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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1659 by ph10, Fri Jun 24 09:01:23 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.40 17-June-2016
8    -------------------------
9    
10    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
11        the match extended over a line boundary.
12    
13    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
14        MSVC static compilation, putting the first patch under a new option.
15    
16    
17    Version 8.39 14-June-2016
18    -------------------------
19    
20    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
21        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
22        misbehaved. This bug was found by the LLVM fuzzer.
23    
24    2.  Similar to the above, if an isolated \E was present between an item and its
25        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
26        bug was found by the LLVM fuzzer.
27    
28    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
29        working correctly in UCP mode.
30    
31    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
32        was set when the pmatch argument was NULL. It now returns REG_INVARG.
33    
34    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
35    
36    6.  An empty \Q\E sequence between an item and its qualifier caused
37        pcre_compile() to misbehave when auto callouts were enabled. This bug was
38        found by the LLVM fuzzer.
39    
40    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
41        space or a #-type comment that was followed by (?-x), which turns off
42        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
43        pcre_compile() assumed that (?-x) applied to the whole pattern and
44        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
45    
46    8.  A call of pcre_copy_named_substring() for a named substring whose number
47        was greater than the space in the ovector could cause a crash.
48    
49    9.  Yet another buffer overflow bug involved duplicate named groups with a
50        group that reset capture numbers (compare 8.38/7 below). Once again, I have
51        just allowed for more memory, even if not needed. (A proper fix is
52        implemented in PCRE2, but it involves a lot of refactoring.)
53    
54    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
55        start of the match to be earlier than the end.
56    
57    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
58    
59    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
60        assertion, caused pcretest to generate incorrect output, and also to read
61        uninitialized memory (detected by ASAN or valgrind).
62    
63    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
64        nested set of parentheses of sufficient size caused an overflow of the
65        compiling workspace (which was diagnosed, but of course is not desirable).
66    
67    14. And yet another buffer overflow bug involving duplicate named groups, this
68        time nested, with a nested back reference. Yet again, I have just allowed
69        for more memory, because anything more needs all the refactoring that has
70        been done for PCRE2. An example pattern that provoked this bug is:
71        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
72        registered as CVE-2016-1283.
73    
74    15. pcretest went into a loop if global matching was requested with an ovector
75        size less than 2. It now gives an error message. This bug was found by
76        afl-fuzz.
77    
78    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
79        ("assertion expected") when (?(?C) was not followed by an opening
80        parenthesis.
81    
82    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
83        actually affect anything, by sheer luck.
84    
85    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
86        static compilation.
87    
88    19. Modified the RunTest script to incorporate a valgrind suppressions file so
89        that certain errors, provoked by the SSE2 instruction set when JIT is used,
90        are ignored.
91    
92    20. A racing condition is fixed in JIT reported by Mozilla.
93    
94    21. Minor code refactor to avoid "array subscript is below array bounds"
95        compiler warning.
96    
97    22. Minor code refactor to avoid "left shift of negative number" warning.
98    
99    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
100        UCP support.
101    
102    24. Refactor to avoid compiler warnings in pcrecpp.cc.
103    
104    25. Refactor to fix a typo in pcre_jit_test.c
105    
106    26. Patch to support compiling pcrecpp.cc with Intel compiler.
107    
108    
109    Version 8.38 23-November-2015
110    -----------------------------
111    
112    1.  If a group that contained a recursive back reference also contained a
113        forward reference subroutine call followed by a non-forward-reference
114        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
115        compile correct code, leading to undefined behaviour or an internally
116        detected error. This bug was discovered by the LLVM fuzzer.
117    
118    2.  Quantification of certain items (e.g. atomic back references) could cause
119        incorrect code to be compiled when recursive forward references were
120        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
121        This bug was discovered by the LLVM fuzzer.
122    
123    3.  A repeated conditional group whose condition was a reference by name caused
124        a buffer overflow if there was more than one group with the given name.
125        This bug was discovered by the LLVM fuzzer.
126    
127    4.  A recursive back reference by name within a group that had the same name as
128        another group caused a buffer overflow. For example:
129        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
130    
131    5.  A forward reference by name to a group whose number is the same as the
132        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
133        a buffer overflow at compile time. This bug was discovered by the LLVM
134        fuzzer.
135    
136    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
137        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
138    
139    7.  Another buffer overflow bug involved duplicate named groups with a
140        reference between their definition, with a group that reset capture
141        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
142        fixed by always allowing for more memory, even if not needed. (A proper fix
143        is implemented in PCRE2, but it involves more refactoring.)
144    
145    8.  There was no check for integer overflow in subroutine calls such as (?123).
146    
147    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
148        being treated as a literal 'l' instead of causing an error.
149    
150    10. There was a buffer overflow if pcre_exec() was called with an ovector of
151        size 1. This bug was found by american fuzzy lop.
152    
153    11. If a non-capturing group containing a conditional group that could match
154        an empty string was repeated, it was not identified as matching an empty
155        string itself. For example: /^(?:(?(1)x|)+)+$()/.
156    
157    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
158        \a and \e in test subject lines.
159    
160    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
161        instead of the EBCDIC value.
162    
163    14. The handling of \c in an EBCDIC environment has been revised so that it is
164        now compatible with the specification in Perl's perlebcdic page.
165    
166    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
167        ASCII/Unicode. This has now been added to the list of characters that are
168        recognized as white space in EBCDIC.
169    
170    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
171        error (correctly) when used outside a class, but did not give an error
172        within a class.
173    
174    17. \h within a class was incorrectly compiled in EBCDIC environments.
175    
176    18. A pattern with an unmatched closing parenthesis that contained a backward
177        assertion which itself contained a forward reference caused buffer
178        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
179    
180    19. JIT should return with error when the compiled pattern requires more stack
181        space than the maximum.
182    
183    20. A possessively repeated conditional group that could match an empty string,
184        for example, /(?(R))*+/, was incorrectly compiled.
185    
186    21. Fix infinite recursion in the JIT compiler when certain patterns such as
187        /(?:|a|){100}x/ are analysed.
188    
189    22. Some patterns with character classes involving [: and \\ were incorrectly
190        compiled and could cause reading from uninitialized memory or an incorrect
191        error diagnosis.
192    
193    23. Pathological patterns containing many nested occurrences of [: caused
194        pcre_compile() to run for a very long time.
195    
196    24. A conditional group with only one branch has an implicit empty alternative
197        branch and must therefore be treated as potentially matching an empty
198        string.
199    
200    25. If (?R was followed by - or + incorrect behaviour happened instead of a
201        diagnostic.
202    
203    26. Arrange to give up on finding the minimum matching length for overly
204        complex patterns.
205    
206    27. Similar to (4) above: in a pattern with duplicated named groups and an
207        occurrence of (?| it is possible for an apparently non-recursive back
208        reference to become recursive if a later named group with the relevant
209        number is encountered. This could lead to a buffer overflow. Wen Guanxing
210        from Venustech ADLAB discovered this bug.
211    
212    28. If pcregrep was given the -q option with -c or -l, or when handling a
213        binary file, it incorrectly wrote output to stdout.
214    
215    29. The JIT compiler did not restore the control verb head in case of *THEN
216        control verbs. This issue was found by Karl Skomski with a custom LLVM
217        fuzzer.
218    
219    30. Error messages for syntax errors following \g and \k were giving inaccurate
220        offsets in the pattern.
221    
222    31. Added a check for integer overflow in conditions (?(<digits>) and
223        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
224        fuzzer.
225    
226    32. Handling recursive references such as (?2) when the reference is to a group
227        later in the pattern uses code that is very hacked about and error-prone.
228        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
229        give an internal error if it is obvious that compiling has gone wrong.
230    
231    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
232        This issue was found by Karl Skomski with a custom LLVM fuzzer.
233    
234    34. The JIT compiler should restore the control chain for empty possessive
235        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
236    
237    35. Match limit check added to JIT recursion. This issue was found by Karl
238        Skomski with a custom LLVM fuzzer.
239    
240    36. Yet another case similar to 27 above has been circumvented by an
241        unconditional allocation of extra memory. This issue is fixed "properly" in
242        PCRE2 by refactoring the way references are handled. Wen Guanxing
243        from Venustech ADLAB discovered this bug.
244    
245    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
246        with a custom LLVM fuzzer.
247    
248    38. Fixed a corner case of range optimization in JIT.
249    
250    39. An incorrect error "overran compiling workspace" was given if there were
251        exactly enough group forward references such that the last one extended
252        into the workspace safety margin. The next one would have expanded the
253        workspace. The test for overflow was not including the safety margin.
254    
255    40. A match limit issue is fixed in JIT which was found by Karl Skomski
256        with a custom LLVM fuzzer.
257    
258    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
259        work under Windows. (Why has it taken so long for anyone to notice?)
260    
261    42. In a character class such as [\W\p{Any}] where both a negative-type escape
262        ("not a word character") and a property escape were present, the property
263        escape was being ignored.
264    
265    43. Fix crash caused by very long (*MARK) or (*THEN) names.
266    
267    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
268        by a single ASCII character in a class item, was incorrectly compiled in
269        UCP mode. The POSIX class got lost, but only if the single character
270        followed it.
271    
272    45. [:punct:] in UCP mode was matching some characters in the range 128-255
273        that should not have been matched.
274    
275    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
276        class, all characters with code points greater than 255 are in the class.
277        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
278        such as \w are turned into Unicode properties), wide characters were not
279        correctly handled, and could fail to match.
280    
281    
282    Version 8.37 28-April-2015
283    --------------------------
284    
285    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
286        for those parentheses to be closed with whatever has been captured so far.
287        However, it was failing to mark any other groups between the hightest
288        capture so far and the currrent group as "unset". Thus, the ovector for
289        those groups contained whatever was previously there. An example is the
290        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
291    
292    2.  If an assertion condition was quantified with a minimum of zero (an odd
293        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
294    
295    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
296        unrecognized modifier, a crash could occur.
297    
298    4.  An attempt to do global matching in pcretest with a zero-length ovector
299        caused a crash.
300    
301    5.  Fixed a memory leak during matching that could occur for a subpattern
302        subroutine call (recursive or otherwise) if the number of captured groups
303        that had to be saved was greater than ten.
304    
305    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
306        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
307        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
308    
309    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
310        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
311    
312    8. When a pattern is compiled, it remembers the highest back reference so that
313       when matching, if the ovector is too small, extra memory can be obtained to
314       use instead. A conditional subpattern whose condition is a check on a
315       capture having happened, such as, for example in the pattern
316       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
317       setting the highest backreference number. This mattered only if pcre_exec()
318       was called with an ovector that was too small to hold the capture, and there
319       was no other kind of back reference (a situation which is probably quite
320       rare). The effect of the bug was that the condition was always treated as
321       FALSE when the capture could not be consulted, leading to a incorrect
322       behaviour by pcre_exec(). This bug has been fixed.
323    
324    9. A reference to a duplicated named group (either a back reference or a test
325       for being set in a conditional) that occurred in a part of the pattern where
326       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
327       to be incorrectly calculated, leading to overwriting.
328    
329    10. A mutually recursive set of back references such as (\2)(\1) caused a
330        segfault at study time (while trying to find the minimum matching length).
331        The infinite loop is now broken (with the minimum length unset, that is,
332        zero).
333    
334    11. If an assertion that was used as a condition was quantified with a minimum
335        of zero, matching went wrong. In particular, if the whole group had
336        unlimited repetition and could match an empty string, a segfault was
337        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
338        assertions to be quantified, but not if they are being used as conditions,
339        so the above pattern is faulted by Perl. PCRE has now been changed so that
340        it also rejects such patterns.
341    
342    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
343        failed to allow the zero-repeat case if pcre2_exec() was called with an
344        ovector too small to capture the group.
345    
346    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
347        Red Hat Product Security:
348    
349        (a) A crash if /K and /F were both set with the option to save the compiled
350        pattern.
351    
352        (b) Another crash if the option to print captured substrings in a callout
353        was combined with setting a null ovector, for example \O\C+ as a subject
354        string.
355    
356    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
357        forward reference repeated a large (but limited) number of times within a
358        repeated outer group that has a zero minimum quantifier, caused incorrect
359        code to be compiled, leading to the error "internal error:
360        previously-checked referenced subpattern not found" when an incorrect
361        memory address was read. This bug was reported as "heap overflow",
362        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
363        CVE-2015-2325.
364    
365    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
366        call within a group that also contained a recursive back reference caused
367        incorrect code to be compiled. This bug was reported as "heap overflow",
368        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
369        number CVE-2015-2326.
370    
371    24. Computing the size of the JIT read-only data in advance has been a source
372        of various issues, and new ones are still appear unfortunately. To fix
373        existing and future issues, size computation is eliminated from the code,
374        and replaced by on-demand memory allocation.
375    
376    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
377        adjacent to the end of the range, and the range contained characters with
378        more than one other case, caused incorrect behaviour when compiled in UTF
379        mode. In that example, the range a-j was left out of the class.
380    
381    26. Fix JIT compilation of conditional blocks, which assertion
382        is converted to (*FAIL). E.g: /(?(?!))/.
383    
384    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
385        discovered by the LLVM fuzzer.
386    
387    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
388        when this assertion was used as a condition, for example (?(?!)a|b). In
389        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
390        error about an unsupported item.
391    
392    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
393        possessification code could take exponential time to complete. A recursion
394        depth limit of 1000 has been imposed to limit the resources used by this
395        optimization.
396    
397    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
398        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
399        because \S ensures they are all in the class. The code for doing this was
400        interacting badly with the code for computing the amount of space needed to
401        compile the pattern, leading to a buffer overflow. This bug was discovered
402        by the LLVM fuzzer.
403    
404    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
405        other kinds of group caused stack overflow at compile time. This bug was
406        discovered by the LLVM fuzzer.
407    
408    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
409        between a subroutine call and its quantifier was incorrectly compiled,
410        leading to buffer overflow or other errors. This bug was discovered by the
411        LLVM fuzzer.
412    
413    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
414        assertion after (?(. The code was failing to check the character after
415        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
416        was discovered by the LLVM fuzzer.
