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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1623 by zherczeg, Thu Feb 4 17:13:10 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 xx-xxxxxx-201x
8    ---------------------------
9    
10    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12        misbehaved. This bug was found by the LLVM fuzzer.
13    
14    2.  Similar to the above, if an isolated \E was present between an item and its
15        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16        bug was found by the LLVM fuzzer.
17    
18    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19        working correctly in UCP mode.
20    
21    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22        was set when the pmatch argument was NULL. It now returns REG_INVARG.
23    
24    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25    
26    6.  An empty \Q\E sequence between an item and its qualifier caused
27        pcre_compile() to misbehave when auto callouts were enabled. This bug was
28        found by the LLVM fuzzer.
29    
30    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
31        space or a #-type comment that was followed by (?-x), which turns off
32        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33        pcre_compile() assumed that (?-x) applied to the whole pattern and
34        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
35    
36    8.  An call of pcre_copy_named_substring() for a named substring whose number
37        was greater than the space in the ovector could cause a crash.
38    
39    9.  Yet another buffer overflow bug involved duplicate named groups with a
40        group that reset capture numbers (compare 8.38/7 below). Once again, I have
41        just allowed for more memory, even if not needed. (A proper fix is
42        implemented in PCRE2, but it involves a lot of refactoring.)
43    
44    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45        start of the match to be earlier than the end.
46    
47    11. JIT code generator for assertion matching is refactored to a separate
48        function.
49    
50    
51    Version 8.38 23-November-2015
52    -----------------------------
53    
54    1.  If a group that contained a recursive back reference also contained a
55        forward reference subroutine call followed by a non-forward-reference
56        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
57        compile correct code, leading to undefined behaviour or an internally
58        detected error. This bug was discovered by the LLVM fuzzer.
59    
60    2.  Quantification of certain items (e.g. atomic back references) could cause
61        incorrect code to be compiled when recursive forward references were
62        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
63        This bug was discovered by the LLVM fuzzer.
64    
65    3.  A repeated conditional group whose condition was a reference by name caused
66        a buffer overflow if there was more than one group with the given name.
67        This bug was discovered by the LLVM fuzzer.
68    
69    4.  A recursive back reference by name within a group that had the same name as
70        another group caused a buffer overflow. For example:
71        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
72    
73    5.  A forward reference by name to a group whose number is the same as the
74        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
75        a buffer overflow at compile time. This bug was discovered by the LLVM
76        fuzzer.
77    
78    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
79        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
80    
81    7.  Another buffer overflow bug involved duplicate named groups with a
82        reference between their definition, with a group that reset capture
83        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
84        fixed by always allowing for more memory, even if not needed. (A proper fix
85        is implemented in PCRE2, but it involves more refactoring.)
86    
87    8.  There was no check for integer overflow in subroutine calls such as (?123).
88    
89    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
90        being treated as a literal 'l' instead of causing an error.
91    
92    10. There was a buffer overflow if pcre_exec() was called with an ovector of
93        size 1. This bug was found by american fuzzy lop.
94    
95    11. If a non-capturing group containing a conditional group that could match
96        an empty string was repeated, it was not identified as matching an empty
97        string itself. For example: /^(?:(?(1)x|)+)+$()/.
98    
99    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
100        \a and \e in test subject lines.
101    
102    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
103        instead of the EBCDIC value.
104    
105    14. The handling of \c in an EBCDIC environment has been revised so that it is
106        now compatible with the specification in Perl's perlebcdic page.
107    
108    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
109        ASCII/Unicode. This has now been added to the list of characters that are
110        recognized as white space in EBCDIC.
111    
112    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
113        error (correctly) when used outside a class, but did not give an error
114        within a class.
115    
116    17. \h within a class was incorrectly compiled in EBCDIC environments.
117    
118    18. A pattern with an unmatched closing parenthesis that contained a backward
119        assertion which itself contained a forward reference caused buffer
120        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
121    
122    19. JIT should return with error when the compiled pattern requires more stack
123        space than the maximum.
124    
125    20. A possessively repeated conditional group that could match an empty string,
126        for example, /(?(R))*+/, was incorrectly compiled.
127    
128    21. Fix infinite recursion in the JIT compiler when certain patterns such as
129        /(?:|a|){100}x/ are analysed.
130    
131    22. Some patterns with character classes involving [: and \\ were incorrectly
132        compiled and could cause reading from uninitialized memory or an incorrect
133        error diagnosis.
134    
135    23. Pathological patterns containing many nested occurrences of [: caused
136        pcre_compile() to run for a very long time.
137    
138    24. A conditional group with only one branch has an implicit empty alternative
139        branch and must therefore be treated as potentially matching an empty
140        string.
141    
142    25. If (?R was followed by - or + incorrect behaviour happened instead of a
143        diagnostic.
144    
145    26. Arrange to give up on finding the minimum matching length for overly
146        complex patterns.
147    
148    27. Similar to (4) above: in a pattern with duplicated named groups and an
149        occurrence of (?| it is possible for an apparently non-recursive back
150        reference to become recursive if a later named group with the relevant
151        number is encountered. This could lead to a buffer overflow. Wen Guanxing
152        from Venustech ADLAB discovered this bug.
153    
154    28. If pcregrep was given the -q option with -c or -l, or when handling a
155        binary file, it incorrectly wrote output to stdout.
156    
157    29. The JIT compiler did not restore the control verb head in case of *THEN
158        control verbs. This issue was found by Karl Skomski with a custom LLVM
159        fuzzer.
160    
161    30. Error messages for syntax errors following \g and \k were giving inaccurate
162        offsets in the pattern.
163    
164    31. Added a check for integer overflow in conditions (?(<digits>) and
165        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
166        fuzzer.
167    
168    32. Handling recursive references such as (?2) when the reference is to a group
169        later in the pattern uses code that is very hacked about and error-prone.
170        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
171        give an internal error if it is obvious that compiling has gone wrong.
172    
173    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
174        This issue was found by Karl Skomski with a custom LLVM fuzzer.
175    
176    34. The JIT compiler should restore the control chain for empty possessive
177        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
178    
179    35. Match limit check added to JIT recursion. This issue was found by Karl
180        Skomski with a custom LLVM fuzzer.
181    
182    36. Yet another case similar to 27 above has been circumvented by an
183        unconditional allocation of extra memory. This issue is fixed "properly" in
184        PCRE2 by refactoring the way references are handled. Wen Guanxing
185        from Venustech ADLAB discovered this bug.
186    
187    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
188        with a custom LLVM fuzzer.
189    
190    38. Fixed a corner case of range optimization in JIT.
