new error number was added without updating the
2061        string.
2062    
2063    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2064    
2065    11. \K was not working if it appeared in an atomic group or in a group that
2066        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2067        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2068        the assertion is positive, but not if it is negative.
2069    
2070    12. Change 11 fortuitously reduced the size of the stack frame used in the
2071        "match()" function of pcre_exec.c by one pointer. Forthcoming
2072        implementation of support for (*MARK) will need an extra pointer on the
2073        stack; I have reserved it now, so that the stack frame size does not
2074        decrease.
2075    
2076    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2077        item in branch that calls a recursion is a subroutine call - as in the
2078        second branch in the above example - was incorrectly given the compile-
2079        time error "recursive call could loop indefinitely" because pcre_compile()
2080        was not correctly checking the subroutine for matching a non-empty string.
2081    
2082    14. The checks for overrunning compiling workspace could trigger after an
2083        overrun had occurred. This is a "should never occur" error, but it can be
2084        triggered by pathological patterns such as hundreds of nested parentheses.
2085        The checks now trigger 100 bytes before the end of the workspace.
2086    
2087    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2088    
2089    
2090    Version 8.01 19-Jan-2010
2091    ------------------------
2092    
2093    1.  If a pattern contained a conditional subpattern with only one branch (in
2094        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2095        computed the wrong minimum data length (which is of course zero for such
2096        subpatterns). This could cause incorrect "no match" results.
2097    
2098    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2099        the pattern is reset in the first branch, pcre_compile() failed with
2100        "internal error: code overflow at offset...". This happened only when
2101        the reset was to the original external option setting. (An optimization
2102        abstracts leading options settings into an external setting, which was the
2103        cause of this.)
2104    
2105    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2106        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2107        assertion pattern did not match (meaning that the assertion was true), it
2108        was incorrectly treated as false if the SKIP had been reached during the
2109        matching. This also applied to assertions used as conditions.
2110    
2111    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2112        assertion subpattern, including such a pattern used as a condition,
2113        unpredictable results occurred, instead of the error return
2114        PCRE_ERROR_DFA_UITEM.
2115    
2116    5.  The C++ GlobalReplace function was not working like Perl for the special
2117        situation when an empty string is matched. It now does the fancy magic
2118        stuff that is necessary.
2119    
2120    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2121        removed. (These were left over from very, very early versions of PCRE.)
2122    
2123    7.  Some cosmetic changes to the code to make life easier when compiling it
2124        as part of something else:
2125    
2126        (a) Change DEBUG to PCRE_DEBUG.
2127    
2128        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2129            called "current" as "current_branch", to prevent a collision with the
2130            Linux macro when compiled as a kernel module.
2131    
2132        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2133            prevent a collision with the Linux macro when compiled as a kernel
2134            module.
2135    
2136    8.  In pcre_compile() there are some checks for integer overflows that used to
2137        cast potentially large values to (double). This has been changed to that
2138        when building, a check for int64_t is made, and if it is found, it is used
2139        instead, thus avoiding the use of floating point arithmetic. (There is no
2140        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2141        double.
2142    
2143    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2144        2005 (difference between two addresses compared to an unsigned value).
2145    
2146    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2147        custom one, because of the following reported problem in Windows:
2148    
2149          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2150              under Win32.
2151          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2152              therefore missing the function definition.
2153          - The compiler thus generates a "C" signature for the test function.
2154          - The linker fails to find the "C" function.
2155          - PCRE fails to configure if asked to do so against libbz2.
2156    
2157    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2158        messages were output:
2159    
2160          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2161          rerunning libtoolize, to keep the correct libtool macros in-tree.
2162          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2163    
2164        I have done both of these things.
2165    
2166    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2167        most of the time, it *can* run out if it is given a pattern that contains a
2168        runaway infinite recursion. I updated the discussion in the pcrestack man
2169        page.
2170    
2171    13. Now that we have gone to the x.xx style of version numbers, the minor
2172        version may start with zero. Using 08 or 09 is a bad idea because users
2173        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2174        interpreted as invalid octal numbers. I've updated the previous comment in
2175        configure.ac, and also added a check that gives an error if 08 or 09 are
2176        used.
2177    
2178    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2179        causing partial matching to fail when the end of the subject matched \W
2180        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2181    
2182    15. There were some discrepancies between the declarations in pcre_internal.h
2183        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2184        their definitions. The declarations used "const uschar *" and the
2185        definitions used USPTR. Even though USPTR is normally defined as "const
2186        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2187        reported that: "This difference in casting confuses some C++ compilers, for
2188        example, SunCC recognizes above declarations as different functions and
2189        generates broken code for hbpcre." I have changed the declarations to use
2190        USPTR.
2191    
2192    16. GNU libtool is named differently on some systems. The autogen.sh script now
2193        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2194        (FreeBSD).
2195    
2196    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2197        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2198        comment: "Figure out how to create a longlong from a string: strtoll and
2199        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2200        instance, but it only takes 2 args instead of 3!"
2201    
2202    18. A subtle bug concerned with back references has been fixed by a change of
2203        specification, with a corresponding code fix. A pattern such as
2204        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2205        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2206        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2207        same bug. Such groups have to be quantified to be useful, or contained
2208        inside another quantified group. (If there's no repetition, the reference
2209        can never match.) The problem arises because, having left the group and
2210        moved on to the rest of the pattern, a later failure that backtracks into
2211        the group uses the captured value from the final iteration of the group
2212        rather than the correct earlier one. I have fixed this in PCRE by forcing
2213        any group that contains a reference to itself to be an atomic group; that
2214        is, there cannot be any backtracking into it once it has completed. This is
2215        similar to recursive and subroutine calls.
2216    
2217    
2218    Version 8.00 19-Oct-09
2219    ----------------------
2220    
2221    1.  The table for translating pcre_compile() error codes into POSIX error codes
2222        was out-of-date, and there was no check on the pcre_compile() error code
2223        being within the table. This could lead to an OK return being given in
2224        error.
2225    
2226    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2227        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2228        in a Windows environment.
2229    
2230    3.  The pcregrep --count option prints the count for each file even when it is
2231        zero, as does GNU grep. However, pcregrep was also printing all files when
2232        --files-with-matches was added. Now, when both options are given, it prints
2233        counts only for those files that have at least one match. (GNU grep just
2234        prints the file name in this circumstance, but including the count seems
2235        more useful - otherwise, why use --count?) Also ensured that the
2236        combination -clh just lists non-zero counts, with no names.