417    
418    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
419        a fixed maximum following a group that contains a subroutine reference was
420        incorrectly compiled and could trigger buffer overflow. This bug was
421        discovered by the LLVM fuzzer.
422    
423    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
424        caused a stack overflow instead of the diagnosis of a non-fixed length
425        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
426    
427    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
428        (e.g. /(?<=\Ka)/) could make pcregrep loop.
429    
430    37. There was a similar problem to 36 in pcretest for global matches.
431    
432    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
433        and a subsequent item in the pattern caused a non-match, backtracking over
434        the repeated \X did not stop, but carried on past the start of the subject,
435        causing reference to random memory and/or a segfault. There were also some
436        other cases where backtracking after \C could crash. This set of bugs was
437        discovered by the LLVM fuzzer.
438    
439    39. The function for finding the minimum length of a matching string could take
440        a very long time if mutual recursion was present many times in a pattern,
441        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
442        method has been implemented. This infelicity was discovered by the LLVM
443        fuzzer.
444    
445    40. Static linking against the PCRE library using the pkg-config module was
446        failing on missing pthread symbols.
447    
448    
449    Version 8.36 26-September-2014
450    ------------------------------
451    
452    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
453        -Wmissing-field-initializers and -Wunused-parameter.
454    
455    2.  The tests for quantifiers being too big (greater than 65535) were being
456        applied after reading the number, and stupidly assuming that integer
457        overflow would give a negative number. The tests are now applied as the
458        numbers are read.
459    
460    3.  Tidy code in pcre_exec.c where two branches that used to be different are
461        now the same.
462    
463    4.  The JIT compiler did not generate match limit checks for certain
464        bracketed expressions with quantifiers. This may lead to exponential
465        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
466        issue should be resolved now.
467    
468    5.  Fixed an issue, which occures when nested alternatives are optimized
469        with table jumps.
470    
471    6.  Inserted two casts and changed some ints to size_t in the light of some
472        reported 64-bit compiler warnings (Bugzilla 1477).
473    
474    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
475        an empty string, which sometimes were behaving incorrectly in the
476        interpreter (though correctly in the JIT matcher). This pcretest input is
477        an example:
478    
479          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
480          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
481    
482        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
483        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
484        for an empty string was breaking the inner loop and carrying on at a lower
485        level, when possessive repeated groups should always return to a higher
486        level as they have no backtrack points in them. The empty string test now
487        occurs at the outer level.
488    
489    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
490        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
491    
492    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
493        doesn't).
494    
495    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
496        to hold a pattern to be incorrectly computed (too small) when there were
497        named back references to duplicated names. This could cause "internal
498        error: code overflow" or "double free or corruption" or other memory
499        handling errors.
500    
501    11. When named subpatterns had the same prefixes, back references could be
502        confused. For example, in this pattern:
503    
504          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
505    
506        the reference to 'Name' was incorrectly treated as a reference to a
507        duplicate name.
508    
509    12. A pattern such as /^s?c/mi8 where the optional character has more than
510        one "other case" was incorrectly compiled such that it would only try to
511        match starting at "c".
512    
513    13. When a pattern starting with \s was studied, VT was not included in the
514        list of possible starting characters; this should have been part of the
515        8.34/18 patch.
516    
517    14. If a character class started [\Qx]... where x is any character, the class
518        was incorrectly terminated at the ].
519    
520    15. If a pattern that started with a caseless match for a character with more
521        than one "other case" was studied, PCRE did not set up the starting code
522        unit bit map for the list of possible characters. Now it does. This is an
523        optimization improvement, not a bug fix.
524    
525    16. The Unicode data tables have been updated to Unicode 7.0.0.
526    
527    17. Fixed a number of memory leaks in pcregrep.
528    
529    18. Avoid a compiler warning (from some compilers) for a function call with
530        a cast that removes "const" from an lvalue by using an intermediate
531        variable (to which the compiler does not object).
532    
533    19. Incorrect code was compiled if a group that contained an internal recursive
534        back reference was optional (had quantifier with a minimum of zero). This
535        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
536        caused segmentation faults because of stack overflows at compile time.
537    
538    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
539        group that is quantified with an indefinite repeat, caused a compile-time
540        loop which used up all the system stack and provoked a segmentation fault.
541        This was not the same bug as 19 above.
542    
543    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
544        Patch by Mike Frysinger.
545    
546    
547    Version 8.35 04-April-2014
548    --------------------------
549    
550    1.  A new flag is set, when property checks are present in an XCLASS.
551        When this flag is not set, PCRE can perform certain optimizations
552        such as studying these XCLASS-es.
553    
554    2.  The auto-possessification of character sets were improved: a normal
555        and an extended character set can be compared now. Furthermore
556        the JIT compiler optimizes more character set checks.
557    
558    3.  Got rid of some compiler warnings for potentially uninitialized variables
559        that show up only when compiled with -O2.
560    
561    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
562        of a match later then the end of the match. The pcretest program was not
563        handling the case sensibly - it was outputting from the start to the next
564        binary zero. It now reports this situation in a message, and outputs the
565        text from the end to the start.
566    
567    5.  Fast forward search is improved in JIT. Instead of the first three
568        characters, any three characters with fixed position can be searched.
569        Search order: first, last, middle.
570    
571    6.  Improve character range checks in JIT. Characters are read by an inprecise
572        function now, which returns with an unknown value if the character code is
573        above a certain threshold (e.g: 256). The only limitation is that the value
574        must be bigger than the threshold as well. This function is useful when
575        the characters above the threshold are handled in the same way.
576    
577    7.  The macros whose names start with RAWUCHAR are placeholders for a future
578        mode in which only the bottom 21 bits of 32-bit data items are used. To
579        make this more memorable for those maintaining the code, the names have
580        been changed to start with UCHAR21, and an extensive comment has been added
581        to their definition.
582    
583    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
584        to the export list in Makefile.am (they were accidentally omitted from the
585        8.34 tarball).
586    
587    9.  The informational output from pcretest used the phrase "starting byte set"
588        which is inappropriate for the 16-bit and 32-bit libraries. As the output
589        for "first char" and "need char" really means "non-UTF-char", I've changed
590        "byte" to "char", and slightly reworded the output. The documentation about
591        these values has also been (I hope) clarified.
592    
593    10. Another JIT related optimization: use table jumps for selecting the correct
594        backtracking path, when more than four alternatives are present inside a
595        bracket.
596    
597    11. Empty match is not possible, when the minimum length is greater than zero,
598        and there is no \K in the pattern. JIT should avoid empty match checks in
599        such cases.
600    
601    12. In a caseless character class with UCP support, when a character with more
602        than one alternative case was not the first character of a range, not all
603        the alternative cases were added to the class. For example, s and \x{17f}
604        are both alternative cases for S: the class [RST] was handled correctly,
605        but [R-T] was not.
606    
607    13. The configure.ac file always checked for pthread support when JIT was
608        enabled. This is not used in Windows, so I have put this test inside a
609        check for the presence of windows.h (which was already tested for).
610    
611    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
612        The algorithm provides a way to skip certain starting offsets, and usually
613        faster than linear prefix searches.
614    
615    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
616        as for 'fr_FR' and 'french'. For some reason, however, it then used the
617        Windows-specific input and output files, which have 'french' screwed in.
618        So this could never have worked. One of the problems with locales is that
619        they aren't always the same. I have now updated RunTest so that it checks
620        the output of the locale test (test 3) against three different output
621        files, and it allows the test to pass if any one of them matches. With luck
622        this should make the test pass on some versions of Solaris where it was
623        failing. Because of the uncertainty, the script did not used to stop if
624        test 3 failed; it now does. If further versions of a French locale ever
625        come to light, they can now easily be added.
626    
627    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
628        there was a message during ./configure, but it did not stop. This now
629        provokes an error. The invalid example in README has been corrected.
630        If a value less than the minimum is given, the minimum value has always
631        been used, but now a warning is given.
632    
633    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
634        was a bug in the test system, which is now fixed. Also, the list of various
635        configurations that are tested for each release did not have one with both
636        16/32 bits and --enable-bar-anycrlf. It now does.
637    
638    18. pcretest was missing "-C bsr" for displaying the \R default setting.
639    
640    19. Little endian PowerPC systems are supported now by the JIT compiler.
641    
642    20. The fast forward newline mechanism could enter to an infinite loop on
643        certain invalid UTF-8 input. Although we don't support these cases
644        this issue can be fixed by a performance optimization.
645    
646    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
647        not take account if existing stack usage. There is now a new global
648        variable called pcre_stack_guard that can be set to point to an external
649        function to check stack availability. It is called at the start of
650        processing every parenthesized group.
651    
652    22. A typo in the code meant that in ungreedy mode the max/min qualifier
653        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
654        not match "ab".
655    
656    23. When UTF was disabled, the JIT program reported some incorrect compile
657        errors. These messages are silenced now.
658    
659    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
660        compiler.
661    
662    25. Change all the temporary files used in RunGrepTest to be different to those
663        used by RunTest so that the tests can be run simultaneously, for example by
664        "make -j check".
665    
666    
667    Version 8.34 15-December-2013
668    -----------------------------
669    
670    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
671        executable memory. Patch inspired by Carsten Klein.
672    
673    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
674        this macro is never tested and has no effect, because the work to support
675        coverage involves only compiling and linking options and special targets in
676        the Makefile. The comment in config.h implied that defining the macro would
677        enable coverage support, which is totally false. There was also support for
678        setting this macro in the CMake files (my fault, I just copied it from
679        configure). SUPPORT_GCOV has now been removed.
680    
681    3.  Make a small performance improvement in strlen16() and strlen32() in
682        pcretest.
683    
684    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
685        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
686        these cases). There was also one in pcretest.c.
687    
688    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
689    
690    6.  In UTF mode, the code for checking whether a group could match an empty
691        string (which is used for indefinitely repeated groups to allow for
692        breaking an infinite loop) was broken when the group contained a repeated
693        negated single-character class with a character that occupied more than one
694        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
695        UTF-8 mode). The effect was undefined: the group might or might not be
696        deemed as matching an empty string, or the program might have crashed.
697    
698    7.  The code for checking whether a group could match an empty string was not
699        recognizing that \h, \H, \v, \V, and \R must match a character.
700    
701    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
702        an empty string. If it can, pcretest shows this in its information output.
703    
704    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
705        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
706        matched by pcre_exec() without using JIT:
707    
708        (a) If the rest of the pattern did not match after a maximal run of
709            grapheme clusters, the code for backing up to try with fewer of them
710            did not always back up over a full grapheme when characters that do not
711            have the modifier quality were involved, e.g. Hangul syllables.
712    
713        (b) If the match point in a subject started with modifier character, and
714            there was no match, the code could incorrectly back up beyond the match
715            point, and potentially beyond the first character in the subject,
716            leading to a segfault or an incorrect match result.
717    
718    10. A conditional group with an assertion condition could lead to PCRE
719        recording an incorrect first data item for a match if no other first data
720        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
721        first data item, and therefore matched "ca" after "c" instead of at the
722        start.
723    
724    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
725        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
726    
727    12. The source of pcregrep now includes z/OS-specific code so that it can be
728        compiled for z/OS as part of the special z/OS distribution.
729    
730    13. Added the -T and -TM options to pcretest.
731    
732    14. The code in pcre_compile.c for creating the table of named capturing groups
733        has been refactored. Instead of creating the table dynamically during the
734        actual compiling pass, the information is remembered during the pre-compile
735        pass (on the stack unless there are more than 20 named groups, in which
736        case malloc() is used) and the whole table is created before the actual
737        compile happens. This has simplified the code (it is now nearly 150 lines
738        shorter) and prepared the way for better handling of references to groups
739        with duplicate names.
740    
741    15. A back reference to a named subpattern when there is more than one of the
742        same name now checks them in the order in which they appear in the pattern.
743        The first one that is set is used for the reference. Previously only the
744        first one was inspected. This change makes PCRE more compatible with Perl.
745    
746    16. Unicode character properties were updated from Unicode 6.3.0.
747    
748    17. The compile-time code for auto-possessification has been refactored, based
749        on a patch by Zoltan Herczeg. It now happens after instead of during
750        compilation. The code is cleaner, and more cases are handled. The option
751        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
752        options in pcretest are provided to set it. It can also be set by
753        (*NO_AUTO_POSSESS) at the start of a pattern.
754    
755    18. The character VT has been added to the default ("C" locale) set of
756        characters that match \s and are generally treated as white space,
757        following this same change in Perl 5.18. There is now no difference between
758        "Perl space" and "POSIX space". Whether VT is treated as white space in
759        other locales depends on the locale.
760    
761    19. The code for checking named groups as conditions, either for being set or
762        for being recursed, has been refactored (this is related to 14 and 15
763        above). Processing unduplicated named groups should now be as fast at
764        numerical groups, and processing duplicated groups should be faster than
765        before.
766    
767    20. Two patches to the CMake build system, by Alexander Barkov:
768    
769          (1) Replace the "source" command by "." in CMakeLists.txt because
770              "source" is a bash-ism.
771    
772          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
773              without these the CMake build does not work on Solaris.
774    
775    21. Perl has changed its handling of \8 and \9. If there is no previously
776        encountered capturing group of those numbers, they are treated as the
777        literal characters 8 and 9 instead of a binary zero followed by the
778        literals. PCRE now does the same.
779    
780    22. Following Perl, added \o{} to specify codepoints in octal, making it
781        possible to specify values greater than 0777 and also making them
782        unambiguous.
783    
784    23. Perl now gives an error for missing closing braces after \x{... instead of
785        treating the string as literal. PCRE now does the same.
786    
787    24. RunTest used to grumble if an inappropriate test was selected explicitly,
788        but just skip it when running all tests. This make it awkward to run ranges
789        of tests when one of them was inappropriate. Now it just skips any
790        inappropriate tests, as it always did when running all tests.
791    
792    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
793        character types such as \d or \w, too many callouts were inserted, and the
794        data that they returned was rubbish.
795    
796    26. In UCP mode, \s was not matching two of the characters that Perl matches,
797        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
798        were matched by \h. The code has now been refactored so that the lists of
799        the horizontal and vertical whitespace characters used for \h and \v (which
800        are defined only in one place) are now also used for \s.