191    
192    39. An incorrect error "overran compiling workspace" was given if there were
193        exactly enough group forward references such that the last one extended
194        into the workspace safety margin. The next one would have expanded the
195        workspace. The test for overflow was not including the safety margin.
196    
197    40. A match limit issue is fixed in JIT which was found by Karl Skomski
198        with a custom LLVM fuzzer.
199    
200    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
201        work under Windows. (Why has it taken so long for anyone to notice?)
202    
203    42. In a character class such as [\W\p{Any}] where both a negative-type escape
204        ("not a word character") and a property escape were present, the property
205        escape was being ignored.
206    
207    43. Fix crash caused by very long (*MARK) or (*THEN) names.
208    
209    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
210        by a single ASCII character in a class item, was incorrectly compiled in
211        UCP mode. The POSIX class got lost, but only if the single character
212        followed it.
213    
214    45. [:punct:] in UCP mode was matching some characters in the range 128-255
215        that should not have been matched.
216    
217    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
218        class, all characters with code points greater than 255 are in the class.
219        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
220        such as \w are turned into Unicode properties), wide characters were not
221        correctly handled, and could fail to match.
222    
223    
224    Version 8.37 28-April-2015
225    --------------------------
226    
227    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
228        for those parentheses to be closed with whatever has been captured so far.
229        However, it was failing to mark any other groups between the hightest
230        capture so far and the currrent group as "unset". Thus, the ovector for
231        those groups contained whatever was previously there. An example is the
232        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
233    
234    2.  If an assertion condition was quantified with a minimum of zero (an odd
235        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
236    
237    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
238        unrecognized modifier, a crash could occur.
239    
240    4.  An attempt to do global matching in pcretest with a zero-length ovector
241        caused a crash.
242    
243    5.  Fixed a memory leak during matching that could occur for a subpattern
244        subroutine call (recursive or otherwise) if the number of captured groups
245        that had to be saved was greater than ten.
246    
247    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
248        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
249        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
250    
251    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
252        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
253    
254    8. When a pattern is compiled, it remembers the highest back reference so that
255       when matching, if the ovector is too small, extra memory can be obtained to
256       use instead. A conditional subpattern whose condition is a check on a
257       capture having happened, such as, for example in the pattern
258       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
259       setting the highest backreference number. This mattered only if pcre_exec()
260       was called with an ovector that was too small to hold the capture, and there
261       was no other kind of back reference (a situation which is probably quite
262       rare). The effect of the bug was that the condition was always treated as
263       FALSE when the capture could not be consulted, leading to a incorrect
264       behaviour by pcre_exec(). This bug has been fixed.
265    
266    9. A reference to a duplicated named group (either a back reference or a test
267       for being set in a conditional) that occurred in a part of the pattern where
268       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
269       to be incorrectly calculated, leading to overwriting.
270    
271    10. A mutually recursive set of back references such as (\2)(\1) caused a
272        segfault at study time (while trying to find the minimum matching length).
273        The infinite loop is now broken (with the minimum length unset, that is,
274        zero).
275    
276    11. If an assertion that was used as a condition was quantified with a minimum
277        of zero, matching went wrong. In particular, if the whole group had
278        unlimited repetition and could match an empty string, a segfault was
279        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
280        assertions to be quantified, but not if they are being used as conditions,
281        so the above pattern is faulted by Perl. PCRE has now been changed so that
282        it also rejects such patterns.
283    
284    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
285        failed to allow the zero-repeat case if pcre2_exec() was called with an
286        ovector too small to capture the group.
287    
288    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
289        Red Hat Product Security:
290    
291        (a) A crash if /K and /F were both set with the option to save the compiled
292        pattern.
293    
294        (b) Another crash if the option to print captured substrings in a callout
295        was combined with setting a null ovector, for example \O\C+ as a subject
296        string.
297    
298    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
299        forward reference repeated a large (but limited) number of times within a
300        repeated outer group that has a zero minimum quantifier, caused incorrect
301        code to be compiled, leading to the error "internal error:
302        previously-checked referenced subpattern not found" when an incorrect
303        memory address was read. This bug was reported as "heap overflow",
304        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
305        CVE-2015-2325.
306    
307    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
308        call within a group that also contained a recursive back reference caused
309        incorrect code to be compiled. This bug was reported as "heap overflow",
310        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
311        number CVE-2015-2326.
312    
313    24. Computing the size of the JIT read-only data in advance has been a source
314        of various issues, and new ones are still appear unfortunately. To fix
315        existing and future issues, size computation is eliminated from the code,
316        and replaced by on-demand memory allocation.
317    
318    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
319        adjacent to the end of the range, and the range contained characters with
320        more than one other case, caused incorrect behaviour when compiled in UTF
321        mode. In that example, the range a-j was left out of the class.
322    
323    26. Fix JIT compilation of conditional blocks, which assertion
324        is converted to (*FAIL). E.g: /(?(?!))/.
325    
326    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
327        discovered by the LLVM fuzzer.
328    
329    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
330        when this assertion was used as a condition, for example (?(?!)a|b). In
331        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
332        error about an unsupported item.
333    
334    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
335        possessification code could take exponential time to complete. A recursion
336        depth limit of 1000 has been imposed to limit the resources used by this
337        optimization.
338    
339    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
340        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
341        because \S ensures they are all in the class. The code for doing this was
342        interacting badly with the code for computing the amount of space needed to
343        compile the pattern, leading to a buffer overflow. This bug was discovered
344        by the LLVM fuzzer.
345    
346    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
347        other kinds of group caused stack overflow at compile time. This bug was
348        discovered by the LLVM fuzzer.
349    
350    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
351        between a subroutine call and its quantifier was incorrectly compiled,
352        leading to buffer overflow or other errors. This bug was discovered by the
353        LLVM fuzzer.
354    
355    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
356        assertion after (?(. The code was failing to check the character after
357        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
358        was discovered by the LLVM fuzzer.
359    
360    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
361        a fixed maximum following a group that contains a subroutine reference was
362        incorrectly compiled and could trigger buffer overflow. This bug was
363        discovered by the LLVM fuzzer.
364    
365    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
366        caused a stack overflow instead of the diagnosis of a non-fixed length
367        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
368    
369    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
370        (e.g. /(?<=\Ka)/) could make pcregrep loop.
371    
372    37. There was a similar problem to 36 in pcretest for global matches.
373    
374    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
375        and a subsequent item in the pattern caused a non-match, backtracking over
376        the repeated \X did not stop, but carried on past the start of the subject,
377        causing reference to random memory and/or a segfault. There were also some
378        other cases where backtracking after \C could crash. This set of bugs was
379        discovered by the LLVM fuzzer.