2237    
2238    4.  The long form of the pcregrep -F option was incorrectly implemented as
2239        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2240        but it seems right to fix it, and I didn't think it was worth preserving
2241        the old behaviour.
2242    
2243    5.  The command line items --regex=pattern and --regexp=pattern were not
2244        recognized by pcregrep, which required --regex pattern or --regexp pattern
2245        (with a space rather than an '='). The man page documented the '=' forms,
2246        which are compatible with GNU grep; these now work.
2247    
2248    6.  No libpcreposix.pc file was created for pkg-config; there was just
2249        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2250    
2251    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2252        when UCP support is not needed, by modifying the Python script that
2253        generates it from Unicode data files. This should not matter if the module
2254        is correctly used as a library, but I received one complaint about 50K of
2255        unwanted data. My guess is that the person linked everything into his
2256        program rather than using a library. Anyway, it does no harm.
2257    
2258    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2259        was a minimum greater than 1 for a wide character in a possessive
2260        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2261        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2262        character. Chaos in the form of incorrect output or a compiling loop could
2263        result.
2264    
2265    9.  The restrictions on what a pattern can contain when partial matching is
2266        requested for pcre_exec() have been removed. All patterns can now be
2267        partially matched by this function. In addition, if there are at least two
2268        slots in the offset vector, the offset of the earliest inspected character
2269        for the match and the offset of the end of the subject are set in them when
2270        PCRE_ERROR_PARTIAL is returned.
2271    
2272    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2273        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2274        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2275        and may be more useful for multi-segment matching.
2276    
2277    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2278        used to be given if ever the end of the subject was reached; now it is
2279        given only if matching could not proceed because another character was
2280        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2281        string "Z", which now yields "no match" instead of "partial match". In the
2282        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2283        final character ended with (*FAIL).
2284    
2285    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2286        if the pattern had a "must contain" character that was already found in the
2287        earlier partial match, unless partial matching was again requested. For
2288        example, with the pattern /dog.(body)?/, the "must contain" character is
2289        "g". If the first part-match was for the string "dog", restarting with
2290        "sbody" failed. This bug has been fixed.
2291    
2292    13. The string returned by pcre_dfa_exec() after a partial match has been
2293        changed so that it starts at the first inspected character rather than the
2294        first character of the match. This makes a difference only if the pattern
2295        starts with a lookbehind assertion or \b or \B (\K is not supported by
2296        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2297        matching functions compatible, and I think it's the right thing to do.
2298    
2299    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2300        so that the demonstration program is easily available in environments where
2301        PCRE has not been installed from source.
2302    
2303    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2304        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2305        library.
2306    
2307    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2308        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2309        is not the first non-POSIX option to be added. Clearly some people find
2310        these options useful.
2311    
2312    17. If a caller to the POSIX matching function regexec() passes a non-zero
2313        value for nmatch with a NULL value for pmatch, the value of
2314        nmatch is forced to zero.
2315    
2316    18. RunGrepTest did not have a test for the availability of the -u option of
2317        the diff command, as RunTest does. It now checks in the same way as
2318        RunTest, and also checks for the -b option.
2319    
2320    19. If an odd number of negated classes containing just a single character
2321        interposed, within parentheses, between a forward reference to a named
2322        subpattern and the definition of the subpattern, compilation crashed with
2323        an internal error, complaining that it could not find the referenced
2324        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2325        [The bug was that it was starting one character too far in when skipping
2326        over the character class, thus treating the ] as data rather than
2327        terminating the class. This meant it could skip too much.]
2328    
2329    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2330        /g option in pcretest when the pattern contains \K, which makes it possible
2331        to have an empty string match not at the start, even when the pattern is
2332        anchored. Updated pcretest and pcredemo to use this option.
2333    
2334    21. If the maximum number of capturing subpatterns in a recursion was greater
2335        than the maximum at the outer level, the higher number was returned, but
2336        with unset values at the outer level. The correct (outer level) value is
2337        now given.
2338    
2339    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2340        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2341        make it do so. The string so far is captured, making this feature
2342        compatible with Perl.
2343    
2344    23. The tests have been re-organized, adding tests 11 and 12, to make it
2345        possible to check the Perl 5.10 features against Perl 5.10.
2346    
2347    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2348        pattern matches a fixed length string. PCRE did not allow this; now it
2349        does. Neither allows recursion.
2350    
2351    25. I finally figured out how to implement a request to provide the minimum
2352        length of subject string that was needed in order to match a given pattern.
2353        (It was back references and recursion that I had previously got hung up
2354        on.) This code has now been added to pcre_study(); it finds a lower bound
2355        to the length of subject needed. It is not necessarily the greatest lower
2356        bound, but using it to avoid searching strings that are too short does give
2357        some useful speed-ups. The value is available to calling programs via
2358        pcre_fullinfo().
2359    
2360    26. While implementing 25, I discovered to my embarrassment that pcretest had
2361        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2362        study optimizations had never been tested with that matching function.
2363        Oops. What is worse, even when it was passed study data, there was a bug in
2364        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2365        were also very few tests of studied patterns with pcre_dfa_exec().
2366    
2367    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2368        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2369        on the other side of the coin, they are no longer allowed to have different
2370        names, because these cannot be distinguished in PCRE, and this has caused
2371        confusion. (This is a difference from Perl.)
2372    
2373    28. When duplicate subpattern names are present (necessarily with different
2374        numbers, as required by 27 above), and a test is made by name in a
2375        conditional pattern, either for a subpattern having been matched, or for
2376        recursion in such a pattern, all the associated numbered subpatterns are
2377        tested, and the overall condition is true if the condition is true for any
2378        one of them. This is the way Perl works, and is also more like the way
2379        testing by number works.
2380    
2381    
2382    Version 7.9 11-Apr-09
2383    ---------------------
2384    
2385    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2386        (pcretest), all targets were linked against these libraries. This included
2387        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2388        libraries. This caused unwanted dependencies to be created. This problem
2389        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2390        pcretest is linked with readline.
2391    
2392    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2393        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2394        moved outside it again, because FALSE and TRUE are already defined in AIX,
2395        but BOOL is not.
2396    
2397    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2398        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2399    
2400    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2401        hyphens) following file names and line numbers when outputting matching
2402        lines. This is not true; no spaces are inserted. I have also clarified the
2403        wording for the --colour (or --color) option.
2404    
2405    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2406        was not coloured; this is different to GNU grep, so I have changed it to be
2407        the same.
2408    
2409    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2410        each matching line was coloured. Now it goes on to look for further matches
2411        of any of the test patterns, which is the same behaviour as GNU grep.