801    
802    27. Add JIT support for the 64 bit TileGX architecture.
803        Patch by Jiong Wang (Tilera Corporation).
804    
805    28. Possessive quantifiers for classes (both explicit and automatically
806        generated) now use special opcodes instead of wrapping in ONCE brackets.
807    
808    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
809        (because it's meaningless), this was not happening when PCRE_CASELESS was
810        set. Not wrong, but inefficient.
811    
812    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
813        properties for \w, \d, etc) is present in a test regex. Otherwise if the
814        test contains no characters greater than 255, Perl doesn't realise it
815        should be using Unicode semantics.
816    
817    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
818        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
819        does in Unicode mode.
820    
821    32. Added the "forbid" facility to pcretest so that putting tests into the
822        wrong test files can sometimes be quickly detected.
823    
824    33. There is now a limit (default 250) on the depth of nesting of parentheses.
825        This limit is imposed to control the amount of system stack used at compile
826        time. It can be changed at build time by --with-parens-nest-limit=xxx or
827        the equivalent in CMake.
828    
829    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
830        errors. Perl warns for these when in warning mode, but PCRE has no facility
831        for giving warnings.
832    
833    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
834        However, this was not working for (?!) because it is optimized to (*FAIL),
835        for which PCRE does not allow quantifiers. The optimization is now disabled
836        when a quantifier follows (?!). I can't see any use for this, but it makes
837        things uniform.
838    
839    36. Perl no longer allows group names to start with digits, so I have made this
840        change also in PCRE. It simplifies the code a bit.
841    
842    37. In extended mode, Perl ignores spaces before a + that indicates a
843        possessive quantifier. PCRE allowed a space before the quantifier, but not
844        before the possessive +. It now does.
845    
846    38. The use of \K (reset reported match start) within a repeated possessive
847        group such as (a\Kb)*+ was not working.
848    
849    40. Document that the same character tables must be used at compile time and
850        run time, and that the facility to pass tables to pcre_exec() and
851        pcre_dfa_exec() is for use only with saved/restored patterns.
852    
853    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
854        features for Builds with MSVC:
855    
856        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
857           builds). The .rc files can be used to set FileDescription and many other
858           attributes.
859    
860        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
861           This allows higher-level build scripts which want .pdb files to avoid
862           hard-coding the exact files needed."
863    
864    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
865        mean "start of word" and "end of word", respectively, as a transition aid.
866    
867    43. A minimizing repeat of a class containing codepoints greater than 255 in
868        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
869        compiled to use the heap for recursion.
870    
871    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
872        is configured.
873    
874    
875    Version 8.33 28-May-2013
876    ------------------------
877    
878    1.  Added 'U' to some constants that are compared to unsigned integers, to
879        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
880        variables that are added to signed variables, to ensure the result is
881        signed and can be negated.
882    
883    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
884        CMake config files.
885    
886    3.  Revise the creation of config.h.generic so that all boolean macros are
887        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
888        overriding via -D on the command line possible.
889    
890    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
891        to unsigned int is reported to make a quite noticeable speed difference in
892        a specific Windows environment. Testing on Linux did also appear to show
893        some benefit (and it is clearly not harmful). Also fixed the definition of
894        Xop which should be unsigned.
895    
896    5.  Related to (4), changing the definition of the intermediate variable cc
897        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
898        improvements.
899    
900    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
901        unnecessary spaces.
902    
903    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
904        and later.
905    
906    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
907    
908    9.  Optimizing fast_forward_start_bits in JIT.
909    
910    10. Adding support for callouts in JIT, and fixing some issues revealed
911        during this work. Namely:
912    
913        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
914    
915        (b) Minimum length was not checked before the matching is started.
916    
917    11. The value of capture_last that is passed to callouts was incorrect in some
918        cases when there was a capture on one path that was subsequently abandoned
919        after a backtrack. Also, the capture_last value is now reset after a
920        recursion, since all captures are also reset in this case.
921    
922    12. The interpreter no longer returns the "too many substrings" error in the
923        case when an overflowing capture is in a branch that is subsequently
924        abandoned after a backtrack.
925    
926    13. In the pathological case when an offset vector of size 2 is used, pcretest
927        now prints out the matched string after a yield of 0 or 1.
928    
929    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
930        Only supported by the JIT compiler at the moment.
931    
932    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
933    
934    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
935        offset of the starting point of the matching process, provided the offsets
936        vector is large enough.
937    
938    17. The \A escape now records a lookbehind value of 1, though its execution
939        does not actually inspect the previous character. This is to ensure that,
940        in partial multi-segment matching, at least one character from the old
941        segment is retained when a new segment is processed. Otherwise, if there
942        are no lookbehinds in the pattern, \A might match incorrectly at the start
943        of a new segment.
944    
945    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
946    
947    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
948        gives some modest performance improvement in 8-bit mode.
949    
950    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
951        be expressed in certain programming languages using Universal Character
952        Names.
953    
954    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
955        which points out that "non characters" are not "characters that may not
956        appear in Unicode strings" but rather "characters that are reserved for
957        internal use and have only local meaning".
958    
959    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
960        there was a conditional group that depended on an assertion, if the
961        assertion was false, the callout that immediately followed the alternation
962        in the condition was skipped when pcre_exec() was used for matching.
963    
964    23. Allow an explicit callout to be inserted before an assertion that is the
965        condition for a conditional group, for compatibility with automatic
966        callouts, which always insert a callout at this point.
967    
968    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
969        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
970    
971    25. (*PRUNE) is now supported by the JIT compiler.
972    
973    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
974    
975    27. Fix the case where there are two or more SKIPs with arguments that may be
976        ignored.
977    
978    28. (*SKIP) is now supported by the JIT compiler.
979    
980    29. (*THEN) is now supported by the JIT compiler.
981    
982    30. Update RunTest with additional test selector options.
983    
984    31. The way PCRE handles backtracking verbs has been changed in two ways.
985    
986        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
987        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
988        backtracking. In some cases this makes it more Perl-compatible, but Perl's
989        rather obscure rules do not always do the same thing.
990    
991        (2) Previously, backtracking verbs were confined within assertions. This is
992        no longer the case for positive assertions, except for (*ACCEPT). Again,
993        this sometimes improves Perl compatibility, and sometimes does not.
994    
995    32. A number of tests that were in test 2 because Perl did things differently
996        have been moved to test 1, because either Perl or PCRE has changed, and
997        these tests are now compatible.
998    
999    32. Backtracking control verbs are now handled in the same way in JIT and
1000        interpreter.
1001    
1002    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1003        contained a forward subroutine reference caused a compile error.
1004    
1005    34. Auto-detect and optimize limited repetitions in JIT.
1006    
1007    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1008        blocking (*UTF) etc.
1009    
1010    36. In the interpreter, maximizing pattern repetitions for characters and
1011        character types now use tail recursion, which reduces stack usage.
1012    
1013    37. The value of the max lookbehind was not correctly preserved if a compiled
1014        and saved regex was reloaded on a host of different endianness.
1015    
1016    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1017        of the compiled pattern block, expand the flags field from 16 to 32 bits
1018        because it was almost full.
1019    
1020    39. Try madvise first before posix_madvise.
1021    
1022    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1023        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1024        original change locked out all matches of empty strings. This has been
1025        changed so that one match of an empty string per line is recognized.
1026        Subsequent searches on the same line (for colouring or for --only-matching,
1027        for example) do not recognize empty strings.
1028    
1029    41. Applied a user patch to fix a number of spelling mistakes in comments.
1030    
1031    42. Data lines longer than 65536 caused pcretest to crash.
1032    
1033    43. Clarified the data type for length and startoffset arguments for pcre_exec
1034        and pcre_dfa_exec in the function-specific man pages, where they were
1035        explicitly stated to be in bytes, never having been updated. I also added
1036        some clarification to the pcreapi man page.
1037    
1038    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1039        a segmentation fault.
1040    
1041    
1042    Version 8.32 30-November-2012
1043    -----------------------------
1044    
1045    1.  Improved JIT compiler optimizations for first character search and single
1046        character iterators.
1047    
1048    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1049        Patch by Daniel Richard G.
1050    
1051    3.  Single character iterator optimizations in the JIT compiler.
1052    
1053    4.  Improved JIT compiler optimizations for character ranges.
1054    
1055    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1056        Reported by Giuseppe D'Angelo.
1057    
1058    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1059        of a line, was being set incorrectly in cases where .* appeared inside
1060        atomic brackets at the start of a pattern, or where there was a subsequent
1061        *PRUNE or *SKIP.
1062    
1063    7.  Improved instruction cache flush for POWER/PowerPC.
1064        Patch by Daniel Richard G.
1065    
1066    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1067        grep:
1068    
1069        (a) There is now no limit to the number of patterns to be matched.
1070    
1071        (b) An error is given if a pattern is too long.
1072    
1073        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1074            are now supported.
1075    
1076        (d) --exclude-from and --include-from (multiple use) have been added.
1077    
1078        (e) Exclusions and inclusions now apply to all files and directories, not
1079            just to those obtained from scanning a directory recursively.
1080    
1081        (f) Multiple uses of -f and --file-list are now supported.
1082    
1083        (g) In a Windows environment, the default for -d has been changed from
1084            "read" (the GNU grep default) to "skip", because otherwise the presence
1085            of a directory in the file list provokes an error.
1086    
1087        (h) The documentation has been revised and clarified in places.
1088    
1089    9.  Improve the matching speed of capturing brackets.
1090    
1091    10. Changed the meaning of \X so that it now matches a Unicode extended
1092        grapheme cluster.
1093    
1094    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1095        out POSIX threads when JIT support is configured.
1096    
1097    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1098    
1099    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1100        structure could go wrong in environments where size_t is not the same size
1101        as int.
1102    
1103    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1104        set.
1105    
1106    15. The EBCDIC support had decayed; later updates to the code had included
1107        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1108        general tidy up of EBCDIC-related issues, and the documentation was also
1109        not quite right. There is now a test that can be run on ASCII systems to
1110        check some of the EBCDIC-related things (but is it not a full test).
1111    
1112    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1113        in a small tidy to the code.
1114    
1115    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1116    
1117    18. If the --only-matching (-o) option in pcregrep is specified multiple
1118        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1119        substrings matched by the 1st and 2nd capturing parentheses. A separating
1120        string can be specified by --om-separator (default empty).
1121    
1122    19. Improving the first n character searches.
1123    
1124    20. Turn case lists for horizontal and vertical white space into macros so that
1125        they are defined only once.
1126    
1127    21. This set of changes together give more compatible Unicode case-folding
1128        behaviour for characters that have more than one other case when UCP
1129        support is available.
1130    
1131        (a) The Unicode property table now has offsets into a new table of sets of
1132            three or more characters that are case-equivalent. The MultiStage2.py
1133            script that generates these tables (the pcre_ucd.c file) now scans
1134            CaseFolding.txt instead of UnicodeData.txt for character case
1135            information.
1136    
1137        (b) The code for adding characters or ranges of characters to a character
1138            class has been abstracted into a generalized function that also handles
1139            case-independence. In UTF-mode with UCP support, this uses the new data
1140            to handle characters with more than one other case.
1141    
1142        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1143            whose other case is greater than 256 are now correctly matched
1144            caselessly. Previously, the high codepoint matched the low one, but not
1145            vice versa.
1146    
1147        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1148            of the new class addition function, using character lists defined as
1149            macros alongside the case definitions of 20 above.
1150    
1151        (e) Caseless back references now work with characters that have more than
1152            one other case.
1153    
1154        (f) General caseless matching of characters with more than one other case
1155            is supported.
1156    
1157    22. Unicode character properties were updated from Unicode 6.2.0
1158    
1159    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1160    
1161    24. Add support for 32-bit character strings, and UTF-32
1162    
1163    25. Major JIT compiler update (code refactoring and bugfixing).
1164        Experimental Sparc 32 support is added.
1165    
1166    26. Applied a modified version of Daniel Richard G's patch to create
1167        pcre.h.generic and config.h.generic by "make" instead of in the
1168        PrepareRelease script.
1169    
1170    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1171        pcre_compile.c when checking for a zero character.
1172    
1173    28. Introducing a native interface for JIT. Through this interface, the compiled
1174        machine code can be directly executed. The purpose of this interface is to
1175        provide fast pattern matching, so several sanity checks are not performed.
1176        However, feature tests are still performed. The new interface provides
1177        1.4x speedup compared to the old one.
1178    
1179    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1180        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1181        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1182    
1183    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1184        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1185        the "old" RFC 2279). Instead, it ended up passing a negative length to
1186        pcre_exec().
1187    
1188    31. Add support for GCC's visibility feature to hide internal functions.
1189    
1190    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1191        "unknown -C option" after outputting 0 or 1.
1192    
1193    33. There is now support for generating a code coverage report for the test
1194        suite in environments where gcc is the compiler and lcov is installed. This
1195        is mainly for the benefit of the developers.
1196    
1197    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1198        unaddressable using valgrind annotations, allowing valgrind to detect
1199        invalid memory accesses. This is mainly for the benefit of the developers.
1200    
1201    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1202    
1203    26. Give configure error if --enable-cpp but no C++ compiler found.
1204    
1205    
1206    Version 8.31 06-July-2012
1207    -------------------------
1208    
1209    1.  Fixing a wrong JIT test case and some compiler warnings.
1210    
1211    2.  Removed a bashism from the RunTest script.
1212    
1213    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1214        to unsigned type, result still unsigned" that was given by an MS compiler
1215        on encountering the code "-sizeof(xxx)".
1216    
1217    4.  Partial matching support is added to the JIT compiler.
1218    
1219    5.  Fixed several bugs concerned with partial matching of items that consist
1220        of more than one character:
1221    
1222        (a) /^(..)\1/ did not partially match "aba" because checking references was
1223            done on an "all or nothing" basis. This also applied to repeated
1224            references.
1225    
1226        (b) \R did not give a hard partial match if \r was found at the end of the
1227            subject.
1228    
1229        (c) \X did not give a hard partial match after matching one or more
1230            characters at the end of the subject.
1231    
1232        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1233            a partial match for the string "\r".