380    
381    39. The function for finding the minimum length of a matching string could take
382        a very long time if mutual recursion was present many times in a pattern,
383        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
384        method has been implemented. This infelicity was discovered by the LLVM
385        fuzzer.
386    
387    40. Static linking against the PCRE library using the pkg-config module was
388        failing on missing pthread symbols.
389    
390    
391    Version 8.36 26-September-2014
392    ------------------------------
393    
394    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
395        -Wmissing-field-initializers and -Wunused-parameter.
396    
397    2.  The tests for quantifiers being too big (greater than 65535) were being
398        applied after reading the number, and stupidly assuming that integer
399        overflow would give a negative number. The tests are now applied as the
400        numbers are read.
401    
402    3.  Tidy code in pcre_exec.c where two branches that used to be different are
403        now the same.
404    
405    4.  The JIT compiler did not generate match limit checks for certain
406        bracketed expressions with quantifiers. This may lead to exponential
407        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
408        issue should be resolved now.
409    
410    5.  Fixed an issue, which occures when nested alternatives are optimized
411        with table jumps.
412    
413    6.  Inserted two casts and changed some ints to size_t in the light of some
414        reported 64-bit compiler warnings (Bugzilla 1477).
415    
416    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
417        an empty string, which sometimes were behaving incorrectly in the
418        interpreter (though correctly in the JIT matcher). This pcretest input is
419        an example:
420    
421          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
422          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
423    
424        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
425        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
426        for an empty string was breaking the inner loop and carrying on at a lower
427        level, when possessive repeated groups should always return to a higher
428        level as they have no backtrack points in them. The empty string test now
429        occurs at the outer level.
430    
431    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
432        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
433    
434    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
435        doesn't).
436    
437    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
438        to hold a pattern to be incorrectly computed (too small) when there were
439        named back references to duplicated names. This could cause "internal
440        error: code overflow" or "double free or corruption" or other memory
441        handling errors.
442    
443    11. When named subpatterns had the same prefixes, back references could be
444        confused. For example, in this pattern:
445    
446          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
447    
448        the reference to 'Name' was incorrectly treated as a reference to a
449        duplicate name.
450    
451    12. A pattern such as /^s?c/mi8 where the optional character has more than
452        one "other case" was incorrectly compiled such that it would only try to
453        match starting at "c".
454    
455    13. When a pattern starting with \s was studied, VT was not included in the
456        list of possible starting characters; this should have been part of the
457        8.34/18 patch.
458    
459    14. If a character class started [\Qx]... where x is any character, the class
460        was incorrectly terminated at the ].
461    
462    15. If a pattern that started with a caseless match for a character with more
463        than one "other case" was studied, PCRE did not set up the starting code
464        unit bit map for the list of possible characters. Now it does. This is an
465        optimization improvement, not a bug fix.
466    
467    16. The Unicode data tables have been updated to Unicode 7.0.0.
468    
469    17. Fixed a number of memory leaks in pcregrep.
470    
471    18. Avoid a compiler warning (from some compilers) for a function call with
472        a cast that removes "const" from an lvalue by using an intermediate
473        variable (to which the compiler does not object).
474    
475    19. Incorrect code was compiled if a group that contained an internal recursive
476        back reference was optional (had quantifier with a minimum of zero). This
477        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
478        caused segmentation faults because of stack overflows at compile time.
479    
480    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
481        group that is quantified with an indefinite repeat, caused a compile-time
482        loop which used up all the system stack and provoked a segmentation fault.
483        This was not the same bug as 19 above.
484    
485    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
486        Patch by Mike Frysinger.
487    
488    
489    Version 8.35 04-April-2014
490    --------------------------
491    
492    1.  A new flag is set, when property checks are present in an XCLASS.
493        When this flag is not set, PCRE can perform certain optimizations
494        such as studying these XCLASS-es.
495    
496    2.  The auto-possessification of character sets were improved: a normal
497        and an extended character set can be compared now. Furthermore
498        the JIT compiler optimizes more character set checks.
499    
500    3.  Got rid of some compiler warnings for potentially uninitialized variables
501        that show up only when compiled with -O2.
502    
503    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
504        of a match later then the end of the match. The pcretest program was not
505        handling the case sensibly - it was outputting from the start to the next
506        binary zero. It now reports this situation in a message, and outputs the
507        text from the end to the start.
508    
509    5.  Fast forward search is improved in JIT. Instead of the first three
510        characters, any three characters with fixed position can be searched.
511        Search order: first, last, middle.
512    
513    6.  Improve character range checks in JIT. Characters are read by an inprecise
514        function now, which returns with an unknown value if the character code is
515        above a certain threshold (e.g: 256). The only limitation is that the value
516        must be bigger than the threshold as well. This function is useful when
517        the characters above the threshold are handled in the same way.
518    
519    7.  The macros whose names start with RAWUCHAR are placeholders for a future
520        mode in which only the bottom 21 bits of 32-bit data items are used. To
521        make this more memorable for those maintaining the code, the names have
522        been changed to start with UCHAR21, and an extensive comment has been added
523        to their definition.
524    
525    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
526        to the export list in Makefile.am (they were accidentally omitted from the
527        8.34 tarball).
528    
529    9.  The informational output from pcretest used the phrase "starting byte set"
530        which is inappropriate for the 16-bit and 32-bit libraries. As the output
531        for "first char" and "need char" really means "non-UTF-char", I've changed
532        "byte" to "char", and slightly reworded the output. The documentation about
533        these values has also been (I hope) clarified.
534    
535    10. Another JIT related optimization: use table jumps for selecting the correct
536        backtracking path, when more than four alternatives are present inside a
537        bracket.
538    
539    11. Empty match is not possible, when the minimum length is greater than zero,
540        and there is no \K in the pattern. JIT should avoid empty match checks in
541        such cases.
542    
543    12. In a caseless character class with UCP support, when a character with more
544        than one alternative case was not the first character of a range, not all
545        the alternative cases were added to the class. For example, s and \x{17f}
546        are both alternative cases for S: the class [RST] was handled correctly,
547        but [R-T] was not.
548    
549    13. The configure.ac file always checked for pthread support when JIT was
550        enabled. This is not used in Windows, so I have put this test inside a
551        check for the presence of windows.h (which was already tested for).
552    
553    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
554        The algorithm provides a way to skip certain starting offsets, and usually
555        faster than linear prefix searches.