2412    
2413    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2414        doesn't make sense to accept an empty string match in pcregrep, so I have
2415        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2416        seems to be how GNU grep behaves. [But see later change 40 for release
2417        8.33.]
2418    
2419    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2420        start or after a newline", because the conditional assertion was not being
2421        correctly handled. The rule now is that both the assertion and what follows
2422        in the first alternative must satisfy the test.
2423    
2424    9.  If auto-callout was enabled in a pattern with a conditional group whose
2425        condition was an assertion, PCRE could crash during matching, both with
2426        pcre_exec() and pcre_dfa_exec().
2427    
2428    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2429        used for matching.
2430    
2431    11. Unicode property support in character classes was not working for
2432        characters (bytes) greater than 127 when not in UTF-8 mode.
2433    
2434    12. Added the -M command line option to pcretest.
2435    
2436    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2437    
2438    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2439    
2440    16. Added comments and documentation about mis-use of no_arg in the C++
2441        wrapper.
2442    
2443    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2444        from Martin Jerabek that uses macro names for all relevant character and
2445        string constants.
2446    
2447    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2448        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2449        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2450        these, but not everybody uses configure.
2451    
2452    19. A conditional group that had only one branch was not being correctly
2453        recognized as an item that could match an empty string. This meant that an
2454        enclosing group might also not be so recognized, causing infinite looping
2455        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2456        with the subject "ab", where knowledge that the repeated group can match
2457        nothing is needed in order to break the loop.
2458    
2459    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2460        exec(), but without supplying a callout function, matching went wrong.
2461    
2462    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2463        leak if the size of the offset vector was greater than 30. When the vector
2464        is smaller, the saved offsets during recursion go onto a local stack
2465        vector, but for larger vectors malloc() is used. It was failing to free
2466        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2467        error, in fact).
2468    
2469    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2470        heapframe that is used only when UTF-8 support is enabled. This caused no
2471        problem, but was untidy.
2472    
2473    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2474        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2475        included within another project.
2476    
2477    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2478        slightly modified by me:
2479    
2480          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2481              not building pcregrep.
2482    
2483          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2484              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2485    
2486    25. Forward references, both numeric and by name, in patterns that made use of
2487        duplicate group numbers, could behave incorrectly or give incorrect errors,
2488        because when scanning forward to find the reference group, PCRE was not
2489        taking into account the duplicate group numbers. A pattern such as
2490        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2491    
2492    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2493        the feature of a custom pointer more persuasive (as requested by a user).
2494    
2495    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2496        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2497        pre-defined.
2498    
2499    28. Added support for (*UTF8) at the start of a pattern.
2500    
2501    29. Arrange for flags added by the "release type" setting in CMake to be shown
2502        in the configuration summary.
2503    
2504    
2505    Version 7.8 05-Sep-08
2506    ---------------------
2507    
2508    1.  Replaced UCP searching code with optimized version as implemented for Ad
2509        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2510        stage table and inline lookup instead of a function, giving speed ups of 2
2511        to 5 times on some simple patterns that I tested. Permission was given to
2512        distribute the MultiStage2.py script that generates the tables (it's not in
2513        the tarball, but is in the Subversion repository).
2514    
2515    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2516        scripts.
2517    
2518    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2519        a group with a zero qualifier. The result of the study could be incorrect,
2520        or the function might crash, depending on the pattern.
2521    
2522    4.  Caseless matching was not working for non-ASCII characters in back
2523        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2524        It now works when Unicode Property Support is available.
2525    
2526    5.  In pcretest, an escape such as \x{de} in the data was always generating
2527        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2528        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2529        truncation.
2530    
2531    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2532    
2533    7.  Added two (int) casts to pcregrep when printing the difference of two
2534        pointers, in case they are 64-bit values.
2535    
2536    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2537        test 2 if it fails.
2538    
2539    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2540        and a #define of that name to empty if it is not externally set. This is to
2541        allow users of MSVC to set it if necessary.
2542    
2543    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2544        the convenience functions in the pcre_get.c source file.
2545    
2546    11. An option change at the start of a pattern that had top-level alternatives
2547        could cause overwriting and/or a crash. This command provoked a crash in
2548        some environments:
2549    
2550          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2551    
2552        This potential security problem was recorded as CVE-2008-2371.
2553    
2554    12. For a pattern where the match had to start at the beginning or immediately
2555        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2556        pcre_dfa_exec() could read past the end of the passed subject if there was
2557        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2558        pcretest so that it places the subject at the end of its malloc-ed buffer.
2559    
2560    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2561        exec() might read past the end of the data buffer in UTF-8 mode.
2562    
2563    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2564        the data contained the byte 0x85 as part of a UTF-8 character within its
2565        first line. This applied both to normal and DFA matching.
2566    
2567    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2568        /^[^d]*?$/8 failed to match "abc".
2569    
2570    16. Added a missing copyright notice to pcrecpp_internal.h.
2571    
2572    17. Make it more clear in the documentation that values returned from
2573        pcre_exec() in ovector are byte offsets, not character counts.
2574    
2575    18. Tidied a few places to stop certain compilers from issuing warnings.
2576    
2577    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2578        supplied by Stefan Weber. I made a further small update for 7.8 because
2579        there is a change of source arrangements: the pcre_searchfuncs.c module is
2580        replaced by pcre_ucd.c.
2581    
2582    
2583    Version 7.7 07-May-08
2584    ---------------------
2585    
2586    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2587        a string to a long long, pretend we don't even have a long long." This is
2588        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2589    
2590    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2591        pre-7.6 versions, which defined a global no_arg variable instead of putting
2592        it in the RE class. (See also #8 below.)
2593    
2594    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2595        Lopes.
2596    
2597    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2598    
2599        (1) The include/exclude patterns were being applied to the whole pathnames
2600            of files, instead of just to the final components.
2601    
2602        (2) If there was more than one level of directory, the subdirectories were
2603            skipped unless they satisfied the include/exclude conditions. This is
2604            inconsistent with GNU grep (and could even be seen as contrary to the
2605            pcregrep specification - which I improved to make it absolutely clear).
2606            The action now is always to scan all levels of directory, and just
2607            apply the include/exclude patterns to regular files.
2608    
2609    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2610        --exclude_dir in the tests to avoid scanning .svn directories.
2611    
2612    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2613        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2614        doesn't support NULs in patterns.
2615    
2616    7.  Added some missing "const"s to declarations of static tables in
2617        pcre_compile.c and pcre_dfa_exec.c.