1234    
1235        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1236            a partial match for a CR character at the end of the subject string.
1237    
1238    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1239        running pcretest, the text "(JIT)" added to the output whenever JIT is
1240        actually used to run the match.
1241    
1242    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1243        or /S+[+] with a digit between 1 and 7.
1244    
1245    8.  OP_NOT now supports any UTF character not just single-byte ones.
1246    
1247    9.  (*MARK) control verb is now supported by the JIT compiler.
1248    
1249    10. The command "./RunTest list" lists the available tests without actually
1250        running any of them. (Because I keep forgetting what they all are.)
1251    
1252    11. Add PCRE_INFO_MAXLOOKBEHIND.
1253    
1254    12. Applied a (slightly modified) user-supplied patch that improves performance
1255        when the heap is used for recursion (compiled with --disable-stack-for-
1256        recursion). Instead of malloc and free for each heap frame each time a
1257        logical recursion happens, frames are retained on a chain and re-used where
1258        possible. This sometimes gives as much as 30% improvement.
1259    
1260    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1261        call.
1262    
1263    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1264    
1265    15. It is now possible to link pcretest with libedit as an alternative to
1266        libreadline.
1267    
1268    16. (*COMMIT) control verb is now supported by the JIT compiler.
1269    
1270    17. The Unicode data tables have been updated to Unicode 6.1.0.
1271    
1272    18. Added --file-list option to pcregrep.
1273    
1274    19. Added binary file support to pcregrep, including the -a, --binary-files,
1275        -I, and --text options.
1276    
1277    20. The madvise function is renamed for posix_madvise for QNX compatibility
1278        reasons. Fixed by Giuseppe D'Angelo.
1279    
1280    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1281    
1282    22. Changed the option for creating symbolic links for 16-bit man pages from
1283        -s to -sf so that re-installing does not cause issues.
1284    
1285    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1286    
1287    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1288        matches in certain environments (the workspace was not being correctly
1289        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1290        some of the workspace data at the beginning of a restart.
1291    
1292    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1293        was not doing so when it should - probably a typo introduced by SVN 528
1294        (change 8.10/14).
1295    
1296    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1297        \w+ when the character tables indicated that \x{c4} was a word character.
1298        There were several related cases, all because the tests for doing a table
1299        lookup were testing for characters less than 127 instead of 255.
1300    
1301    27. If a pattern contains capturing parentheses that are not used in a match,
1302        their slots in the ovector are set to -1. For those that are higher than
1303        any matched groups, this happens at the end of processing. In the case when
1304        there were back references that the ovector was too small to contain
1305        (causing temporary malloc'd memory to be used during matching), and the
1306        highest capturing number was not used, memory off the end of the ovector
1307        was incorrectly being set to -1. (It was using the size of the temporary
1308        memory instead of the true size.)
1309    
1310    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1311        ovector size, it uses memory at the end of the block that it has got.
1312    
1313    29. Check for an overlong MARK name and give an error at compile time. The
1314        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1315    
1316    30. JIT compiler update.
1317    
1318    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1319        Rill for the patch.
1320    
1321    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1322    
1323    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1324    
1325    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1326        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1327        was enabled.
1328    
1329    35. Improve JIT code generation for greedy plus quantifier.
1330    
1331    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1332        1 to "aa" instead of to an empty string. The bug affected repeated groups
1333        that could potentially match an empty string.
1334    
1335    37. Optimizing single character iterators in JIT.
1336    
1337    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1338        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1339        codepoints that are too big for the mode are faulted, and in a UTF mode,
1340        disallowed codepoints are also faulted.
1341    
1342    39. If PCRE was compiled with UTF support, in three places in the DFA
1343        matcher there was code that should only have been obeyed in UTF mode, but
1344        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1345        processing when bytes with values greater than 127 were present. In 16-bit
1346        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1347        both cases the values are those that cannot be the first data item in a UTF
1348        character. The three items that might have provoked this were recursions,
1349        possessively repeated groups, and atomic groups.
1350    
1351    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1352        and pcregrep, because some OS require shared objects to be explicitly
1353        passed to ld, causing the link step to fail if they are not.
1354    
1355    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1356        mode, patterns that started with \h* or \R* might be incorrectly matched.
1357    
1358    
1359    Version 8.30 04-February-2012
1360    -----------------------------
1361    
1362    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1363        name is defined in ctype.h.
1364    
1365    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1366        only in quite long subpatterns.
1367    
1368    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1369        since it was replaced by pcre_fullinfo() in February 2000.
1370    
1371    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1372        match a (*MARK), and the match failed at the start of the subject, a
1373        reference to memory before the start of the subject could occur. This bug
1374        was introduced by fix 17 of release 8.21.
1375    
1376    5.  A reference to an unset group with zero minimum repetition was giving
1377        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1378        /(another)?(\1?)test/ matched against "hello world test". This bug was
1379        introduced in release 8.13.
1380    
1381    6.  Add support for 16-bit character strings (a large amount of work involving
1382        many changes and refactorings).
1383    
1384    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1385        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1386        from a file.
1387    
1388    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1389        rounding is not applied in this particular case).
1390    
1391    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1392        if they appear, or are escaped, in patterns.
1393    
1394    10. Get rid of a number of -Wunused-but-set-variable warnings.
1395    
1396    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1397        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1398        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1399        also returns the mark "x". This bug applied to capturing parentheses,
1400        non-capturing parentheses, and atomic parentheses. It also applied to some
1401        assertions.
1402    
1403    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1404        information out of configure.ac instead of relying on pcre.h.generic, which
1405        is not stored in the repository.
1406    
1407    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1408        -lreadline.
1409    
1410    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1411    
1412    15. Applied Graycode's patch to put the top-level frame on the stack rather
1413        than the heap when not using the stack for recursion. This gives a
1414        performance improvement in many cases when recursion is not deep.
1415    
1416    16. Experimental code added to "pcretest -C" to output the stack frame size.
1417    
1418    
1419    Version 8.21 12-Dec-2011
1420    ------------------------
1421    
1422    1.  Updating the JIT compiler.
1423    
1424    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1425        are added as well.
1426    
1427    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1428        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1429        calling _pcre_jit_exec. Some extra comments are added.
1430    
1431    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1432        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1433        was introduced by change 18 for 8.20.
1434    
1435    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1436        ECMA-262 standard.
1437    
1438    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1439        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1440        This bug was probably introduced by change 9 of 8.13.
1441    
1442    7.  While fixing 6 above, I noticed that a number of other items were being
1443        incorrectly rejected as "not fixed length". This arose partly because newer
1444        opcodes had not been added to the fixed-length checking code. I have (a)
1445        corrected the bug and added tests for these items, and (b) arranged for an
1446        error to occur if an unknown opcode is encountered while checking for fixed
1447        length instead of just assuming "not fixed length". The items that were
1448        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1449        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1450        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1451    
1452    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1453        being incorrectly compiled and would have given unpredicatble results.
1454    
1455    9.  A possessively repeated subpattern with minimum repeat count greater than
1456        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1457        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1458        the first (A) could occur when it should not.
1459    
1460    10. Add a cast and remove a redundant test from the code.
1461    
1462    11. JIT should use pcre_malloc/pcre_free for allocation.
1463    
1464    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1465        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1466        is anything other than /usr, -L is still shown).
1467    
1468    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1469    
1470    14. Perl does not support \N without a following name in a [] class; PCRE now
1471        also gives an error.
1472    
1473    15. If a forward reference was repeated with an upper limit of around 2000,
1474        it caused the error "internal error: overran compiling workspace". The
1475        maximum number of forward references (including repeats) was limited by the
1476        internal workspace, and dependent on the LINK_SIZE. The code has been
1477        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1478        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1479        of around 200,000 forward references. While doing this, I also speeded up
1480        the filling in of repeated forward references.
1481    
1482    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1483        incorrectly expecting the subject to contain another "a" after the start.
1484    
1485    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1486        in the match, the SKIP should be ignored. This was not happening; instead
1487        the SKIP was being treated as NOMATCH. For patterns such as
1488        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1489        tested.
1490    
1491    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1492        now much more compatible with Perl, in particular in cases where the result
1493        is a non-match for a non-anchored pattern. For example, if
1494        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1495        "m", where previously it did not return a name. A side effect of this
1496        change is that for partial matches, the last encountered mark name is
1497        returned, as for non matches. A number of tests that were previously not
1498        Perl-compatible have been moved into the Perl-compatible test files. The
1499        refactoring has had the pleasing side effect of removing one argument from
1500        the match() function, thus reducing its stack requirements.
1501    
1502    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1503        subsequent uses of /S (without +) incorrectly behaved like /S+.
1504    
1505    21. Retrieve executable code size support for the JIT compiler and fixing
1506        some warnings.
1507    
1508    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1509        not work when the shorter character appeared right at the end of the
1510        subject string.
1511    
1512    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1513        systems.
1514    
1515    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1516        output it when the /M option is used in pcretest.
1517    
1518    25. The CheckMan script was not being included in the distribution. Also, added
1519        an explicit "perl" to run Perl scripts from the PrepareRelease script
1520        because this is reportedly needed in Windows.
1521    
1522    26. If study data was being save in a file and studying had not found a set of
1523        "starts with" bytes for the pattern, the data written to the file (though
1524        never used) was taken from uninitialized memory and so caused valgrind to
1525        complain.
1526    
1527    27. Updated RunTest.bat as provided by Sheri Pierce.
1528    
1529    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1530    
1531    29. Computation of memory usage for the table of capturing group names was
1532        giving an unnecessarily large value.
1533    
1534    
1535    Version 8.20 21-Oct-2011
1536    ------------------------
1537    
1538    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1539        a POSIX class. After further experiments with Perl, which convinced me that
1540        Perl has bugs and confusions, a closing square bracket is no longer allowed
1541        in a POSIX name. This bug also affected patterns with classes that started
1542        with full stops.
1543    
1544    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1545        captured substring, but while checking the failing first alternative,
1546        substring 1 is temporarily captured. If the output vector supplied to
1547        pcre_exec() was not big enough for this capture, the yield of the function
1548        was still zero ("insufficient space for captured substrings"). This cannot
1549        be totally fixed without adding another stack variable, which seems a lot
1550        of expense for a edge case. However, I have improved the situation in cases
1551        such as /(a)(b)x|abc/ matched against "abc", where the return code
1552        indicates that fewer than the maximum number of slots in the ovector have
1553        been set.
1554    
1555    3.  Related to (2) above: when there are more back references in a pattern than
1556        slots in the output vector, pcre_exec() uses temporary memory during
1557        matching, and copies in the captures as far as possible afterwards. It was
1558        using the entire output vector, but this conflicts with the specification
1559        that only 2/3 is used for passing back captured substrings. Now it uses
1560        only the first 2/3, for compatibility. This is, of course, another edge
1561        case.
1562    
1563    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1564        main code base, and can be used by building with --enable-jit. When this is
1565        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1566        runtime --no-jit option is given.
1567    
1568    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1569        ovector, the return from the function was zero, implying that there were
1570        other matches that did not fit. The correct "exactly full" value is now
1571        returned.
1572    
1573    6.  If a subpattern that was called recursively or as a subroutine contained
1574        (*PRUNE) or any other control that caused it to give a non-standard return,
1575        invalid errors such as "Error -26 (nested recursion at the same subject
1576        position)" or even infinite loops could occur.
1577    
1578    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1579        computing the minimum length on reaching *ACCEPT, and so ended up with the
1580        wrong value of 1 rather than 0. Further investigation indicates that
1581        computing a minimum subject length in the presence of *ACCEPT is difficult
1582        (think back references, subroutine calls), and so I have changed the code
1583        so that no minimum is registered for a pattern that contains *ACCEPT.
1584    
1585    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1586        it was not handled as intended. [But see 16 below.]
1587    
1588    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1589        Sheri Pierce.
1590    
1591    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1592        the first byte in a match must be "a".
1593    
1594    11. Change 17 for 8.13 increased the recursion depth for patterns like
1595        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1596        pattern contains any instances of (*THEN). If it does not, the old
1597        optimizations are restored. It would be nice to do this on a per-group
1598        basis, but at the moment that is not feasible.
1599    
1600    12. In some environments, the output of pcretest -C is CRLF terminated. This
1601        broke RunTest's code that checks for the link size. A single white space
1602        character after the value is now allowed for.
1603    
1604    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1605        For "fr", it uses the Windows-specific input and output files.
1606    
1607    14. If (*THEN) appeared in a group that was called recursively or as a
1608        subroutine, it did not work as intended. [But see next item.]
1609    
1610    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1611        pattern fragments (but not containing any | characters). If A and B are
1612        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1613        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1614        D. In other words, Perl considers parentheses that do not contain any |
1615        characters to be part of a surrounding alternative, whereas PCRE was
1616        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1617        differently. PCRE now behaves in the same way as Perl, except in the case
1618        of subroutine/recursion calls such as (?1) which have in any case always
1619        been different (but PCRE had them first :-).
1620    
1621    16. Related to 15 above: Perl does not treat the | in a conditional group as
1622        creating alternatives. Such a group is treated in the same way as an
1623        ordinary group without any | characters when processing (*THEN). PCRE has
1624        been changed to match Perl's behaviour.
1625    
1626    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1627        RunGrepTest script failed.
1628    
1629    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1630        inevitable for groups that contain captures, but it can lead to a lot of
1631        stack use in large patterns. The old behaviour has been restored for atomic
1632        groups that do not contain any capturing parentheses.
1633    
1634    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1635        suppress the check for a minimum subject length at run time. (If it was
1636        given to pcre_exec() or pcre_dfa_exec() it did work.)
1637    
1638    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1639        fail to work when decoding hex characters in data strings in EBCDIC
1640        environments.
1641    
1642    21. It appears that in at least one Mac OS environment, the isxdigit() function
1643        is implemented as a macro that evaluates to its argument more than once,
1644        contravening the C 90 Standard (I haven't checked a later standard). There
1645        was an instance in pcretest which caused it to go wrong when processing
1646        \x{...} escapes in subject strings. The has been rewritten to avoid using
1647        things like p++ in the argument of isxdigit().