556    
557    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
558        as for 'fr_FR' and 'french'. For some reason, however, it then used the
559        Windows-specific input and output files, which have 'french' screwed in.
560        So this could never have worked. One of the problems with locales is that
561        they aren't always the same. I have now updated RunTest so that it checks
562        the output of the locale test (test 3) against three different output
563        files, and it allows the test to pass if any one of them matches. With luck
564        this should make the test pass on some versions of Solaris where it was
565        failing. Because of the uncertainty, the script did not used to stop if
566        test 3 failed; it now does. If further versions of a French locale ever
567        come to light, they can now easily be added.
568    
569    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
570        there was a message during ./configure, but it did not stop. This now
571        provokes an error. The invalid example in README has been corrected.
572        If a value less than the minimum is given, the minimum value has always
573        been used, but now a warning is given.
574    
575    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
576        was a bug in the test system, which is now fixed. Also, the list of various
577        configurations that are tested for each release did not have one with both
578        16/32 bits and --enable-bar-anycrlf. It now does.
579    
580    18. pcretest was missing "-C bsr" for displaying the \R default setting.
581    
582    19. Little endian PowerPC systems are supported now by the JIT compiler.
583    
584    20. The fast forward newline mechanism could enter to an infinite loop on
585        certain invalid UTF-8 input. Although we don't support these cases
586        this issue can be fixed by a performance optimization.
587    
588    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
589        not take account if existing stack usage. There is now a new global
590        variable called pcre_stack_guard that can be set to point to an external
591        function to check stack availability. It is called at the start of
592        processing every parenthesized group.
593    
594    22. A typo in the code meant that in ungreedy mode the max/min qualifier
595        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
596        not match "ab".
597    
598    23. When UTF was disabled, the JIT program reported some incorrect compile
599        errors. These messages are silenced now.
600    
601    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
602        compiler.
603    
604    25. Change all the temporary files used in RunGrepTest to be different to those
605        used by RunTest so that the tests can be run simultaneously, for example by
606        "make -j check".
607    
608    
609    Version 8.34 15-December-2013
610    -----------------------------
611    
612    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
613        executable memory. Patch inspired by Carsten Klein.
614    
615    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
616        this macro is never tested and has no effect, because the work to support
617        coverage involves only compiling and linking options and special targets in
618        the Makefile. The comment in config.h implied that defining the macro would
619        enable coverage support, which is totally false. There was also support for
620        setting this macro in the CMake files (my fault, I just copied it from
621        configure). SUPPORT_GCOV has now been removed.
622    
623    3.  Make a small performance improvement in strlen16() and strlen32() in
624        pcretest.
625    
626    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
627        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
628        these cases). There was also one in pcretest.c.
629    
630    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
631    
632    6.  In UTF mode, the code for checking whether a group could match an empty
633        string (which is used for indefinitely repeated groups to allow for
634        breaking an infinite loop) was broken when the group contained a repeated
635        negated single-character class with a character that occupied more than one
636        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
637        UTF-8 mode). The effect was undefined: the group might or might not be
638        deemed as matching an empty string, or the program might have crashed.
639    
640    7.  The code for checking whether a group could match an empty string was not
641        recognizing that \h, \H, \v, \V, and \R must match a character.
642    
643    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
644        an empty string. If it can, pcretest shows this in its information output.
645    
646    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
647        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
648        matched by pcre_exec() without using JIT:
649    
650        (a) If the rest of the pattern did not match after a maximal run of
651            grapheme clusters, the code for backing up to try with fewer of them
652            did not always back up over a full grapheme when characters that do not
653            have the modifier quality were involved, e.g. Hangul syllables.
654    
655        (b) If the match point in a subject started with modifier character, and
656            there was no match, the code could incorrectly back up beyond the match
657            point, and potentially beyond the first character in the subject,
658            leading to a segfault or an incorrect match result.
659    
660    10. A conditional group with an assertion condition could lead to PCRE
661        recording an incorrect first data item for a match if no other first data
662        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
663        first data item, and therefore matched "ca" after "c" instead of at the
664        start.
665    
666    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
667        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
668    
669    12. The source of pcregrep now includes z/OS-specific code so that it can be
670        compiled for z/OS as part of the special z/OS distribution.
671    
672    13. Added the -T and -TM options to pcretest.
673    
674    14. The code in pcre_compile.c for creating the table of named capturing groups
675        has been refactored. Instead of creating the table dynamically during the
676        actual compiling pass, the information is remembered during the pre-compile
677        pass (on the stack unless there are more than 20 named groups, in which
678        case malloc() is used) and the whole table is created before the actual
679        compile happens. This has simplified the code (it is now nearly 150 lines
680        shorter) and prepared the way for better handling of references to groups
681        with duplicate names.
682    
683    15. A back reference to a named subpattern when there is more than one of the
684        same name now checks them in the order in which they appear in the pattern.
685        The first one that is set is used for the reference. Previously only the
686        first one was inspected. This change makes PCRE more compatible with Perl.
687    
688    16. Unicode character properties were updated from Unicode 6.3.0.
689    
690    17. The compile-time code for auto-possessification has been refactored, based
691        on a patch by Zoltan Herczeg. It now happens after instead of during
692        compilation. The code is cleaner, and more cases are handled. The option
693        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
694        options in pcretest are provided to set it. It can also be set by
695        (*NO_AUTO_POSSESS) at the start of a pattern.
696    
697    18. The character VT has been added to the default ("C" locale) set of
698        characters that match \s and are generally treated as white space,
699        following this same change in Perl 5.18. There is now no difference between
700        "Perl space" and "POSIX space". Whether VT is treated as white space in
701        other locales depends on the locale.
702    
703    19. The code for checking named groups as conditions, either for being set or
704        for being recursed, has been refactored (this is related to 14 and 15
705        above). Processing unduplicated named groups should now be as fast at
706        numerical groups, and processing duplicated groups should be faster than
707        before.
708    
709    20. Two patches to the CMake build system, by Alexander Barkov:
710    
711          (1) Replace the "source" command by "." in CMakeLists.txt because
712              "source" is a bash-ism.
713    
714          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
715              without these the CMake build does not work on Solaris.
716    
717    21. Perl has changed its handling of \8 and \9. If there is no previously
718        encountered capturing group of those numbers, they are treated as the
719        literal characters 8 and 9 instead of a binary zero followed by the
720        literals. PCRE now does the same.
721    
722    22. Following Perl, added \o{} to specify codepoints in octal, making it
723        possible to specify values greater than 0777 and also making them
724        unambiguous.
725    
726    23. Perl now gives an error for missing closing braces after \x{... instead of
727        treating the string as literal. PCRE now does the same.