2618    
2619    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2620        caused by fix #2  above. (Subsequently also a second patch to fix the
2621        first patch. And a third patch - this was a messy problem.)
2622    
2623    9.  Applied Craig's patch to remove the use of push_back().
2624    
2625    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2626        matching function regexec().
2627    
2628    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2629        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2630        references. PCRE supports relative numbers with this syntax (I don't think
2631        Oniguruma does).
2632    
2633    12. Previously, a group with a zero repeat such as (...){0} was completely
2634        omitted from the compiled regex. However, this means that if the group
2635        was called as a subroutine from elsewhere in the pattern, things went wrong
2636        (an internal error was given). Such groups are now left in the compiled
2637        pattern, with a new opcode that causes them to be skipped at execution
2638        time.
2639    
2640    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2641        to the way PCRE behaves:
2642    
2643        (a) A lone ] character is dis-allowed (Perl treats it as data).
2644    
2645        (b) A back reference to an unmatched subpattern matches an empty string
2646            (Perl fails the current match path).
2647    
2648        (c) A data ] in a character class must be notated as \] because if the
2649            first data character in a class is ], it defines an empty class. (In
2650            Perl it is not possible to have an empty class.) The empty class []
2651            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2652            The negative empty class [^] matches any one character, independently
2653            of the DOTALL setting.
2654    
2655    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2656        non-existent subpattern following a character class starting with ']' and
2657        containing () gave an internal compiling error instead of "reference to
2658        non-existent subpattern". Fortunately, when the pattern did exist, the
2659        compiled code was correct. (When scanning forwards to check for the
2660        existence of the subpattern, it was treating the data ']' as terminating
2661        the class, so got the count wrong. When actually compiling, the reference
2662        was subsequently set up correctly.)
2663    
2664    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2665        it was being rejected as not supported by pcre_dfa_exec(), even though
2666        other assertions are supported. I have made pcre_dfa_exec() support
2667        (*FAIL).
2668    
2669    16. The implementation of 13c above involved the invention of a new opcode,
2670        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2671        cannot be changed at match time, I realized I could make a small
2672        improvement to matching performance by compiling OP_ALLANY instead of
2673        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2674        on the OP_ANY path.
2675    
2676    17. Compiling pcretest on Windows with readline support failed without the
2677        following two fixes: (1) Make the unistd.h include conditional on
2678        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2679    
2680    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2681        ncurses library to be included for pcretest when ReadLine support is
2682        requested, but also to allow for it to be overridden. This patch came from
2683        Daniel Bergström.
2684    
2685    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2686        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2687        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2688        spotting this.
2689    
2690    
2691    Version 7.6 28-Jan-08
2692    ---------------------
2693    
2694    1.  A character class containing a very large number of characters with
2695        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2696        overflow.
2697    
2698    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2699        HAVE_LONG_LONG is not defined.
2700    
2701    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2702        bring them up to date and include new features. This patch includes:
2703    
2704        - Fixed PH's badly added libz and libbz2 support.
2705        - Fixed a problem with static linking.
2706        - Added pcredemo. [But later removed - see 7 below.]
2707        - Fixed dftables problem and added an option.
2708        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2709            HAVE_LONG_LONG.
2710        - Added readline support for pcretest.
2711        - Added an listing of the option settings after cmake has run.
2712    
2713    4.  A user submitted a patch to Makefile that makes it easy to create
2714        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2715        Makefile.am that cause it to include this special target, without
2716        affecting anything else. Note that the same mingw target plus all
2717        the other distribution libraries and programs are now supported
2718        when configuring with CMake (see 6 below) instead of with
2719        Configure/Make.
2720    
2721    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2722        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2723        exported in the Windows port". It has not yet been confirmed that the patch
2724        solves the problem, but it does no harm.
2725    
2726    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2727        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2728        with CMake, and also correct the comment about stack recursion.
2729    
2730    7.  Remove the automatic building of pcredemo from the ./configure system and
2731        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2732        of a program that users should build themselves after PCRE is installed, so
2733        building it automatically is not really right. What is more, it gave
2734        trouble in some build environments.
2735    
2736    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2737    
2738    
2739    Version 7.5 10-Jan-08
2740    ---------------------
2741    
2742    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2743        values in parens when parsing an RE using the C++ wrapper."
2744    
2745    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2746        Characters greater than 255 were excluded from the class instead of being
2747        included.
2748    
2749    3.  The same bug as (2) above applied to negated POSIX classes such as
2750        [:^space:].
2751    
2752    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2753        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2754        I have changed it.
2755    
2756    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2757        first named subpattern) and a construct such as (?&a) would reference the
2758        first named subpattern whose name started with "a" (in other words, the
2759        length check was missing). Both these problems are fixed. "Subpattern name
2760        expected" is now given for (?&) (a zero-length name), and this patch also
2761        makes it give the same error for \k'' (previously it complained that that
2762        was a reference to a non-existent subpattern).
2763    
2764    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2765        this is right because (?- can be followed by option settings as well as by
2766        digits. I have, however, made the messages clearer.
2767    
2768    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2769        than the number used in the conditional) now cause a compile-time error.
2770        This is actually not compatible with Perl, which accepts such patterns, but
2771        treats the conditional as always being FALSE (as PCRE used to), but it
2772        seems to me that giving a diagnostic is better.
2773    
2774    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2775        and messages.
2776    
2777    9.  Fix two occurrences of "backslash" in comments that should have been
2778        "backspace".
2779    
2780    10. Remove two redundant lines of code that can never be obeyed (their function
2781        was moved elsewhere).
2782    
2783    11. The program that makes PCRE's Unicode character property table had a bug
2784        which caused it to generate incorrect table entries for sequences of
2785        characters that have the same character type, but are in different scripts.
2786        It amalgamated them into a single range, with the script of the first of
2787        them. In other words, some characters were in the wrong script. There were
2788        thirteen such cases, affecting characters in the following ranges:
2789    
2790          U+002b0 - U+002c1
2791          U+0060c - U+0060d
2792          U+0061e - U+00612
2793          U+0064b - U+0065e
2794          U+0074d - U+0076d
2795          U+01800 - U+01805
2796          U+01d00 - U+01d77
2797          U+01d9b - U+01dbf
2798          U+0200b - U+0200f
2799          U+030fc - U+030fe
2800          U+03260 - U+0327f
2801          U+0fb46 - U+0fbb1
2802          U+10450 - U+1049d
2803    
2804    12. The -o option (show only the matching part of a line) for pcregrep was not
2805        compatible with GNU grep in that, if there was more than one match in a
2806        line, it showed only the first of them. It now behaves in the same way as
2807        GNU grep.
2808    
2809    13. If the -o and -v options were combined for pcregrep, it printed a blank
2810        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2811        does the same. The return code can be used to tell if there were any
2812        non-matching lines.