1648    
1649    
1650    Version 8.13 16-Aug-2011
1651    ------------------------
1652    
1653    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1654    
1655    2.  Two minor typos in pcre_internal.h have been fixed.
1656    
1657    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1658        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1659        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1660    
1661    4.  There were a number of related bugs in the code for matching backrefences
1662        caselessly in UTF-8 mode when codes for the characters concerned were
1663        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1664        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1665        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1666        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1667        2-byte code at the end of the subject (it thought there wasn't enough data
1668        left).
1669    
1670    5.  Comprehensive information about what went wrong is now returned by
1671        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1672        as the output vector has at least 2 elements. The offset of the start of
1673        the failing character and a reason code are placed in the vector.
1674    
1675    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1676        now returned is for the first byte of the failing character, instead of the
1677        last byte inspected. This is an incompatible change, but I hope it is small
1678        enough not to be a problem. It makes the returned offset consistent with
1679        pcre_exec() and pcre_dfa_exec().
1680    
1681    7.  pcretest now gives a text phrase as well as the error number when
1682        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1683        failure, the offset and reason code are output.
1684    
1685    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1686        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1687        back over a single character (\n). This seems wrong (because it treated the
1688        two characters as a single entity when going forwards), conflicts with the
1689        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1690        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1691        has been changed.
1692    
1693    9.  Some internal refactoring has changed the processing so that the handling
1694        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1695        time (the PCRE_DOTALL option was changed this way some time ago: version
1696        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1697        which was always a bit of a fudge. It also means that there is one less
1698        argument for the match() function, which reduces its stack requirements
1699        slightly. This change also fixes an incompatibility with Perl: the pattern
1700        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1701    
1702    10. More internal refactoring has drastically reduced the number of recursive
1703        calls to match() for possessively repeated groups such as (abc)++ when
1704        using pcre_exec().
1705    
1706    11. While implementing 10, a number of bugs in the handling of groups were
1707        discovered and fixed:
1708    
1709        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1710        (a|)*(?1) gave a compile-time internal error.
1711        ((a|)+)+  did not notice that the outer group could match an empty string.
1712        (^a|^)+   was not marked as anchored.
1713        (.*a|.*)+ was not marked as matching at start or after a newline.
1714    
1715    12. Yet more internal refactoring has removed another argument from the match()
1716        function. Special calls to this function are now indicated by setting a
1717        value in a variable in the "match data" data block.
1718    
1719    13. Be more explicit in pcre_study() instead of relying on "default" for
1720        opcodes that mean there is no starting character; this means that when new
1721        ones are added and accidentally left out of pcre_study(), testing should
1722        pick them up.
1723    
1724    14. The -s option of pcretest has been documented for ages as being an old
1725        synonym of -m (show memory usage). I have changed it to mean "force study
1726        for every regex", that is, assume /S for every regex. This is similar to -i
1727        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1728        using it. It makes it easier to run collections of tests with and without
1729        study enabled, and thereby test pcre_study() more easily. All the standard
1730        tests are now run with and without -s (but some patterns can be marked as
1731        "never study" - see 20 below).
1732    
1733    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1734        restoration of the capturing data to the outer values was not happening
1735        correctly.
1736    
1737    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1738        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1739        pattern had matched an empty string, and so incorrectly returned a no
1740        match.
1741    
1742    17. There was optimizing code for the last branch of non-capturing parentheses,
1743        and also for the obeyed branch of a conditional subexpression, which used
1744        tail recursion to cut down on stack usage. Unfortunately, now that there is
1745        the possibility of (*THEN) occurring in these branches, tail recursion is
1746        no longer possible because the return has to be checked for (*THEN). These
1747        two optimizations have therefore been removed. [But see 8.20/11 above.]
1748    
1749    18. If a pattern containing \R was studied, it was assumed that \R always
1750        matched two bytes, thus causing the minimum subject length to be
1751        incorrectly computed because \R can also match just one byte.
1752    
1753    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1754        was incorrectly computed.
1755    
1756    20. If /S is present twice on a test pattern in pcretest input, it now
1757        *disables* studying, thereby overriding the use of -s on the command line
1758        (see 14 above). This is necessary for one or two tests to keep the output
1759        identical in both cases.
1760    
1761    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1762        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1763    
1764    22. When an atomic group that contained a capturing parenthesis was
1765        successfully matched, but the branch in which it appeared failed, the
1766        capturing was not being forgotten if a higher numbered group was later
1767        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1768        group 1 to "a", when in fact it should be unset. This applied to multi-
1769        branched capturing and non-capturing groups, repeated or not, and also to
1770        positive assertions (capturing in negative assertions does not happen
1771        in PCRE) and also to nested atomic groups.
1772    
1773    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1774        subject after a captured substring, to make it easier to tell which of a
1775        number of identical substrings has been captured.
1776    
1777    24. The way atomic groups are processed by pcre_exec() has been changed so that
1778        if they are repeated, backtracking one repetition now resets captured
1779        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1780        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1781        "aaa". Previously, it would have been "a". As part of this code
1782        refactoring, the way recursive calls are handled has also been changed.
1783    
1784    25. If an assertion condition captured any substrings, they were not passed
1785        back unless some other capturing happened later. For example, if
1786        (?(?=(a))a) was matched against "a", no capturing was returned.
1787    
1788    26. When studying a pattern that contained subroutine calls or assertions,
1789        the code for finding the minimum length of a possible match was handling
1790        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1791        group 1 called group 2 while simultaneously a separate group 2 called group
1792        1). A stack overflow occurred in this case. I have fixed this by limiting
1793        the recursion depth to 10.
1794    
1795    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1796        Fortmann. This supports explicit test numbers on the command line, and has
1797        argument validation and error reporting.
1798    
1799    28. An instance of \X with an unlimited repeat could fail if at any point the
1800        first character it looked at was a mark character.
1801    
1802    29. Some minor code refactoring concerning Unicode properties and scripts
1803        should reduce the stack requirement of match() slightly.
1804    
1805    30. Added the '=' option to pcretest to check the setting of unused capturing
1806        slots at the end of the pattern, which are documented as being -1, but are
1807        not included in the return count.
1808    
1809    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1810        compiled something random. Now it gives a compile-time error (as does
1811        Perl).
1812    
1813    32. A *MARK encountered during the processing of a positive assertion is now
1814        recorded and passed back (compatible with Perl).
1815    
1816    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1817        had alternative anchored branches, the search for a second match in a line
1818        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1819        matched the line "0102" twice. The same bug affected patterns that started
1820        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1821        twice.
1822    
1823    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1824        does, and because of capturing effects, quantifying parenthesized
1825        assertions may at times be useful. Quantifiers are now allowed for
1826        parenthesized assertions.
1827    
1828    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1829    
1830    36. \g was being checked for fancy things in a character class, when it should
1831        just be a literal "g".
1832    
1833    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1834        appearance of a nested POSIX class supersedes an apparent external class.
1835        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1836        unescaped square brackets may also appear as part of class names. For
1837        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1838        more like Perl. (But see 8.20/1 above.)
1839    
1840    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1841        was because it thought it was \N{name}, which is not supported).
1842    
1843    39. Add minix to OS list not supporting the -S option in pcretest.
1844    
1845    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1846        cannot analyze patterns in sufficient detail to catch mutual recursions
1847        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1848        subgroup is called recursively as a subpattern for a second time at the
1849        same position in the subject string. In previous releases this might have
1850        been caught by the recursion limit, or it might have run out of stack.
1851    
1852    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1853        happen only once. PCRE was, however incorrectly giving a compile time error
1854        "recursive call could loop indefinitely" because it cannot analyze the
1855        pattern in sufficient detail. The compile time test no longer happens when
1856        PCRE is compiling a conditional subpattern, but actual runaway loops are
1857        now caught at runtime (see 40 above).
1858    
1859    42. It seems that Perl allows any characters other than a closing parenthesis
1860        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1861        has been changed to be the same.
1862    
1863    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1864        as not to get warnings when autogen.sh is called. Also changed
1865        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1866    
1867    44. To help people who use pcregrep to scan files containing exceedingly long
1868        lines, the following changes have been made:
1869    
1870        (a) The default value of the buffer size parameter has been increased from
1871            8K to 20K. (The actual buffer used is three times this size.)
1872    
1873        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1874            PCRE is built.
1875    
1876        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1877            to be set at run time.
1878    
1879        (d) Numerical values in pcregrep options can be followed by K or M, for
1880            example --buffer-size=50K.
1881    
1882        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1883            given and the return code is set to 2.
1884    
1885    45. Add a pointer to the latest mark to the callout data block.
1886    
1887    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1888        partial match of an empty string instead of no match. This was specific to
1889        the use of ".".
1890    
1891    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1892        complete match instead of a partial match. This bug was dependent on both
1893        the PCRE_UTF8 and PCRE_DOTALL options being set.
1894    
1895    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1896        starting byte set, because \b was not being ignored.
1897    
1898    
1899    Version 8.12 15-Jan-2011
1900    ------------------------
1901    
1902    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1903        checks for such things as part of the documentation building process.
1904    
1905    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1906        --match-limit and --recursion-limit options (added for 8.11). In
1907        particular, this made one of the standard tests fail. (The integer value
1908        went into the wrong half of a long int.)
1909    
1910    3.  If the --colour option was given to pcregrep with -v (invert match), it
1911        did strange things, either producing crazy output, or crashing. It should,
1912        of course, ignore a request for colour when reporting lines that do not
1913        match.
1914    
1915    4.  Another pcregrep bug caused similar problems if --colour was specified with
1916        -M (multiline) and the pattern match finished with a line ending.
1917    
1918    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1919        matched in multiline mode, the following line was shown as part of the
1920        match. This seems wrong, so I have changed it.
1921    
1922    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1923        the check for further matches in the same line (so they could be coloured)
1924        to overrun the end of the current line. If another match was found, it was
1925        incorrectly shown (and then shown again when found in the next line).
1926    
1927    7.  If pcregrep was compiled under Windows, there was a reference to the
1928        function pcregrep_exit() before it was defined. I am assuming this was
1929        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1930        reported by a user. I've moved the definition above the reference.
1931    
1932    
1933    Version 8.11 10-Dec-2010
1934    ------------------------
1935    
1936    1.  (*THEN) was not working properly if there were untried alternatives prior
1937        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1938        backtracked to try for "b" instead of moving to the next alternative branch
1939        at the same level (in this case, to look for "c"). The Perl documentation
1940        is clear that when (*THEN) is backtracked onto, it goes to the "next
1941        alternative in the innermost enclosing group".
1942    
1943    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1944        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1945        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1946        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1947        (*THEN).
1948    
1949    3.  If \s appeared in a character class, it removed the VT character from
1950        the class, even if it had been included by some previous item, for example
1951        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1952        of \s, but is part of the POSIX "space" class.)
1953    
1954    4.  A partial match never returns an empty string (because you can always
1955        match an empty string at the end of the subject); however the checking for
1956        an empty string was starting at the "start of match" point. This has been
1957        changed to the "earliest inspected character" point, because the returned
1958        data for a partial match starts at this character. This means that, for
1959        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1960        (previously it gave "no match").
1961    
1962    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1963        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1964        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1965        has an implication that the given string is incomplete (because a partial
1966        match is preferred over a full match). For this reason, these items now
1967        give a partial match in this situation. [Aside: previously, the one case
1968        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1969        match rather than a full match, which was wrong by the old rules, but is
1970        now correct.]
1971    
1972    6.  There was a bug in the handling of #-introduced comments, recognized when
1973        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1974        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1975        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1976        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1977        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1978        places in pcre_compile().
1979    
1980    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1981        comments when looking ahead for named forward references to subpatterns,
1982        the only newline sequence it recognized was NL. It now handles newlines
1983        according to the set newline convention.
1984    
1985    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1986        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1987        cater for a lack of strerror(). These oversights have been fixed.
1988    
1989    9.  Added --match-limit and --recursion-limit to pcregrep.
1990    
1991    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1992    
1993    11. When the -o option was used, pcregrep was setting a return code of 1, even
1994        when matches were found, and --line-buffered was not being honoured.
1995    
1996    12. Added an optional parentheses number to the -o and --only-matching options
1997        of pcregrep.
1998    
1999    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2000        can match an empty string. The code to do it in pcretest and pcredemo
2001        needed fixing:
2002    
2003        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2004            only one byte after an empty string match just before CRLF (this case
2005            just got forgotten; "any" and "anycrlf" were OK).
2006    
2007        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2008            mode when an empty string match preceded an ASCII character followed by
2009            a non-ASCII character. (The code for advancing by one character rather
2010            than one byte was nonsense.)
2011    
2012        (c) The pcredemo.c sample program did not have any code at all to handle
2013            the cases when CRLF is a valid newline sequence.
2014    
2015    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2016        as a starting offset was within the subject string. There is now a new
2017        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2018        negative or greater than the length of the string. In order to test this,
2019        pcretest is extended to allow the setting of negative starting offsets.
2020    
2021    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2022        starting offset points to the beginning of a UTF-8 character was
2023        unnecessarily clumsy. I tidied it up.
2024    
2025    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2026        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2027    
2028    17. Nobody had reported that the --include_dir option, which was added in
2029        release 7.7 should have been called --include-dir (hyphen, not underscore)
2030        for compatibility with GNU grep. I have changed it to --include-dir, but
2031        left --include_dir as an undocumented synonym, and the same for
2032        --exclude-dir, though that is not available in GNU grep, at least as of
2033        release 2.5.4.
2034    
2035    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2036        characters from a string of bytes) have been redefined so as not to use
2037        loops, in order to improve performance in some environments. At the same
2038        time, I abstracted some of the common code into auxiliary macros to save
2039        repetition (this should not affect the compiled code).
2040    
2041    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2042        compile-time error is now given if \c is not followed by an ASCII
2043        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2044        different, and any byte value is allowed.)
2045    
2046    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2047        START_OPTIMIZE option, which is now allowed at compile time - but just
2048        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2049        to pcregrep and other applications that have no direct access to PCRE
2050        options. The new /Y option in pcretest sets this option when calling
2051        pcre_compile().
2052    
2053    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2054        back references. Groups containing recursive back references were forced to
2055        be atomic by that change, but in the case of named groups, the amount of
2056        memory required was incorrectly computed, leading to "Failed: internal
2057        error: code overflow". This has been fixed.