728    
729    24. RunTest used to grumble if an inappropriate test was selected explicitly,
730        but just skip it when running all tests. This make it awkward to run ranges
731        of tests when one of them was inappropriate. Now it just skips any
732        inappropriate tests, as it always did when running all tests.
733    
734    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
735        character types such as \d or \w, too many callouts were inserted, and the
736        data that they returned was rubbish.
737    
738    26. In UCP mode, \s was not matching two of the characters that Perl matches,
739        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
740        were matched by \h. The code has now been refactored so that the lists of
741        the horizontal and vertical whitespace characters used for \h and \v (which
742        are defined only in one place) are now also used for \s.
743    
744    27. Add JIT support for the 64 bit TileGX architecture.
745        Patch by Jiong Wang (Tilera Corporation).
746    
747    28. Possessive quantifiers for classes (both explicit and automatically
748        generated) now use special opcodes instead of wrapping in ONCE brackets.
749    
750    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
751        (because it's meaningless), this was not happening when PCRE_CASELESS was
752        set. Not wrong, but inefficient.
753    
754    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
755        properties for \w, \d, etc) is present in a test regex. Otherwise if the
756        test contains no characters greater than 255, Perl doesn't realise it
757        should be using Unicode semantics.
758    
759    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
760        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
761        does in Unicode mode.
762    
763    32. Added the "forbid" facility to pcretest so that putting tests into the
764        wrong test files can sometimes be quickly detected.
765    
766    33. There is now a limit (default 250) on the depth of nesting of parentheses.
767        This limit is imposed to control the amount of system stack used at compile
768        time. It can be changed at build time by --with-parens-nest-limit=xxx or
769        the equivalent in CMake.
770    
771    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
772        errors. Perl warns for these when in warning mode, but PCRE has no facility
773        for giving warnings.
774    
775    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
776        However, this was not working for (?!) because it is optimized to (*FAIL),
777        for which PCRE does not allow quantifiers. The optimization is now disabled
778        when a quantifier follows (?!). I can't see any use for this, but it makes
779        things uniform.
780    
781    36. Perl no longer allows group names to start with digits, so I have made this
782        change also in PCRE. It simplifies the code a bit.
783    
784    37. In extended mode, Perl ignores spaces before a + that indicates a
785        possessive quantifier. PCRE allowed a space before the quantifier, but not
786        before the possessive +. It now does.
787    
788    38. The use of \K (reset reported match start) within a repeated possessive
789        group such as (a\Kb)*+ was not working.
790    
791    40. Document that the same character tables must be used at compile time and
792        run time, and that the facility to pass tables to pcre_exec() and
793        pcre_dfa_exec() is for use only with saved/restored patterns.
794    
795    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
796        features for Builds with MSVC:
797    
798        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
799           builds). The .rc files can be used to set FileDescription and many other
800           attributes.
801    
802        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
803           This allows higher-level build scripts which want .pdb files to avoid
804           hard-coding the exact files needed."
805    
806    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
807        mean "start of word" and "end of word", respectively, as a transition aid.
808    
809    43. A minimizing repeat of a class containing codepoints greater than 255 in
810        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
811        compiled to use the heap for recursion.
812    
813    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
814        is configured.
815    
816    
817    Version 8.33 28-May-2013
818    ------------------------
819    
820    1.  Added 'U' to some constants that are compared to unsigned integers, to
821        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
822        variables that are added to signed variables, to ensure the result is
823        signed and can be negated.
824    
825    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
826        CMake config files.
827    
828    3.  Revise the creation of config.h.generic so that all boolean macros are
829        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
830        overriding via -D on the command line possible.
831    
832    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
833        to unsigned int is reported to make a quite noticeable speed difference in
834        a specific Windows environment. Testing on Linux did also appear to show
835        some benefit (and it is clearly not harmful). Also fixed the definition of
836        Xop which should be unsigned.
837    
838    5.  Related to (4), changing the definition of the intermediate variable cc
839        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
840        improvements.
841    
842    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
843        unnecessary spaces.
844    
845    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
846        and later.
847    
848    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
849    
850    9.  Optimizing fast_forward_start_bits in JIT.
851    
852    10. Adding support for callouts in JIT, and fixing some issues revealed
853        during this work. Namely:
854    
855        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
856    
857        (b) Minimum length was not checked before the matching is started.
858    
859    11. The value of capture_last that is passed to callouts was incorrect in some
860        cases when there was a capture on one path that was subsequently abandoned
861        after a backtrack. Also, the capture_last value is now reset after a
862        recursion, since all captures are also reset in this case.
863    
864    12. The interpreter no longer returns the "too many substrings" error in the
865        case when an overflowing capture is in a branch that is subsequently
866        abandoned after a backtrack.
867    
868    13. In the pathological case when an offset vector of size 2 is used, pcretest
869        now prints out the matched string after a yield of 0 or 1.
870    
871    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
872        Only supported by the JIT compiler at the moment.
873    
874    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
875    
876    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
877        offset of the starting point of the matching process, provided the offsets
878        vector is large enough.
879    
880    17. The \A escape now records a lookbehind value of 1, though its execution
881        does not actually inspect the previous character. This is to ensure that,
882        in partial multi-segment matching, at least one character from the old
883        segment is retained when a new segment is processed. Otherwise, if there
884        are no lookbehinds in the pattern, \A might match incorrectly at the start
885        of a new segment.
886    
887    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
888    
889    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
890        gives some modest performance improvement in 8-bit mode.
891    
892    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
893        be expressed in certain programming languages using Universal Character
894        Names.
895    
896    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
897        which points out that "non characters" are not "characters that may not
898        appear in Unicode strings" but rather "characters that are reserved for
899        internal use and have only local meaning".
900    
901    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
902        there was a conditional group that depended on an assertion, if the
903        assertion was false, the callout that immediately followed the alternation
904        in the condition was skipped when pcre_exec() was used for matching.
905    
906    23. Allow an explicit callout to be inserted before an assertion that is the
907        condition for a conditional group, for compatibility with automatic
908        callouts, which always insert a callout at this point.
909    
910    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
911        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
912    
913    25. (*PRUNE) is now supported by the JIT compiler.
914    
915    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
916    
917    27. Fix the case where there are two or more SKIPs with arguments that may be
918        ignored.
919    
920    28. (*SKIP) is now supported by the JIT compiler.
921    
922    29. (*THEN) is now supported by the JIT compiler.
923    
924    30. Update RunTest with additional test selector options.
925    
926    31. The way PCRE handles backtracking verbs has been changed in two ways.