2813    
2814    14. Added --file-offsets and --line-offsets to pcregrep.
2815    
2816    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2817        infinitely looping recursion. The bug was that positive lookaheads were not
2818        being skipped when checking for a possible empty match (negative lookaheads
2819        and both kinds of lookbehind were skipped).
2820    
2821    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2822        inclusion of <windows.h> to before rather than after the definition of
2823        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2824    
2825    17. Specifying a possessive quantifier with a specific limit for a Unicode
2826        character property caused pcre_compile() to compile bad code, which led at
2827        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2828        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2829        caused the error; without that there was no problem.
2830    
2831    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2832    
2833    19. Added --enable-pcretest-libreadline.
2834    
2835    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2836        RE::GlobalReplace(). As a result, the number of replacements returned was
2837        double what it should be. I removed one of the increments, but Craig sent a
2838        later patch that removed the other one (the right fix) and added unit tests
2839        that check the return values (which was not done before).
2840    
2841    21. Several CMake things:
2842    
2843        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2844            the names libpcre and libpcreposix, not just pcre and pcreposix.
2845    
2846        (2) The above change means that pcretest and pcregrep are now correctly
2847            linked with the newly-built libraries, not previously installed ones.
2848    
2849        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2850    
2851    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2852        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2853        UTF-8 newline character). The key issue is that the pattern starts .*;
2854        this means that the match must be either at the beginning, or after a
2855        newline. The bug was in the code for advancing after a failed match and
2856        checking that the new position followed a newline. It was not taking
2857        account of UTF-8 characters correctly.
2858    
2859    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2860        character classes. PCRE was not treating the sequence [:...:] as a
2861        character class unless the ... were all letters. Perl, however, seems to
2862        allow any characters between [: and :], though of course it rejects as
2863        unknown any "names" that contain non-letters, because all the known class
2864        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2865        for example, whereas PCRE did not - it did not recognize a POSIX character
2866        class. This seemed a bit dangerous, so the code has been changed to be
2867        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2868        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2869        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2870        Perl does, and where it didn't before.
2871    
2872    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2873        Windows environments %n is disabled by default.
2874    
2875    
2876    Version 7.4 21-Sep-07
2877    ---------------------
2878    
2879    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2880        means that a class such as [\s] counted as "explicit reference to CR or
2881        LF". That isn't really right - the whole point of the change was to try to
2882        help when there was an actual mention of one of the two characters. So now
2883        the change happens only if \r or \n (or a literal CR or LF) character is
2884        encountered.
2885    
2886    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2887        of both had grown to the point where there were only 3 bits left.
2888        Fortunately, there was spare space in the data structure, and so I have
2889        moved the internal flags into a new 16-bit field to free up more option
2890        bits.
2891    
2892    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2893        but did not set the internal JCHANGED flag - either of these is enough to
2894        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2895        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2896        start sets both bits.
2897    
2898    4.  Added options (at build time, compile time, exec time) to change \R from
2899        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2900    
2901    5.  doc/pcresyntax.html was missing from the distribution.
2902    
2903    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2904        compatibility, even though it is no longer used.
2905    
2906    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2907        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2908        windows.h file is present (where different names are used). [This was
2909        reversed later after testing - see 16 below.]
2910    
2911    8.  Changed all #include <config.h> to #include "config.h". There were also
2912        some further <pcre.h> cases that I changed to "pcre.h".
2913    
2914    9.  When pcregrep was used with the --colour option, it missed the line ending
2915        sequence off the lines that it output.
2916    
2917    10. It was pointed out to me that arrays of string pointers cause lots of
2918        relocations when a shared library is dynamically loaded. A technique of
2919        using a single long string with a table of offsets can drastically reduce
2920        these. I have refactored PCRE in four places to do this. The result is
2921        dramatic:
2922    
2923          Originally:                          290
2924          After changing UCP table:            187
2925          After changing error message table:   43
2926          After changing table of "verbs"       36
2927          After changing table of Posix names   22
2928    
2929        Thanks to the folks working on Gregex for glib for this insight.
2930    
2931    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2932        unicode-properties was also set.
2933    
2934    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2935    
2936    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2937        checked only for CRLF.
2938    
2939    14. Added casts to pcretest.c to avoid compiler warnings.
2940    
2941    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2942    
2943    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2944        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2945        entirely. This removes changes made in 7 above.
2946    
2947    17. The CMake files have been updated, and there is now more information about
2948        building with CMake in the NON-UNIX-USE document.
2949    
2950    
2951    Version 7.3 28-Aug-07
2952    ---------------------
2953    
2954     1. In the rejigging of the build system that eventually resulted in 7.1, the
2955        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2956        brackets there is not right, since it causes compilers to look for an
2957        installed pcre.h, not the version that is in the source that is being
2958        compiled (which of course may be different). I have changed it back to:
2959    
2960          #include "pcre.h"
2961    
2962        I have a vague recollection that the change was concerned with compiling in
2963        different directories, but in the new build system, that is taken care of
2964        by the VPATH setting the Makefile.
2965    
2966     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2967        when the subject happened to end in the byte 0x85 (e.g. if the last
2968        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2969        characters but of course it shouldn't be taken as a newline when it is part
2970        of another character. The bug was that, for an unlimited repeat of . in
2971        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2972        characters when looking for a newline.
2973    
2974     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2975    
2976     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2977        in debug output.
2978    
2979     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2980        long printing in the pcrecpp unittest when running under MinGW.
2981    
2982     6. ESC_K was left out of the EBCDIC table.
2983    
2984     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2985        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2986        limit also applies to "virtual nesting" when a pattern is recursive, and in
2987        this case 1000 isn't so big. I have been able to remove this limit at the
2988        expense of backing off one optimization in certain circumstances. Normally,
2989        when pcre_exec() would call its internal match() function recursively and
2990        immediately return the result unconditionally, it uses a "tail recursion"
2991        feature to save stack. However, when a subpattern that can match an empty
2992        string has an unlimited repetition quantifier, it no longer makes this
2993        optimization. That gives it a stack frame in which to save the data for
2994        checking that an empty string has been matched. Previously this was taken
2995        from the 1000-entry workspace that had been reserved. So now there is no
2996        explicit limit, but more stack is used.