2058    
2059    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2060        pcretest.c, to avoid build problems in some Borland environments.
2061    
2062    
2063    Version 8.10 25-Jun-2010
2064    ------------------------
2065    
2066    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2067        THEN.
2068    
2069    2.  (*ACCEPT) was not working when inside an atomic group.
2070    
2071    3.  Inside a character class, \B is treated as a literal by default, but
2072        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2073        causes the error). The code is unchanged, but I tidied the documentation.
2074    
2075    4.  Inside a character class, PCRE always treated \R and \X as literals,
2076        whereas Perl faults them if its -w option is set. I have changed PCRE so
2077        that it faults them when PCRE_EXTRA is set.
2078    
2079    5.  Added support for \N, which always matches any character other than
2080        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2081    
2082    6.  When compiling pcregrep with newer versions of gcc which may have
2083        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2084        declared with attribute warn_unused_result" were given. Just casting the
2085        result to (void) does not stop the warnings; a more elaborate fudge is
2086        needed. I've used a macro to implement this.
2087    
2088    7.  Minor change to pcretest.c to avoid a compiler warning.
2089    
2090    8.  Added four artifical Unicode properties to help with an option to make
2091        \s etc use properties (see next item). The new properties are: Xan
2092        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2093    
2094    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2095        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2096        this option. Modified pcretest to add /W to test this facility. Added
2097        REG_UCP to make it available via the POSIX interface.
2098    
2099    10. Added --line-buffered to pcregrep.
2100    
2101    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2102        studied, and the match started with a letter with a code point greater than
2103        127 whose first byte was different to the first byte of the other case of
2104        the letter, the other case of this starting letter was not recognized
2105        (#976).
2106    
2107    12. If a pattern that was studied started with a repeated Unicode property
2108        test, for example, \p{Nd}+, there was the theoretical possibility of
2109        setting up an incorrect bitmap of starting bytes, but fortunately it could
2110        not have actually happened in practice until change 8 above was made (it
2111        added property types that matched character-matching opcodes).
2112    
2113    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2114        possible starting bytes for non-anchored patterns.
2115    
2116    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2117        \R, and also a number of cases that involve Unicode properties, both
2118        explicit and implicit when PCRE_UCP is set.
2119    
2120    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2121        input, it could crash or give wrong results if characters with values
2122        greater than 0xc0 were present in the subject string. (Detail: it assumed
2123        UTF-8 input when processing these items.)
2124    
2125    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2126        size_t is 64-bit (#991).
2127    
2128    17. Added a check for running out of memory when PCRE is compiled with
2129        --disable-stack-for-recursion (#990).
2130    
2131    18. If the last data line in a file for pcretest does not have a newline on
2132        the end, a newline was missing in the output.
2133    
2134    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2135        less than 128) in its various bitmaps. However, there is a facility for
2136        generating tables according to the current locale when PCRE is compiled. It
2137        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2138        characters, are recognized by isspace() and therefore were getting set in
2139        these tables, and indeed these tables seem to approximate to ISO 8859. This
2140        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2141        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2142        which of course cannot start UTF-8 characters. I have changed the code so
2143        that only real ASCII characters (less than 128) and the correct starting
2144        bytes for UTF-8 encodings are set for characters greater than 127 when in
2145        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2146        altogether.)
2147    
2148    20. Added the /T option to pcretest so as to be able to run tests with non-
2149        standard character tables, thus making it possible to include the tests
2150        used for 19 above in the standard set of tests.
2151    
2152    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2153        reference to a subpattern the other side of a comment that contains an
2154        opening parenthesis caused either an internal compiling error, or a
2155        reference to the wrong subpattern.
2156    
2157    
2158    Version 8.02 19-Mar-2010
2159    ------------------------
2160    
2161    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2162    
2163    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2164        configured.
2165    
2166    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2167        original author of that file, following a query about its status.
2168    
2169    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2170        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2171    
2172    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2173        quantifier applied to a forward-referencing subroutine call, could compile
2174        incorrect code or give the error "internal error: previously-checked
2175        referenced subpattern not found".
2176    
2177    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2178        variables to point to external functions. For these systems, therefore,
2179        pcre_malloc etc. are now initialized to local functions that call the
2180        relevant global functions.
2181    
2182    7.  There were two entries missing in the vectors called coptable and poptable
2183        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2184        I've fixed the data, and added a kludgy way of testing at compile time that
2185        the lengths are correct (equal to the number of opcodes).
2186    
2187    8.  Following on from 7, I added a similar kludge to check the length of the
2188        eint vector in pcreposix.c.
2189    
2190    9.  Error texts for pcre_compile() are held as one long string to avoid too
2191        much relocation at load time. To find a text, the string is searched,
2192        counting zeros. There was no check for running off the end of the string,
2193        which could happen if a new error number was added without updating the
2194        string.
2195    
2196    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2197    
2198    11. \K was not working if it appeared in an atomic group or in a group that
2199        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2200        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2201        the assertion is positive, but not if it is negative.
2202    
2203    12. Change 11 fortuitously reduced the size of the stack frame used in the
2204        "match()" function of pcre_exec.c by one pointer. Forthcoming
2205        implementation of support for (*MARK) will need an extra pointer on the
2206        stack; I have reserved it now, so that the stack frame size does not
2207        decrease.
2208    
2209    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2210        item in branch that calls a recursion is a subroutine call - as in the
2211        second branch in the above example - was incorrectly given the compile-
2212        time error "recursive call could loop indefinitely" because pcre_compile()
2213        was not correctly checking the subroutine for matching a non-empty string.
2214    
2215    14. The checks for overrunning compiling workspace could trigger after an
2216        overrun had occurred. This is a "should never occur" error, but it can be
2217        triggered by pathological patterns such as hundreds of nested parentheses.
2218        The checks now trigger 100 bytes before the end of the workspace.
2219    
2220    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2221    
2222    
2223    Version 8.01 19-Jan-2010
2224    ------------------------
2225    
2226    1.  If a pattern contained a conditional subpattern with only one branch (in
2227        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2228        computed the wrong minimum data length (which is of course zero for such
2229        subpatterns). This could cause incorrect "no match" results.
2230    
2231    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2232        the pattern is reset in the first branch, pcre_compile() failed with
2233        "internal error: code overflow at offset...". This happened only when
2234        the reset was to the original external option setting. (An optimization
2235        abstracts leading options settings into an external setting, which was the
2236        cause of this.)
2237    
2238    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2239        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2240        assertion pattern did not match (meaning that the assertion was true), it
2241        was incorrectly treated as false if the SKIP had been reached during the
2242        matching. This also applied to assertions used as conditions.
2243    
2244    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2245        assertion subpattern, including such a pattern used as a condition,
2246        unpredictable results occurred, instead of the error return
2247        PCRE_ERROR_DFA_UITEM.
2248    
2249    5.  The C++ GlobalReplace function was not working like Perl for the special
2250        situation when an empty string is matched. It now does the fancy magic
2251        stuff that is necessary.
2252    
2253    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2254        removed. (These were left over from very, very early versions of PCRE.)
2255    
2256    7.  Some cosmetic changes to the code to make life easier when compiling it
2257        as part of something else:
2258    
2259        (a) Change DEBUG to PCRE_DEBUG.
2260    
2261        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2262            called "current" as "current_branch", to prevent a collision with the
2263            Linux macro when compiled as a kernel module.
2264    
2265        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2266            prevent a collision with the Linux macro when compiled as a kernel
2267            module.
2268    
2269    8.  In pcre_compile() there are some checks for integer overflows that used to
2270        cast potentially large values to (double). This has been changed to that
2271        when building, a check for int64_t is made, and if it is found, it is used
2272        instead, thus avoiding the use of floating point arithmetic. (There is no
2273        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2274        double.
2275    
2276    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2277        2005 (difference between two addresses compared to an unsigned value).
2278    
2279    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2280        custom one, because of the following reported problem in Windows:
2281    
2282          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2283              under Win32.
2284          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2285              therefore missing the function definition.
2286          - The compiler thus generates a "C" signature for the test function.
2287          - The linker fails to find the "C" function.
2288          - PCRE fails to configure if asked to do so against libbz2.
2289    
2290    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2291        messages were output:
2292    
2293          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2294          rerunning libtoolize, to keep the correct libtool macros in-tree.
2295          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2296    
2297        I have done both of these things.
2298    
2299    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2300        most of the time, it *can* run out if it is given a pattern that contains a
2301        runaway infinite recursion. I updated the discussion in the pcrestack man
2302        page.
2303    
2304    13. Now that we have gone to the x.xx style of version numbers, the minor
2305        version may start with zero. Using 08 or 09 is a bad idea because users
2306        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2307        interpreted as invalid octal numbers. I've updated the previous comment in
2308        configure.ac, and also added a check that gives an error if 08 or 09 are
2309        used.
2310    
2311    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2312        causing partial matching to fail when the end of the subject matched \W
2313        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2314    
2315    15. There were some discrepancies between the declarations in pcre_internal.h
2316        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2317        their definitions. The declarations used "const uschar *" and the
2318        definitions used USPTR. Even though USPTR is normally defined as "const
2319        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2320        reported that: "This difference in casting confuses some C++ compilers, for
2321        example, SunCC recognizes above declarations as different functions and
2322        generates broken code for hbpcre." I have changed the declarations to use
2323        USPTR.
2324    
2325    16. GNU libtool is named differently on some systems. The autogen.sh script now
2326        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2327        (FreeBSD).
2328    
2329    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2330        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2331        comment: "Figure out how to create a longlong from a string: strtoll and
2332        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2333        instance, but it only takes 2 args instead of 3!"
2334    
2335    18. A subtle bug concerned with back references has been fixed by a change of
2336        specification, with a corresponding code fix. A pattern such as
2337        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2338        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2339        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2340        same bug. Such groups have to be quantified to be useful, or contained
2341        inside another quantified group. (If there's no repetition, the reference
2342        can never match.) The problem arises because, having left the group and
2343        moved on to the rest of the pattern, a later failure that backtracks into
2344        the group uses the captured value from the final iteration of the group
2345        rather than the correct earlier one. I have fixed this in PCRE by forcing
2346        any group that contains a reference to itself to be an atomic group; that
2347        is, there cannot be any backtracking into it once it has completed. This is
2348        similar to recursive and subroutine calls.
2349    
2350    
2351    Version 8.00 19-Oct-09
2352    ----------------------
2353    
2354    1.  The table for translating pcre_compile() error codes into POSIX error codes
2355        was out-of-date, and there was no check on the pcre_compile() error code
2356        being within the table. This could lead to an OK return being given in
2357        error.
2358    
2359    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2360        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2361        in a Windows environment.
2362    
2363    3.  The pcregrep --count option prints the count for each file even when it is
2364        zero, as does GNU grep. However, pcregrep was also printing all files when
2365        --files-with-matches was added. Now, when both options are given, it prints
2366        counts only for those files that have at least one match. (GNU grep just
2367        prints the file name in this circumstance, but including the count seems
2368        more useful - otherwise, why use --count?) Also ensured that the
2369        combination -clh just lists non-zero counts, with no names.
2370    
2371    4.  The long form of the pcregrep -F option was incorrectly implemented as
2372        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2373        but it seems right to fix it, and I didn't think it was worth preserving
2374        the old behaviour.
2375    
2376    5.  The command line items --regex=pattern and --regexp=pattern were not
2377        recognized by pcregrep, which required --regex pattern or --regexp pattern
2378        (with a space rather than an '='). The man page documented the '=' forms,
2379        which are compatible with GNU grep; these now work.
2380    
2381    6.  No libpcreposix.pc file was created for pkg-config; there was just
2382        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2383    
2384    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2385        when UCP support is not needed, by modifying the Python script that
2386        generates it from Unicode data files. This should not matter if the module
2387        is correctly used as a library, but I received one complaint about 50K of
2388        unwanted data. My guess is that the person linked everything into his
2389        program rather than using a library. Anyway, it does no harm.
2390    
2391    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2392        was a minimum greater than 1 for a wide character in a possessive
2393        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2394        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2395        character. Chaos in the form of incorrect output or a compiling loop could
2396        result.
2397    
2398    9.  The restrictions on what a pattern can contain when partial matching is
2399        requested for pcre_exec() have been removed. All patterns can now be
2400        partially matched by this function. In addition, if there are at least two
2401        slots in the offset vector, the offset of the earliest inspected character
2402        for the match and the offset of the end of the subject are set in them when
2403        PCRE_ERROR_PARTIAL is returned.
2404    
2405    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2406        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2407        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2408        and may be more useful for multi-segment matching.
2409    
2410    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2411        used to be given if ever the end of the subject was reached; now it is
2412        given only if matching could not proceed because another character was
2413        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2414        string "Z", which now yields "no match" instead of "partial match". In the
2415        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2416        final character ended with (*FAIL).
2417    
2418    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2419        if the pattern had a "must contain" character that was already found in the
2420        earlier partial match, unless partial matching was again requested. For
2421        example, with the pattern /dog.(body)?/, the "must contain" character is
2422        "g". If the first part-match was for the string "dog", restarting with
2423        "sbody" failed. This bug has been fixed.
2424    
2425    13. The string returned by pcre_dfa_exec() after a partial match has been
2426        changed so that it starts at the first inspected character rather than the
2427        first character of the match. This makes a difference only if the pattern
2428        starts with a lookbehind assertion or \b or \B (\K is not supported by
2429        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2430        matching functions compatible, and I think it's the right thing to do.
2431    
2432    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2433        so that the demonstration program is easily available in environments where
2434        PCRE has not been installed from source.
2435    
2436    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2437        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2438        library.
2439    
2440    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2441        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2442        is not the first non-POSIX option to be added. Clearly some people find
2443        these options useful.
2444    
2445    17. If a caller to the POSIX matching function regexec() passes a non-zero
2446        value for nmatch with a NULL value for pmatch, the value of
2447        nmatch is forced to zero.