927    
928        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
929        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
930        backtracking. In some cases this makes it more Perl-compatible, but Perl's
931        rather obscure rules do not always do the same thing.
932    
933        (2) Previously, backtracking verbs were confined within assertions. This is
934        no longer the case for positive assertions, except for (*ACCEPT). Again,
935        this sometimes improves Perl compatibility, and sometimes does not.
936    
937    32. A number of tests that were in test 2 because Perl did things differently
938        have been moved to test 1, because either Perl or PCRE has changed, and
939        these tests are now compatible.
940    
941    32. Backtracking control verbs are now handled in the same way in JIT and
942        interpreter.
943    
944    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
945        contained a forward subroutine reference caused a compile error.
946    
947    34. Auto-detect and optimize limited repetitions in JIT.
948    
949    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
950        blocking (*UTF) etc.
951    
952    36. In the interpreter, maximizing pattern repetitions for characters and
953        character types now use tail recursion, which reduces stack usage.
954    
955    37. The value of the max lookbehind was not correctly preserved if a compiled
956        and saved regex was reloaded on a host of different endianness.
957    
958    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
959        of the compiled pattern block, expand the flags field from 16 to 32 bits
960        because it was almost full.
961    
962    39. Try madvise first before posix_madvise.
963    
964    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
965        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
966        original change locked out all matches of empty strings. This has been
967        changed so that one match of an empty string per line is recognized.
968        Subsequent searches on the same line (for colouring or for --only-matching,
969        for example) do not recognize empty strings.
970    
971    41. Applied a user patch to fix a number of spelling mistakes in comments.
972    
973    42. Data lines longer than 65536 caused pcretest to crash.
974    
975    43. Clarified the data type for length and startoffset arguments for pcre_exec
976        and pcre_dfa_exec in the function-specific man pages, where they were
977        explicitly stated to be in bytes, never having been updated. I also added
978        some clarification to the pcreapi man page.
979    
980    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
981        a segmentation fault.
982    
983    
984    Version 8.32 30-November-2012
985    -----------------------------
986    
987    1.  Improved JIT compiler optimizations for first character search and single
988        character iterators.
989    
990    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
991        Patch by Daniel Richard G.
992    
993    3.  Single character iterator optimizations in the JIT compiler.
994    
995    4.  Improved JIT compiler optimizations for character ranges.
996    
997    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
998        Reported by Giuseppe D'Angelo.
999    
1000    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1001        of a line, was being set incorrectly in cases where .* appeared inside
1002        atomic brackets at the start of a pattern, or where there was a subsequent
1003        *PRUNE or *SKIP.
1004    
1005    7.  Improved instruction cache flush for POWER/PowerPC.
1006        Patch by Daniel Richard G.
1007    
1008    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1009        grep:
1010    
1011        (a) There is now no limit to the number of patterns to be matched.
1012    
1013        (b) An error is given if a pattern is too long.
1014    
1015        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1016            are now supported.
1017    
1018        (d) --exclude-from and --include-from (multiple use) have been added.
1019    
1020        (e) Exclusions and inclusions now apply to all files and directories, not
1021            just to those obtained from scanning a directory recursively.
1022    
1023        (f) Multiple uses of -f and --file-list are now supported.
1024    
1025        (g) In a Windows environment, the default for -d has been changed from
1026            "read" (the GNU grep default) to "skip", because otherwise the presence
1027            of a directory in the file list provokes an error.
1028    
1029        (h) The documentation has been revised and clarified in places.
1030    
1031    9.  Improve the matching speed of capturing brackets.
1032    
1033    10. Changed the meaning of \X so that it now matches a Unicode extended
1034        grapheme cluster.
1035    
1036    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1037        out POSIX threads when JIT support is configured.
1038    
1039    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1040    
1041    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1042        structure could go wrong in environments where size_t is not the same size
1043        as int.
1044    
1045    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1046        set.
1047    
1048    15. The EBCDIC support had decayed; later updates to the code had included
1049        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1050        general tidy up of EBCDIC-related issues, and the documentation was also
1051        not quite right. There is now a test that can be run on ASCII systems to
1052        check some of the EBCDIC-related things (but is it not a full test).
1053    
1054    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1055        in a small tidy to the code.
1056    
1057    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1058    
1059    18. If the --only-matching (-o) option in pcregrep is specified multiple
1060        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1061        substrings matched by the 1st and 2nd capturing parentheses. A separating
1062        string can be specified by --om-separator (default empty).
1063    
1064    19. Improving the first n character searches.
1065    
1066    20. Turn case lists for horizontal and vertical white space into macros so that
1067        they are defined only once.
1068    
1069    21. This set of changes together give more compatible Unicode case-folding
1070        behaviour for characters that have more than one other case when UCP
1071        support is available.
1072    
1073        (a) The Unicode property table now has offsets into a new table of sets of
1074            three or more characters that are case-equivalent. The MultiStage2.py
1075            script that generates these tables (the pcre_ucd.c file) now scans
1076            CaseFolding.txt instead of UnicodeData.txt for character case
1077            information.
1078    
1079        (b) The code for adding characters or ranges of characters to a character
1080            class has been abstracted into a generalized function that also handles
1081            case-independence. In UTF-mode with UCP support, this uses the new data
1082            to handle characters with more than one other case.
1083    
1084        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1085            whose other case is greater than 256 are now correctly matched
1086            caselessly. Previously, the high codepoint matched the low one, but not
1087            vice versa.
1088    
1089        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1090            of the new class addition function, using character lists defined as
1091            macros alongside the case definitions of 20 above.
1092    
1093        (e) Caseless back references now work with characters that have more than
1094            one other case.
1095    
1096        (f) General caseless matching of characters with more than one other case
1097            is supported.
1098    
1099    22. Unicode character properties were updated from Unicode 6.2.0
1100    
1101    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1102    
1103    24. Add support for 32-bit character strings, and UTF-32
1104    
1105    25. Major JIT compiler update (code refactoring and bugfixing).
1106        Experimental Sparc 32 support is added.
1107    
1108    26. Applied a modified version of Daniel Richard G's patch to create
1109        pcre.h.generic and config.h.generic by "make" instead of in the
1110        PrepareRelease script.
1111    
1112    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1113        pcre_compile.c when checking for a zero character.
1114    
1115    28. Introducing a native interface for JIT. Through this interface, the compiled
1116        machine code can be directly executed. The purpose of this interface is to
1117        provide fast pattern matching, so several sanity checks are not performed.
1118        However, feature tests are still performed. The new interface provides
1119        1.4x speedup compared to the old one.