2997    
2998     8. Applied Daniel's patches to solve problems with the import/export magic
2999        syntax that is required for Windows, and which was going wrong for the
3000        pcreposix and pcrecpp parts of the library. These were overlooked when this
3001        problem was solved for the main library.
3002    
3003     9. There were some crude static tests to avoid integer overflow when computing
3004        the size of patterns that contain repeated groups with explicit upper
3005        limits. As the maximum quantifier is 65535, the maximum group length was
3006        set at 30,000 so that the product of these two numbers did not overflow a
3007        32-bit integer. However, it turns out that people want to use groups that
3008        are longer than 30,000 bytes (though not repeat them that many times).
3009        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3010        made it possible to implement the integer overflow checks in a much more
3011        dynamic way, which I have now done. The artificial limitation on group
3012        length has been removed - we now have only the limit on the total length of
3013        the compiled pattern, which depends on the LINK_SIZE setting.
3014    
3015    10. Fixed a bug in the documentation for get/copy named substring when
3016        duplicate names are permitted. If none of the named substrings are set, the
3017        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3018        empty string.
3019    
3020    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3021        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3022        because the ] is interpreted as the first data character and the
3023        terminating ] is not found. PCRE has been made compatible with Perl in this
3024        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3025        cause memory overwriting.
3026    
3027    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3028        string has been matched (to stop an infinite loop). It was not recognizing
3029        a conditional subpattern that could match an empty string if that
3030        subpattern was within another subpattern. For example, it looped when
3031        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3032        condition was not nested. This bug has been fixed.
3033    
3034    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3035        past the start of the subject in the presence of bytes with the top bit
3036        set, for example "\x8aBCD".
3037    
3038    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3039        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3040    
3041    14. Optimized (?!) to (*FAIL).
3042    
3043    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3044        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3045        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3046        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3047        does: it's just the validity check that is more restrictive.
3048    
3049    16. Inserted checks for integer overflows during escape sequence (backslash)
3050        processing, and also fixed erroneous offset values for syntax errors during
3051        backslash processing.
3052    
3053    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3054        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3055    
3056    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3057        caused an overrun.
3058    
3059    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3060        something other than just ASCII characters) inside a group that had an
3061        unlimited repeat caused a loop at compile time (while checking to see
3062        whether the group could match an empty string).
3063    
3064    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3065        [\P{Any}] did so. (Error in the code for printing property names.)
3066    
3067    21. An orphan \E inside a character class could cause a crash.
3068    
3069    22. A repeated capturing bracket such as (A)? could cause a wild memory
3070        reference during compilation.
3071    
3072    23. There are several functions in pcre_compile() that scan along a compiled
3073        expression for various reasons (e.g. to see if it's fixed length for look
3074        behind). There were bugs in these functions when a repeated \p or \P was
3075        present in the pattern. These operators have additional parameters compared
3076        with \d, etc, and these were not being taken into account when moving along
3077        the compiled data. Specifically:
3078    
3079        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3080            length.
3081    
3082        (b) An item such as \pL+ within a repeated group could cause crashes or
3083            loops.
3084    
3085        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3086            "reference to non-existent subpattern" error.
3087    
3088        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3089    
3090    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3091        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3092    
3093    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3094    
3095    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3096        character were causing crashes (broken optimization).
3097    
3098    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3099        \p or \P) caused a compile-time loop.
3100    
3101    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3102        break. For example, the unstudied pattern [\r\n]A does not match the string
3103        "\r\nA" because change 7.0/46 below moves the current point on by two
3104        characters after failing to match at the start. However, the pattern \nA
3105        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3106        the same is true. There doesn't seem any very clean way out of this, but
3107        what I have chosen to do makes the common cases work: PCRE now takes note
3108        of whether there can be an explicit match for \r or \n anywhere in the
3109        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3110        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3111        pattern has explicit CR or LF references.
3112    
3113    29. Added (*CR) etc for changing newline setting at start of pattern.
3114    
3115    
3116    Version 7.2 19-Jun-07
3117    ---------------------
3118    
3119     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3120        which is apparently normally available under Windows.
3121    
3122     2. Re-jig the pcregrep tests with different newline settings in an attempt
3123        to make them independent of the local environment's newline setting.
3124    
3125     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3126    
3127     4. Some of the "internals" tests were previously cut out when the link size
3128        was not 2, because the output contained actual offsets. The recent new
3129        "Z" feature of pcretest means that these can be cut out, making the tests
3130        usable with all link sizes.
3131    
3132     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3133        stack recursion. This gives a massive performance boost under BSD, but just
3134        a small improvement under Linux. However, it saves one field in the frame
3135        in all cases.
3136    
3137     6. Added more features from the forthcoming Perl 5.10:
3138    
3139        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3140            recursion call. It refers to the nth most recently opened parentheses.
3141    
3142        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3143            to be opened parentheses.
3144    
3145        (c) Conditions that refer to capturing parentheses can be specified
3146            relatively, for example, (?(-2)... or (?(+3)...
3147    
3148        (d) \K resets the start of the current match so that everything before
3149            is not part of it.
3150    
3151        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3152    
3153        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3154            reference syntax.
3155    
3156        (g) (?| introduces a group in which the numbering of parentheses in each
3157            alternative starts with the same number.
3158    
3159        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3160    
3161     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3162        PCRE_INFO_JCHANGED.
3163    
3164     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3165        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3166        for detecting groups that can match an empty string.
3167    
3168     9. A pattern with a very large number of alternatives (more than several
3169        hundred) was running out of internal workspace during the pre-compile
3170        phase, where pcre_compile() figures out how much memory will be needed. A
3171        bit of new cunning has reduced the workspace needed for groups with
3172        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3173        workspace instead of running out of the 4096 that are available.
3174    
3175    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3176    
3177    11. Applied patch from Google to remove an optimization that didn't quite work.
3178        The report of the bug said:
3179    
3180          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3181          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3182          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3183    
3184    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3185        it matched the wrong number of bytes.
3186    
3187    
3188    Version 7.1 24-Apr-07
3189    ---------------------
3190    
3191     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3192        that is more "standard", making use of automake and other Autotools. There
3193        is some re-arrangement of the files and adjustment of comments consequent
3194        on this.
3195    
3196     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3197        for recursive directory scanning broke on some systems because the files
3198        are not scanned in any specific order and on different systems the order
3199        was different. A call to "sort" has been inserted into RunGrepTest for the
3200        approprate test as a short-term fix. In the longer term there may be an
3201        alternative.