2448    
2449    18. RunGrepTest did not have a test for the availability of the -u option of
2450        the diff command, as RunTest does. It now checks in the same way as
2451        RunTest, and also checks for the -b option.
2452    
2453    19. If an odd number of negated classes containing just a single character
2454        interposed, within parentheses, between a forward reference to a named
2455        subpattern and the definition of the subpattern, compilation crashed with
2456        an internal error, complaining that it could not find the referenced
2457        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2458        [The bug was that it was starting one character too far in when skipping
2459        over the character class, thus treating the ] as data rather than
2460        terminating the class. This meant it could skip too much.]
2461    
2462    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2463        /g option in pcretest when the pattern contains \K, which makes it possible
2464        to have an empty string match not at the start, even when the pattern is
2465        anchored. Updated pcretest and pcredemo to use this option.
2466    
2467    21. If the maximum number of capturing subpatterns in a recursion was greater
2468        than the maximum at the outer level, the higher number was returned, but
2469        with unset values at the outer level. The correct (outer level) value is
2470        now given.
2471    
2472    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2473        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2474        make it do so. The string so far is captured, making this feature
2475        compatible with Perl.
2476    
2477    23. The tests have been re-organized, adding tests 11 and 12, to make it
2478        possible to check the Perl 5.10 features against Perl 5.10.
2479    
2480    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2481        pattern matches a fixed length string. PCRE did not allow this; now it
2482        does. Neither allows recursion.
2483    
2484    25. I finally figured out how to implement a request to provide the minimum
2485        length of subject string that was needed in order to match a given pattern.
2486        (It was back references and recursion that I had previously got hung up
2487        on.) This code has now been added to pcre_study(); it finds a lower bound
2488        to the length of subject needed. It is not necessarily the greatest lower
2489        bound, but using it to avoid searching strings that are too short does give
2490        some useful speed-ups. The value is available to calling programs via
2491        pcre_fullinfo().
2492    
2493    26. While implementing 25, I discovered to my embarrassment that pcretest had
2494        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2495        study optimizations had never been tested with that matching function.
2496        Oops. What is worse, even when it was passed study data, there was a bug in
2497        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2498        were also very few tests of studied patterns with pcre_dfa_exec().
2499    
2500    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2501        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2502        on the other side of the coin, they are no longer allowed to have different
2503        names, because these cannot be distinguished in PCRE, and this has caused
2504        confusion. (This is a difference from Perl.)
2505    
2506    28. When duplicate subpattern names are present (necessarily with different
2507        numbers, as required by 27 above), and a test is made by name in a
2508        conditional pattern, either for a subpattern having been matched, or for
2509        recursion in such a pattern, all the associated numbered subpatterns are
2510        tested, and the overall condition is true if the condition is true for any
2511        one of them. This is the way Perl works, and is also more like the way
2512        testing by number works.
2513    
2514    
2515    Version 7.9 11-Apr-09
2516    ---------------------
2517    
2518    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2519        (pcretest), all targets were linked against these libraries. This included
2520        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2521        libraries. This caused unwanted dependencies to be created. This problem
2522        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2523        pcretest is linked with readline.
2524    
2525    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2526        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2527        moved outside it again, because FALSE and TRUE are already defined in AIX,
2528        but BOOL is not.
2529    
2530    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2531        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2532    
2533    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2534        hyphens) following file names and line numbers when outputting matching
2535        lines. This is not true; no spaces are inserted. I have also clarified the
2536        wording for the --colour (or --color) option.
2537    
2538    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2539        was not coloured; this is different to GNU grep, so I have changed it to be
2540        the same.
2541    
2542    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2543        each matching line was coloured. Now it goes on to look for further matches
2544        of any of the test patterns, which is the same behaviour as GNU grep.
2545    
2546    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2547        doesn't make sense to accept an empty string match in pcregrep, so I have
2548        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2549        seems to be how GNU grep behaves. [But see later change 40 for release
2550        8.33.]
2551    
2552    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2553        start or after a newline", because the conditional assertion was not being
2554        correctly handled. The rule now is that both the assertion and what follows
2555        in the first alternative must satisfy the test.
2556    
2557    9.  If auto-callout was enabled in a pattern with a conditional group whose
2558        condition was an assertion, PCRE could crash during matching, both with
2559        pcre_exec() and pcre_dfa_exec().
2560    
2561    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2562        used for matching.
2563    
2564    11. Unicode property support in character classes was not working for
2565        characters (bytes) greater than 127 when not in UTF-8 mode.
2566    
2567    12. Added the -M command line option to pcretest.
2568    
2569    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2570    
2571    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2572    
2573    16. Added comments and documentation about mis-use of no_arg in the C++
2574        wrapper.
2575    
2576    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2577        from Martin Jerabek that uses macro names for all relevant character and
2578        string constants.
2579    
2580    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2581        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2582        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2583        these, but not everybody uses configure.
2584    
2585    19. A conditional group that had only one branch was not being correctly
2586        recognized as an item that could match an empty string. This meant that an
2587        enclosing group might also not be so recognized, causing infinite looping
2588        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2589        with the subject "ab", where knowledge that the repeated group can match
2590        nothing is needed in order to break the loop.
2591    
2592    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2593        exec(), but without supplying a callout function, matching went wrong.
2594    
2595    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2596        leak if the size of the offset vector was greater than 30. When the vector
2597        is smaller, the saved offsets during recursion go onto a local stack
2598        vector, but for larger vectors malloc() is used. It was failing to free
2599        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2600        error, in fact).
2601    
2602    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2603        heapframe that is used only when UTF-8 support is enabled. This caused no
2604        problem, but was untidy.
2605    
2606    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2607        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2608        included within another project.
2609    
2610    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2611        slightly modified by me:
2612    
2613          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2614              not building pcregrep.
2615    
2616          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2617              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2618    
2619    25. Forward references, both numeric and by name, in patterns that made use of
2620        duplicate group numbers, could behave incorrectly or give incorrect errors,
2621        because when scanning forward to find the reference group, PCRE was not
2622        taking into account the duplicate group numbers. A pattern such as
2623        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2624    
2625    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2626        the feature of a custom pointer more persuasive (as requested by a user).
2627    
2628    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2629        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2630        pre-defined.
2631    
2632    28. Added support for (*UTF8) at the start of a pattern.
2633    
2634    29. Arrange for flags added by the "release type" setting in CMake to be shown
2635        in the configuration summary.
2636    
2637    
2638    Version 7.8 05-Sep-08
2639    ---------------------
2640    
2641    1.  Replaced UCP searching code with optimized version as implemented for Ad
2642        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2643        stage table and inline lookup instead of a function, giving speed ups of 2
2644        to 5 times on some simple patterns that I tested. Permission was given to
2645        distribute the MultiStage2.py script that generates the tables (it's not in
2646        the tarball, but is in the Subversion repository).
2647    
2648    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2649        scripts.
2650    
2651    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2652        a group with a zero qualifier. The result of the study could be incorrect,
2653        or the function might crash, depending on the pattern.
2654    
2655    4.  Caseless matching was not working for non-ASCII characters in back
2656        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2657        It now works when Unicode Property Support is available.
2658    
2659    5.  In pcretest, an escape such as \x{de} in the data was always generating
2660        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2661        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2662        truncation.
2663    
2664    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2665    
2666    7.  Added two (int) casts to pcregrep when printing the difference of two
2667        pointers, in case they are 64-bit values.
2668    
2669    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2670        test 2 if it fails.
2671    
2672    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2673        and a #define of that name to empty if it is not externally set. This is to
2674        allow users of MSVC to set it if necessary.
2675    
2676    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2677        the convenience functions in the pcre_get.c source file.
2678    
2679    11. An option change at the start of a pattern that had top-level alternatives
2680        could cause overwriting and/or a crash. This command provoked a crash in
2681        some environments:
2682    
2683          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2684    
2685        This potential security problem was recorded as CVE-2008-2371.
2686    
2687    12. For a pattern where the match had to start at the beginning or immediately
2688        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2689        pcre_dfa_exec() could read past the end of the passed subject if there was
2690        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2691        pcretest so that it places the subject at the end of its malloc-ed buffer.
2692    
2693    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2694        exec() might read past the end of the data buffer in UTF-8 mode.
2695    
2696    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2697        the data contained the byte 0x85 as part of a UTF-8 character within its
2698        first line. This applied both to normal and DFA matching.
2699    
2700    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2701        /^[^d]*?$/8 failed to match "abc".
2702    
2703    16. Added a missing copyright notice to pcrecpp_internal.h.
2704    
2705    17. Make it more clear in the documentation that values returned from
2706        pcre_exec() in ovector are byte offsets, not character counts.
2707    
2708    18. Tidied a few places to stop certain compilers from issuing warnings.
2709    
2710    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2711        supplied by Stefan Weber. I made a further small update for 7.8 because
2712        there is a change of source arrangements: the pcre_searchfuncs.c module is
2713        replaced by pcre_ucd.c.
2714    
2715    
2716    Version 7.7 07-May-08
2717    ---------------------
2718    
2719    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2720        a string to a long long, pretend we don't even have a long long." This is
2721        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2722    
2723    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2724        pre-7.6 versions, which defined a global no_arg variable instead of putting
2725        it in the RE class. (See also #8 below.)
2726    
2727    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2728        Lopes.
2729    
2730    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2731    
2732        (1) The include/exclude patterns were being applied to the whole pathnames
2733            of files, instead of just to the final components.
2734    
2735        (2) If there was more than one level of directory, the subdirectories were
2736            skipped unless they satisfied the include/exclude conditions. This is
2737            inconsistent with GNU grep (and could even be seen as contrary to the
2738            pcregrep specification - which I improved to make it absolutely clear).
2739            The action now is always to scan all levels of directory, and just
2740            apply the include/exclude patterns to regular files.
2741    
2742    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2743        --exclude_dir in the tests to avoid scanning .svn directories.
2744    
2745    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2746        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2747        doesn't support NULs in patterns.
2748    
2749    7.  Added some missing "const"s to declarations of static tables in
2750        pcre_compile.c and pcre_dfa_exec.c.
2751    
2752    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2753        caused by fix #2  above. (Subsequently also a second patch to fix the
2754        first patch. And a third patch - this was a messy problem.)
2755    
2756    9.  Applied Craig's patch to remove the use of push_back().
2757    
2758    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2759        matching function regexec().
2760    
2761    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2762        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2763        references. PCRE supports relative numbers with this syntax (I don't think
2764        Oniguruma does).
2765    
2766    12. Previously, a group with a zero repeat such as (...){0} was completely
2767        omitted from the compiled regex. However, this means that if the group
2768        was called as a subroutine from elsewhere in the pattern, things went wrong
2769        (an internal error was given). Such groups are now left in the compiled
2770        pattern, with a new opcode that causes them to be skipped at execution
2771        time.
2772    
2773    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2774        to the way PCRE behaves:
2775    
2776        (a) A lone ] character is dis-allowed (Perl treats it as data).
2777    
2778        (b) A back reference to an unmatched subpattern matches an empty string
2779            (Perl fails the current match path).
2780    
2781        (c) A data ] in a character class must be notated as \] because if the
2782            first data character in a class is ], it defines an empty class. (In
2783            Perl it is not possible to have an empty class.) The empty class []
2784            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2785            The negative empty class [^] matches any one character, independently
2786            of the DOTALL setting.
2787    
2788    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2789        non-existent subpattern following a character class starting with ']' and
2790        containing () gave an internal compiling error instead of "reference to
2791        non-existent subpattern". Fortunately, when the pattern did exist, the
2792        compiled code was correct. (When scanning forwards to check for the
2793        existence of the subpattern, it was treating the data ']' as terminating
2794        the class, so got the count wrong. When actually compiling, the reference
2795        was subsequently set up correctly.)
2796    
2797    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2798        it was being rejected as not supported by pcre_dfa_exec(), even though
2799        other assertions are supported. I have made pcre_dfa_exec() support
2800        (*FAIL).
2801    
2802    16. The implementation of 13c above involved the invention of a new opcode,
2803        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2804        cannot be changed at match time, I realized I could make a small
2805        improvement to matching performance by compiling OP_ALLANY instead of
2806        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2807        on the OP_ANY path.
2808    
2809    17. Compiling pcretest on Windows with readline support failed without the
2810        following two fixes: (1) Make the unistd.h include conditional on
2811        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2812    
2813    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2814        ncurses library to be included for pcretest when ReadLine support is
2815        requested, but also to allow for it to be overridden. This patch came from
2816        Daniel Bergström.
2817    
2818    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2819        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2820        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2821        spotting this.
2822    
2823    
2824    Version 7.6 28-Jan-08
2825    ---------------------
2826    
2827    1.  A character class containing a very large number of characters with
2828        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2829        overflow.
2830    
2831    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2832        HAVE_LONG_LONG is not defined.
2833    
2834    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2835        bring them up to date and include new features. This patch includes:
2836    
2837        - Fixed PH's badly added libz and libbz2 support.
2838        - Fixed a problem with static linking.
2839        - Added pcredemo. [But later removed - see 7 below.]
2840        - Fixed dftables problem and added an option.
2841        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2842            HAVE_LONG_LONG.
2843        - Added readline support for pcretest.
2844        - Added an listing of the option settings after cmake has run.
2845    
2846    4.  A user submitted a patch to Makefile that makes it easy to create
2847        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2848        Makefile.am that cause it to include this special target, without
2849        affecting anything else. Note that the same mingw target plus all
2850        the other distribution libraries and programs are now supported
2851        when configuring with CMake (see 6 below) instead of with
2852        Configure/Make.
2853    
2854    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2855        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2856        exported in the Windows port". It has not yet been confirmed that the patch
2857        solves the problem, but it does no harm.
2858    
2859    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2860        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2861        with CMake, and also correct the comment about stack recursion.
2862    
2863    7.  Remove the automatic building of pcredemo from the ./configure system and
2864        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2865        of a program that users should build themselves after PCRE is installed, so
2866        building it automatically is not really right. What is more, it gave
2867        trouble in some build environments.
2868    
2869    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2870    
2871    
2872    Version 7.5 10-Jan-08
2873    ---------------------
2874    
2875    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2876        values in parens when parsing an RE using the C++ wrapper."
2877    
2878    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2879        Characters greater than 255 were excluded from the class instead of being
2880        included.