1120    
1121    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1122        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1123        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1124    
1125    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1126        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1127        the "old" RFC 2279). Instead, it ended up passing a negative length to
1128        pcre_exec().
1129    
1130    31. Add support for GCC's visibility feature to hide internal functions.
1131    
1132    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1133        "unknown -C option" after outputting 0 or 1.
1134    
1135    33. There is now support for generating a code coverage report for the test
1136        suite in environments where gcc is the compiler and lcov is installed. This
1137        is mainly for the benefit of the developers.
1138    
1139    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1140        unaddressable using valgrind annotations, allowing valgrind to detect
1141        invalid memory accesses. This is mainly for the benefit of the developers.
1142    
1143    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1144    
1145    26. Give configure error if --enable-cpp but no C++ compiler found.
1146    
1147    
1148    Version 8.31 06-July-2012
1149    -------------------------
1150    
1151    1.  Fixing a wrong JIT test case and some compiler warnings.
1152    
1153    2.  Removed a bashism from the RunTest script.
1154    
1155    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1156        to unsigned type, result still unsigned" that was given by an MS compiler
1157        on encountering the code "-sizeof(xxx)".
1158    
1159    4.  Partial matching support is added to the JIT compiler.
1160    
1161    5.  Fixed several bugs concerned with partial matching of items that consist
1162        of more than one character:
1163    
1164        (a) /^(..)\1/ did not partially match "aba" because checking references was
1165            done on an "all or nothing" basis. This also applied to repeated
1166            references.
1167    
1168        (b) \R did not give a hard partial match if \r was found at the end of the
1169            subject.
1170    
1171        (c) \X did not give a hard partial match after matching one or more
1172            characters at the end of the subject.
1173    
1174        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1175            a partial match for the string "\r".
1176    
1177        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1178            a partial match for a CR character at the end of the subject string.
1179    
1180    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1181        running pcretest, the text "(JIT)" added to the output whenever JIT is
1182        actually used to run the match.
1183    
1184    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1185        or /S+[+] with a digit between 1 and 7.
1186    
1187    8.  OP_NOT now supports any UTF character not just single-byte ones.
1188    
1189    9.  (*MARK) control verb is now supported by the JIT compiler.
1190    
1191    10. The command "./RunTest list" lists the available tests without actually
1192        running any of them. (Because I keep forgetting what they all are.)
1193    
1194    11. Add PCRE_INFO_MAXLOOKBEHIND.
1195    
1196    12. Applied a (slightly modified) user-supplied patch that improves performance
1197        when the heap is used for recursion (compiled with --disable-stack-for-
1198        recursion). Instead of malloc and free for each heap frame each time a
1199        logical recursion happens, frames are retained on a chain and re-used where
1200        possible. This sometimes gives as much as 30% improvement.
1201    
1202    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1203        call.
1204    
1205    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1206    
1207    15. It is now possible to link pcretest with libedit as an alternative to
1208        libreadline.
1209    
1210    16. (*COMMIT) control verb is now supported by the JIT compiler.
1211    
1212    17. The Unicode data tables have been updated to Unicode 6.1.0.
1213    
1214    18. Added --file-list option to pcregrep.
1215    
1216    19. Added binary file support to pcregrep, including the -a, --binary-files,
1217        -I, and --text options.
1218    
1219    20. The madvise function is renamed for posix_madvise for QNX compatibility
1220        reasons. Fixed by Giuseppe D'Angelo.
1221    
1222    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1223    
1224    22. Changed the option for creating symbolic links for 16-bit man pages from
1225        -s to -sf so that re-installing does not cause issues.
1226    
1227    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1228    
1229    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1230        matches in certain environments (the workspace was not being correctly
1231        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1232        some of the workspace data at the beginning of a restart.
1233    
1234    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1235        was not doing so when it should - probably a typo introduced by SVN 528
1236        (change 8.10/14).
1237    
1238    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1239        \w+ when the character tables indicated that \x{c4} was a word character.
1240        There were several related cases, all because the tests for doing a table
1241        lookup were testing for characters less than 127 instead of 255.
1242    
1243    27. If a pattern contains capturing parentheses that are not used in a match,
1244        their slots in the ovector are set to -1. For those that are higher than
1245        any matched groups, this happens at the end of processing. In the case when
1246        there were back references that the ovector was too small to contain
1247        (causing temporary malloc'd memory to be used during matching), and the
1248        highest capturing number was not used, memory off the end of the ovector
1249        was incorrectly being set to -1. (It was using the size of the temporary
1250        memory instead of the true size.)
1251    
1252    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1253        ovector size, it uses memory at the end of the block that it has got.
1254    
1255    29. Check for an overlong MARK name and give an error at compile time. The
1256        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1257    
1258    30. JIT compiler update.
1259    
1260    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1261        Rill for the patch.
1262    
1263    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1264    
1265    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1266    
1267    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1268        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1269        was enabled.
1270    
1271    35. Improve JIT code generation for greedy plus quantifier.
1272    
1273    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1274        1 to "aa" instead of to an empty string. The bug affected repeated groups
1275        that could potentially match an empty string.
1276    
1277    37. Optimizing single character iterators in JIT.
1278    
1279    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1280        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1281        codepoints that are too big for the mode are faulted, and in a UTF mode,
1282        disallowed codepoints are also faulted.
1283    
1284    39. If PCRE was compiled with UTF support, in three places in the DFA
1285        matcher there was code that should only have been obeyed in UTF mode, but
1286        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1287        processing when bytes with values greater than 127 were present. In 16-bit
1288        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1289        both cases the values are those that cannot be the first data item in a UTF
1290        character. The three items that might have provoked this were recursions,
1291        possessively repeated groups, and atomic groups.
1292    
1293    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1294        and pcregrep, because some OS require shared objects to be explicitly
1295        passed to ld, causing the link step to fail if they are not.
1296    
1297    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1298        mode, patterns that started with \h* or \R* might be incorrectly matched.
1299    
1300    
1301    Version 8.30 04-February-2012
1302    -----------------------------
1303    
1304    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1305        name is defined in ctype.h.
1306    
1307    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1308        only in quite long subpatterns.
1309    
1310    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1311        since it was replaced by pcre_fullinfo() in February 2000.
1312    
1313    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1314        match a (*MARK), and the match failed at the start of the subject, a
1315        reference to memory before the start of the subject could occur. This bug
1316        was introduced by fix 17 of release 8.21.
1317    
1318    5.  A reference to an unset group with zero minimum repetition was giving
1319        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1320        /(another)?(\1?)test/ matched against "hello world test". This bug was
1321        introduced in release 8.13.