3202    
3203     3. I had an email from Eric Raymond about problems translating some of PCRE's
3204        man pages to HTML (despite the fact that I distribute HTML pages, some
3205        people do their own conversions for various reasons). The problems
3206        concerned the use of low-level troff macros .br and .in. I have therefore
3207        removed all such uses from the man pages (some were redundant, some could
3208        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3209        HTML has been updated to handle .nf/.fi and to complain if it encounters
3210        .br or .in.
3211    
3212     4. Updated comments in configure.ac that get placed in config.h.in and also
3213        arranged for config.h to be included in the distribution, with the name
3214        config.h.generic, for the benefit of those who have to compile without
3215        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3216    
3217     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3218        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3219        makevp.bat for the current PCRE, using the additional files
3220        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3221    
3222     6. A Windows user reported a minor discrepancy with test 2, which turned out
3223        to be caused by a trailing space on an input line that had got lost in his
3224        copy. The trailing space was an accident, so I've just removed it.
3225    
3226     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3227        that is needed.
3228    
3229     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3230        as "const" (a) because they are and (b) because it helps the PHP
3231        maintainers who have recently made a script to detect big data structures
3232        in the php code that should be moved to the .rodata section. I remembered
3233        to update Builducptable as well, so it won't revert if ucptable.h is ever
3234        re-created.
3235    
3236     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3237        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3238        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3239        support is not required. This saves 1.5-2K of code, which is important in
3240        some applications.
3241    
3242        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3243        so as not to refer to the tables, even though these functions will never be
3244        called when UTF-8 support is disabled. Otherwise there are problems with a
3245        shared library.
3246    
3247    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3248    
3249        (a) It was defining its arguments as char * instead of void *.
3250    
3251        (b) It was assuming that all moves were upwards in memory; this was true
3252            a long time ago when I wrote it, but is no longer the case.
3253    
3254        The emulated memove() is provided for those environments that have neither
3255        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3256        is clearly not the case, as these two bugs were recently reported.
3257    
3258    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3259        and Detrail to create the HTML documentation, the .txt form of the man
3260        pages, and it removes trailing spaces from listed files. It also creates
3261        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3262        case, it wraps all the #defines with #ifndefs. This script should be run
3263        before "make dist".
3264    
3265    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3266        with Unicode property support.
3267    
3268        (a) For a maximizing quantifier, if the two different cases of the
3269            character were of different lengths in their UTF-8 codings (there are
3270            some cases like this - I found 11), and the matching function had to
3271            back up over a mixture of the two cases, it incorrectly assumed they
3272            were both the same length.
3273    
3274        (b) When PCRE was configured to use the heap rather than the stack for
3275            recursion during matching, it was not correctly preserving the data for
3276            the other case of a UTF-8 character when checking ahead for a match
3277            while processing a minimizing repeat. If the check also involved
3278            matching a wide character, but failed, corruption could cause an
3279            erroneous result when trying to check for a repeat of the original
3280            character.
3281    
3282    13. Some tidying changes to the testing mechanism:
3283    
3284        (a) The RunTest script now detects the internal link size and whether there
3285            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3286            values substituted by "configure". (The RunGrepTest script already did
3287            this for UTF-8.) The configure.ac script no longer substitutes the
3288            relevant variables.
3289    
3290        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3291            with length and offset values. This means that the output is different
3292            for different internal link sizes. Test 2 is skipped for link sizes
3293            other than 2 because of this, bypassing the problem. Unfortunately,
3294            there was also a test in test 3 (the locale tests) that used /B and
3295            failed for link sizes other than 2. Rather than cut the whole test out,
3296            I have added a new /Z option to pcretest that replaces the length and
3297            offset values with spaces. This is now used to make test 3 independent
3298            of link size. (Test 2 will be tidied up later.)
3299    
3300    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3301        segmentation fault instead of returning the appropriate error message.
3302    
3303    15. In multiline mode when the newline sequence was set to "any", the pattern
3304        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3305        This doesn't seem right; it now treats the CRLF combination as the line
3306        ending, and so does not match in that case. It's only a pattern such as ^$
3307        that would hit this one: something like ^ABC$ would have failed after \r
3308        and then tried again after \r\n.
3309    
3310    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3311        in an attempt to make files that differ only in their line terminators
3312        compare equal. This works on Linux.
3313    
3314    17. Under certain error circumstances pcregrep might try to free random memory
3315        as it exited. This is now fixed, thanks to valgrind.
3316    
3317    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3318        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3319        was because its rules for how to advance for /g after matching an empty
3320        string at the end of a line did not allow for this case. They now check for
3321        it specially.
3322    
3323    20. pcretest is supposed to handle patterns and data of any length, by
3324        extending its buffers when necessary. It was getting this wrong when the
3325        buffer for a data line had to be extended.
3326    
3327    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3328        CRLF as a newline sequence.
3329    
3330    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3331        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3332        I have nevertheless tidied it up.
3333    
3334    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3335    
3336    24. Added a man page for pcre-config.
3337    
3338    
3339    Version 7.0 19-Dec-06
3340    ---------------------
3341    
3342     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3343        moving to gcc 4.1.1.
3344    
3345     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3346        sys/time.h, which is documented as needed for this function. It doesn't
3347        seem to matter on Linux, but it showed up on some releases of OS X.
3348    
3349     3. It seems that there are systems where bytes whose values are greater than
3350        127 match isprint() in the "C" locale. The "C" locale should be the
3351        default when a C program starts up. In most systems, only ASCII printing
3352        characters match isprint(). This difference caused the output from pcretest
3353        to vary, making some of the tests fail. I have changed pcretest so that:
3354    
3355        (a) When it is outputting text in the compiled version of a pattern, bytes
3356            other than 32-126 are always shown as hex escapes.
3357    
3358        (b) When it is outputting text that is a matched part of a subject string,
3359            it does the same, unless a different locale has been set for the match
3360            (using the /L modifier). In this case, it uses isprint() to decide.
3361    
3362     4. Fixed a major bug that caused incorrect computation of the amount of memory
3363        required for a compiled pattern when options that changed within the
3364        pattern affected the logic of the preliminary scan that determines the
3365        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3366        that the computed length was too small. The symptoms of this bug were
3367        either the PCRE error "internal error: code overflow" from pcre_compile(),
3368        or a glibc crash with a message such as "pcretest: free(): invalid next
3369        size (fast)". Examples of patterns that provoked this bug (shown in
3370        pcretest format) are:
3371    
3372          /(?-x: )/x
3373          /(?x)(?-x: \s*#\s*)/
3374          /((?i)[\x{c0}])/8
3375          /(?i:[\x{c0}])/8
3376    
3377        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3378        is now done differently.