2881    
2882    3.  The same bug as (2) above applied to negated POSIX classes such as
2883        [:^space:].
2884    
2885    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2886        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2887        I have changed it.
2888    
2889    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2890        first named subpattern) and a construct such as (?&a) would reference the
2891        first named subpattern whose name started with "a" (in other words, the
2892        length check was missing). Both these problems are fixed. "Subpattern name
2893        expected" is now given for (?&) (a zero-length name), and this patch also
2894        makes it give the same error for \k'' (previously it complained that that
2895        was a reference to a non-existent subpattern).
2896    
2897    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2898        this is right because (?- can be followed by option settings as well as by
2899        digits. I have, however, made the messages clearer.
2900    
2901    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2902        than the number used in the conditional) now cause a compile-time error.
2903        This is actually not compatible with Perl, which accepts such patterns, but
2904        treats the conditional as always being FALSE (as PCRE used to), but it
2905        seems to me that giving a diagnostic is better.
2906    
2907    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2908        and messages.
2909    
2910    9.  Fix two occurrences of "backslash" in comments that should have been
2911        "backspace".
2912    
2913    10. Remove two redundant lines of code that can never be obeyed (their function
2914        was moved elsewhere).
2915    
2916    11. The program that makes PCRE's Unicode character property table had a bug
2917        which caused it to generate incorrect table entries for sequences of
2918        characters that have the same character type, but are in different scripts.
2919        It amalgamated them into a single range, with the script of the first of
2920        them. In other words, some characters were in the wrong script. There were
2921        thirteen such cases, affecting characters in the following ranges:
2922    
2923          U+002b0 - U+002c1
2924          U+0060c - U+0060d
2925          U+0061e - U+00612
2926          U+0064b - U+0065e
2927          U+0074d - U+0076d
2928          U+01800 - U+01805
2929          U+01d00 - U+01d77
2930          U+01d9b - U+01dbf
2931          U+0200b - U+0200f
2932          U+030fc - U+030fe
2933          U+03260 - U+0327f
2934          U+0fb46 - U+0fbb1
2935          U+10450 - U+1049d
2936    
2937    12. The -o option (show only the matching part of a line) for pcregrep was not
2938        compatible with GNU grep in that, if there was more than one match in a
2939        line, it showed only the first of them. It now behaves in the same way as
2940        GNU grep.
2941    
2942    13. If the -o and -v options were combined for pcregrep, it printed a blank
2943        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2944        does the same. The return code can be used to tell if there were any
2945        non-matching lines.
2946    
2947    14. Added --file-offsets and --line-offsets to pcregrep.
2948    
2949    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2950        infinitely looping recursion. The bug was that positive lookaheads were not
2951        being skipped when checking for a possible empty match (negative lookaheads
2952        and both kinds of lookbehind were skipped).
2953    
2954    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2955        inclusion of <windows.h> to before rather than after the definition of
2956        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2957    
2958    17. Specifying a possessive quantifier with a specific limit for a Unicode
2959        character property caused pcre_compile() to compile bad code, which led at
2960        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2961        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2962        caused the error; without that there was no problem.
2963    
2964    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2965    
2966    19. Added --enable-pcretest-libreadline.
2967    
2968    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2969        RE::GlobalReplace(). As a result, the number of replacements returned was
2970        double what it should be. I removed one of the increments, but Craig sent a
2971        later patch that removed the other one (the right fix) and added unit tests
2972        that check the return values (which was not done before).
2973    
2974    21. Several CMake things:
2975    
2976        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2977            the names libpcre and libpcreposix, not just pcre and pcreposix.
2978    
2979        (2) The above change means that pcretest and pcregrep are now correctly
2980            linked with the newly-built libraries, not previously installed ones.
2981    
2982        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2983    
2984    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2985        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2986        UTF-8 newline character). The key issue is that the pattern starts .*;
2987        this means that the match must be either at the beginning, or after a
2988        newline. The bug was in the code for advancing after a failed match and
2989        checking that the new position followed a newline. It was not taking
2990        account of UTF-8 characters correctly.
2991    
2992    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2993        character classes. PCRE was not treating the sequence [:...:] as a
2994        character class unless the ... were all letters. Perl, however, seems to
2995        allow any characters between [: and :], though of course it rejects as
2996        unknown any "names" that contain non-letters, because all the known class
2997        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2998        for example, whereas PCRE did not - it did not recognize a POSIX character
2999        class. This seemed a bit dangerous, so the code has been changed to be
3000        closer to Perl. The behaviour is not identical to Perl, because PCRE will
3001        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3002        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3003        Perl does, and where it didn't before.
3004    
3005    24. Rewrite so as to remove the single use of %n from pcregrep because in some
3006        Windows environments %n is disabled by default.
3007    
3008    
3009    Version 7.4 21-Sep-07
3010    ---------------------
3011    
3012    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
3013        means that a class such as [\s] counted as "explicit reference to CR or
3014        LF". That isn't really right - the whole point of the change was to try to
3015        help when there was an actual mention of one of the two characters. So now
3016        the change happens only if \r or \n (or a literal CR or LF) character is
3017        encountered.
3018    
3019    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
3020        of both had grown to the point where there were only 3 bits left.
3021        Fortunately, there was spare space in the data structure, and so I have
3022        moved the internal flags into a new 16-bit field to free up more option
3023        bits.
3024    
3025    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3026        but did not set the internal JCHANGED flag - either of these is enough to
3027        control the way the "get" function works - but the PCRE_INFO_JCHANGED
3028        facility is supposed to tell if (?J) was ever used, so now (?J) at the
3029        start sets both bits.
3030    
3031    4.  Added options (at build time, compile time, exec time) to change \R from
3032        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3033    
3034    5.  doc/pcresyntax.html was missing from the distribution.
3035    
3036    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3037        compatibility, even though it is no longer used.
3038    
3039    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3040        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3041        windows.h file is present (where different names are used). [This was
3042        reversed later after testing - see 16 below.]
3043    
3044    8.  Changed all #include <config.h> to #include "config.h". There were also
3045        some further <pcre.h> cases that I changed to "pcre.h".
3046    
3047    9.  When pcregrep was used with the --colour option, it missed the line ending
3048        sequence off the lines that it output.
3049    
3050    10. It was pointed out to me that arrays of string pointers cause lots of
3051        relocations when a shared library is dynamically loaded. A technique of
3052        using a single long string with a table of offsets can drastically reduce
3053        these. I have refactored PCRE in four places to do this. The result is
3054        dramatic:
3055    
3056          Originally:                          290
3057          After changing UCP table:            187
3058          After changing error message table:   43
3059          After changing table of "verbs"       36
3060          After changing table of Posix names   22
3061    
3062        Thanks to the folks working on Gregex for glib for this insight.
3063    
3064    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3065        unicode-properties was also set.
3066    
3067    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3068    
3069    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3070        checked only for CRLF.
3071    
3072    14. Added casts to pcretest.c to avoid compiler warnings.
3073    
3074    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3075    
3076    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3077        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3078        entirely. This removes changes made in 7 above.
3079    
3080    17. The CMake files have been updated, and there is now more information about
3081        building with CMake in the NON-UNIX-USE document.
3082    
3083    
3084    Version 7.3 28-Aug-07
3085    ---------------------
3086    
3087     1. In the rejigging of the build system that eventually resulted in 7.1, the
3088        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3089        brackets there is not right, since it causes compilers to look for an
3090        installed pcre.h, not the version that is in the source that is being
3091        compiled (which of course may be different). I have changed it back to:
3092    
3093          #include "pcre.h"
3094    
3095        I have a vague recollection that the change was concerned with compiling in
3096        different directories, but in the new build system, that is taken care of
3097        by the VPATH setting the Makefile.
3098    
3099     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3100        when the subject happened to end in the byte 0x85 (e.g. if the last
3101        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3102        characters but of course it shouldn't be taken as a newline when it is part
3103        of another character. The bug was that, for an unlimited repeat of . in
3104        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3105        characters when looking for a newline.
3106    
3107     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3108    
3109     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3110        in debug output.
3111    
3112     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3113        long printing in the pcrecpp unittest when running under MinGW.
3114    
3115     6. ESC_K was left out of the EBCDIC table.
3116    
3117     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3118        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3119        limit also applies to "virtual nesting" when a pattern is recursive, and in
3120        this case 1000 isn't so big. I have been able to remove this limit at the
3121        expense of backing off one optimization in certain circumstances. Normally,
3122        when pcre_exec() would call its internal match() function recursively and
3123        immediately return the result unconditionally, it uses a "tail recursion"
3124        feature to save stack. However, when a subpattern that can match an empty
3125        string has an unlimited repetition quantifier, it no longer makes this
3126        optimization. That gives it a stack frame in which to save the data for
3127        checking that an empty string has been matched. Previously this was taken
3128        from the 1000-entry workspace that had been reserved. So now there is no
3129        explicit limit, but more stack is used.
3130    
3131     8. Applied Daniel's patches to solve problems with the import/export magic
3132        syntax that is required for Windows, and which was going wrong for the
3133        pcreposix and pcrecpp parts of the library. These were overlooked when this
3134        problem was solved for the main library.
3135    
3136     9. There were some crude static tests to avoid integer overflow when computing
3137        the size of patterns that contain repeated groups with explicit upper
3138        limits. As the maximum quantifier is 65535, the maximum group length was
3139        set at 30,000 so that the product of these two numbers did not overflow a
3140        32-bit integer. However, it turns out that people want to use groups that
3141        are longer than 30,000 bytes (though not repeat them that many times).
3142        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3143        made it possible to implement the integer overflow checks in a much more
3144        dynamic way, which I have now done. The artificial limitation on group
3145        length has been removed - we now have only the limit on the total length of
3146        the compiled pattern, which depends on the LINK_SIZE setting.
3147    
3148    10. Fixed a bug in the documentation for get/copy named substring when
3149        duplicate names are permitted. If none of the named substrings are set, the
3150        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3151        empty string.
3152    
3153    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3154        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3155        because the ] is interpreted as the first data character and the
3156        terminating ] is not found. PCRE has been made compatible with Perl in this
3157        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3158        cause memory overwriting.
3159    
3160    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3161        string has been matched (to stop an infinite loop). It was not recognizing
3162        a conditional subpattern that could match an empty string if that
3163        subpattern was within another subpattern. For example, it looped when
3164        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3165        condition was not nested. This bug has been fixed.
3166    
3167    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3168        past the start of the subject in the presence of bytes with the top bit
3169        set, for example "\x8aBCD".
3170    
3171    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3172        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3173    
3174    14. Optimized (?!) to (*FAIL).
3175    
3176    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3177        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3178        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3179        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3180        does: it's just the validity check that is more restrictive.
3181    
3182    16. Inserted checks for integer overflows during escape sequence (backslash)
3183        processing, and also fixed erroneous offset values for syntax errors during
3184        backslash processing.
3185    
3186    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3187        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3188    
3189    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3190        caused an overrun.
3191    
3192    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3193        something other than just ASCII characters) inside a group that had an
3194        unlimited repeat caused a loop at compile time (while checking to see
3195        whether the group could match an empty string).
3196    
3197    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3198        [\P{Any}] did so. (Error in the code for printing property names.)
3199    
3200    21. An orphan \E inside a character class could cause a crash.
3201    
3202    22. A repeated capturing bracket such as (A)? could cause a wild memory
3203        reference during compilation.
3204    
3205    23. There are several functions in pcre_compile() that scan along a compiled
3206        expression for various reasons (e.g. to see if it's fixed length for look
3207        behind). There were bugs in these functions when a repeated \p or \P was
3208        present in the pattern. These operators have additional parameters compared
3209        with \d, etc, and these were not being taken into account when moving along
3210        the compiled data. Specifically:
3211    
3212        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3213            length.
3214    
3215        (b) An item such as \pL+ within a repeated group could cause crashes or
3216            loops.
3217    
3218        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3219            "reference to non-existent subpattern" error.
3220    
3221        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3222    
3223    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3224        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3225    
3226    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3227    
3228    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3229        character were causing crashes (broken optimization).
3230    
3231    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3232        \p or \P) caused a compile-time loop.
3233    
3234    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3235        break. For example, the unstudied pattern [\r\n]A does not match the string
3236        "\r\nA" because change 7.0/46 below moves the current point on by two
3237        characters after failing to match at the start. However, the pattern \nA
3238        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3239        the same is true. There doesn't seem any very clean way out of this, but
3240        what I have chosen to do makes the common cases work: PCRE now takes note
3241        of whether there can be an explicit match for \r or \n anywhere in the
3242        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3243        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3244        pattern has explicit CR or LF references.
3245    
3246    29. Added (*CR) etc for changing newline setting at start of pattern.
3247    
3248    
3249    Version 7.2 19-Jun-07
3250  ---------------------  ---------------------
3251    
3252   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3266  Version 7.2 01-May-07
3266      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3267      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3268      in all cases.      in all cases.
3269    
3270   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3271    
3272      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3273          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3274    
3275      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3276          to be opened parentheses.          to be opened parentheses.
3277    
3278      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3279          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3280    
3281        (d) \K resets the start of the current match so that everything before
3282            is not part of it.
3283    
3284        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3285    
3286        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3287            reference syntax.
3288    
3289        (g) (?| introduces a group in which the numbering of parentheses in each
3290            alternative starts with the same number.
3291    
3292        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3293    
3294     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3295        PCRE_INFO_JCHANGED.
3296    
3297     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3298        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3299        for detecting groups that can match an empty string.
3300    
3301     9. A pattern with a very large number of alternatives (more than several
3302        hundred) was running out of internal workspace during the pre-compile
3303        phase, where pcre_compile() figures out how much memory will be needed. A
3304        bit of new cunning has reduced the workspace needed for groups with
3305        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3306        workspace instead of running out of the 4096 that are available.
3307    
3308    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3309    
3310    11. Applied patch from Google to remove an optimization that didn't quite work.
3311        The report of the bug said:
3312    
3313          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3314          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3315          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3316    
3317    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3318        it matched the wrong number of bytes.
3319    
3320    
3321  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3322  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1659

  ViewVC Help
Powered by ViewVC 1.1.5