3379    
3380     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3381        wrapper classes; (b) implement a new function in the C++ scanner that is
3382        more efficient than the old way of doing things because it avoids levels of
3383        recursion in the regex matching; (c) add a paragraph to the documentation
3384        for the FullMatch() function.
3385    
3386     6. The escape sequence \n was being treated as whatever was defined as
3387        "newline". Not only was this contrary to the documentation, which states
3388        that \n is character 10 (hex 0A), but it also went horribly wrong when
3389        "newline" was defined as CRLF. This has been fixed.
3390    
3391     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3392        was being set to -1 for the "end of line" case (supposedly a value that no
3393        character can have). Though this value is never used (the check for end of
3394        line is "zero bytes in current character"), it caused compiler complaints.
3395        I've changed it to 0xffffffff.
3396    
3397     8. In pcre_version.c, the version string was being built by a sequence of
3398        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3399        string (as it is for production releases) called a macro with an empty
3400        argument. The C standard says the result of this is undefined. The gcc
3401        compiler treats it as an empty string (which was what was wanted) but it is
3402        reported that Visual C gives an error. The source has been hacked around to
3403        avoid this problem.
3404    
3405     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3406        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3407        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3408        of them did).
3409    
3410    10. Originally, pcretest opened its input and output without "b"; then I was
3411        told that "b" was needed in some environments, so it was added for release
3412        5.0 to both the input and output. (It makes no difference on Unix-like
3413        systems.) Later I was told that it is wrong for the input on Windows. I've
3414        now abstracted the modes into two macros, to make it easier to fiddle with
3415        them, and removed "b" from the input mode under Windows.
3416    
3417    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3418    
3419    12. Added -help and --help to pcretest as an official way of being reminded
3420        of the options.
3421    
3422    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3423        and pcrecpp.cc because they annoy compilers at high warning levels.
3424    
3425    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3426    
3427    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3428        scripts are not C programs :-) and which was not noticed because it works
3429        on Linux.
3430    
3431    16. pcretest is supposed to handle any length of pattern and data line (as one
3432        line or as a continued sequence of lines) by extending its input buffer if
3433        necessary. This feature was broken for very long pattern lines, leading to
3434        a string of junk being passed to pcre_compile() if the pattern was longer
3435        than about 50K.
3436    
3437    17. I have done a major re-factoring of the way pcre_compile() computes the
3438        amount of memory needed for a compiled pattern. Previously, there was code
3439        that made a preliminary scan of the pattern in order to do this. That was
3440        OK when PCRE was new, but as the facilities have expanded, it has become
3441        harder and harder to keep it in step with the real compile phase, and there
3442        have been a number of bugs (see for example, 4 above). I have now found a
3443        cunning way of running the real compile function in a "fake" mode that
3444        enables it to compute how much memory it would need, while actually only
3445        ever using a few hundred bytes of working memory and without too many
3446        tests of the mode. This should make future maintenance and development
3447        easier. A side effect of this work is that the limit of 200 on the nesting
3448        depth of parentheses has been removed (though this was never a serious
3449        limitation, I suspect). However, there is a downside: pcre_compile() now
3450        runs more slowly than before (30% or more, depending on the pattern). I
3451        hope this isn't a big issue. There is no effect on runtime performance.
3452    
3453    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3454        newline (only possible for the last line of a file) and it was a
3455        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3456    
3457    19. Added additional timing features to pcretest. (1) The -tm option now times
3458        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3459        separate command line item, by a number that specifies the number of
3460        repeats to use when timing. The default is 50000; this gives better
3461        precision, but takes uncomfortably long for very large patterns.
3462    
3463    20. Extended pcre_study() to be more clever in cases where a branch of a
3464        subpattern has no definite first character. For example, (a*|b*)[cd] would
3465        previously give no result from pcre_study(). Now it recognizes that the
3466        first character must be a, b, c, or d.
3467    
3468    21. There was an incorrect error "recursive call could loop indefinitely" if
3469        a subpattern (or the entire pattern) that was being tested for matching an
3470        empty string contained only one non-empty item after a nested subpattern.
3471        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3472        incorrectly, because the \d was being skipped in the check.
3473    
3474    22. The pcretest program now has a new pattern option /B and a command line
3475        option -b, which is equivalent to adding /B to every pattern. This causes
3476        it to show the compiled bytecode, without the additional information that
3477        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3478        is the same as /B/I).
3479    
3480    23. A new optimization is now able automatically to treat some sequences such
3481        as a*b as a*+b. More specifically, if something simple (such as a character
3482        or a simple class like \d) has an unlimited quantifier, and is followed by
3483        something that cannot possibly match the quantified thing, the quantifier
3484        is automatically "possessified".
3485    
3486    24. A recursive reference to a subpattern whose number was greater than 39
3487        went wrong under certain circumstances in UTF-8 mode. This bug could also
3488        have affected the operation of pcre_study().
3489    
3490    25. Realized that a little bit of performance could be had by replacing
3491        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3492    
3493    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3494    
3495    27. Possessive quantifiers such as a++ were previously implemented by turning
3496        them into atomic groups such as ($>a+). Now they have their own opcodes,
3497        which improves performance. This includes the automatically created ones
3498        from 23 above.
3499    
3500    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3501        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3502        the first matched character to be a colon. This applied both to named and
3503        numbered groups.
3504    
3505    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3506    
3507    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3508        building PCRE on Windows easier, so I have included it in the distribution.
3509    
3510    31. There is now a check in pcretest against a ridiculously large number being
3511        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3512        loop, the loop is abandoned.
3513    
3514    32. Forward references to subpatterns in conditions such as (?(2)...) where
3515        subpattern 2 is defined later cause pcre_compile() to search forwards in
3516        the pattern for the relevant set of parentheses. This search went wrong
3517        when there were unescaped parentheses in a character class, parentheses
3518        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3519    
3520    33. "Subroutine" calls and backreferences were previously restricted to
3521        referencing subpatterns earlier in the regex. This restriction has now
3522        been removed.
3523    
3524    34. Added a number of extra features that are going to be in Perl 5.10. On the
3525        whole, these are just syntactic alternatives for features that PCRE had
3526        previously implemented using the Python syntax or my own invention. The
3527        other formats are all retained for compatibility.
3528    
3529        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3530            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3531            also .NET compatible.
3532    
3533        (b) A recursion or subroutine call to a named group can now be defined as
3534            (?&name) as well as (?P>name).
3535    
3536        (c) A backreference to a named group can now be defined as \k<name> or
3537            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3538            5.10, are also .NET compatible.