1322    
1323    6.  Add support for 16-bit character strings (a large amount of work involving
1324        many changes and refactorings).
1325    
1326    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1327        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1328        from a file.
1329    
1330    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1331        rounding is not applied in this particular case).
1332    
1333    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1334        if they appear, or are escaped, in patterns.
1335    
1336    10. Get rid of a number of -Wunused-but-set-variable warnings.
1337    
1338    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1339        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1340        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1341        also returns the mark "x". This bug applied to capturing parentheses,
1342        non-capturing parentheses, and atomic parentheses. It also applied to some
1343        assertions.
1344    
1345    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1346        information out of configure.ac instead of relying on pcre.h.generic, which
1347        is not stored in the repository.
1348    
1349    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1350        -lreadline.
1351    
1352    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1353    
1354    15. Applied Graycode's patch to put the top-level frame on the stack rather
1355        than the heap when not using the stack for recursion. This gives a
1356        performance improvement in many cases when recursion is not deep.
1357    
1358    16. Experimental code added to "pcretest -C" to output the stack frame size.
1359    
1360    
1361    Version 8.21 12-Dec-2011
1362    ------------------------
1363    
1364    1.  Updating the JIT compiler.
1365    
1366    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1367        are added as well.
1368    
1369    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1370        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1371        calling _pcre_jit_exec. Some extra comments are added.
1372    
1373    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1374        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1375        was introduced by change 18 for 8.20.
1376    
1377    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1378        ECMA-262 standard.
1379    
1380    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1381        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1382        This bug was probably introduced by change 9 of 8.13.
1383    
1384    7.  While fixing 6 above, I noticed that a number of other items were being
1385        incorrectly rejected as "not fixed length". This arose partly because newer
1386        opcodes had not been added to the fixed-length checking code. I have (a)
1387        corrected the bug and added tests for these items, and (b) arranged for an
1388        error to occur if an unknown opcode is encountered while checking for fixed
1389        length instead of just assuming "not fixed length". The items that were
1390        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1391        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1392        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1393    
1394    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1395        being incorrectly compiled and would have given unpredicatble results.
1396    
1397    9.  A possessively repeated subpattern with minimum repeat count greater than
1398        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1399        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1400        the first (A) could occur when it should not.
1401    
1402    10. Add a cast and remove a redundant test from the code.
1403    
1404    11. JIT should use pcre_malloc/pcre_free for allocation.
1405    
1406    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1407        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1408        is anything other than /usr, -L is still shown).
1409    
1410    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1411    
1412    14. Perl does not support \N without a following name in a [] class; PCRE now
1413        also gives an error.
1414    
1415    15. If a forward reference was repeated with an upper limit of around 2000,
1416        it caused the error "internal error: overran compiling workspace". The
1417        maximum number of forward references (including repeats) was limited by the
1418        internal workspace, and dependent on the LINK_SIZE. The code has been
1419        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1420        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1421        of around 200,000 forward references. While doing this, I also speeded up
1422        the filling in of repeated forward references.
1423    
1424    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1425        incorrectly expecting the subject to contain another "a" after the start.
1426    
1427    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1428        in the match, the SKIP should be ignored. This was not happening; instead
1429        the SKIP was being treated as NOMATCH. For patterns such as
1430        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1431        tested.
1432    
1433    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1434        now much more compatible with Perl, in particular in cases where the result
1435        is a non-match for a non-anchored pattern. For example, if
1436        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1437        "m", where previously it did not return a name. A side effect of this
1438        change is that for partial matches, the last encountered mark name is
1439        returned, as for non matches. A number of tests that were previously not
1440        Perl-compatible have been moved into the Perl-compatible test files. The
1441        refactoring has had the pleasing side effect of removing one argument from
1442        the match() function, thus reducing its stack requirements.
1443    
1444    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1445        subsequent uses of /S (without +) incorrectly behaved like /S+.
1446    
1447    21. Retrieve executable code size support for the JIT compiler and fixing
1448        some warnings.
1449    
1450    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1451        not work when the shorter character appeared right at the end of the
1452        subject string.
1453    
1454    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1455        systems.
1456    
1457    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1458        output it when the /M option is used in pcretest.
1459    
1460    25. The CheckMan script was not being included in the distribution. Also, added
1461        an explicit "perl" to run Perl scripts from the PrepareRelease script
1462        because this is reportedly needed in Windows.
1463    
1464    26. If study data was being save in a file and studying had not found a set of
1465        "starts with" bytes for the pattern, the data written to the file (though
1466        never used) was taken from uninitialized memory and so caused valgrind to
1467        complain.
1468    
1469    27. Updated RunTest.bat as provided by Sheri Pierce.
1470    
1471    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1472    
1473    29. Computation of memory usage for the table of capturing group names was
1474        giving an unnecessarily large value.
1475    
1476    
1477    Version 8.20 21-Oct-2011
1478    ------------------------
1479    
1480    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1481        a POSIX class. After further experiments with Perl, which convinced me that
1482        Perl has bugs and confusions, a closing square bracket is no longer allowed
1483        in a POSIX name. This bug also affected patterns with classes that started
1484        with full stops.
1485    
1486    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1487        captured substring, but while checking the failing first alternative,
1488        substring 1 is temporarily captured. If the output vector supplied to
1489        pcre_exec() was not big enough for this capture, the yield of the function
1490        was still zero ("insufficient space for captured substrings"). This cannot
1491        be totally fixed without adding another stack variable, which seems a lot
1492        of expense for a edge case. However, I have improved the situation in cases
1493        such as /(a)(b)x|abc/ matched against "abc", where the return code
1494        indicates that fewer than the maximum number of slots in the ovector have
1495        been set.
1496    
1497    3.  Related to (2) above: when there are more back references in a pattern than
1498        slots in the output vector, pcre_exec() uses temporary memory during
1499        matching, and copies in the captures as far as possible afterwards. It was
1500        using the entire output vector, but this conflicts with the specification
1501        that only 2/3 is used for passing back captured substrings. Now it uses
1502        only the first 2/3, for compatibility. This is, of course, another edge
1503        case.
1504    
1505    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1506        main code base, and can be used by building with --enable-jit. When this is
1507        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1508        runtime --no-jit option is given.
1509    
1510    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1511        ovector, the return from the function was zero, implying that there were
1512        other matches that did not fit. The correct "exactly full" value is now
1513        returned.
1514    
1515    6.  If a subpattern that was called recursively or as a subroutine contained
1516        (*PRUNE) or any other control that caused it to give a non-standard return,
1517        invalid errors such as "Error -26 (nested recursion at the same subject
1518        position)" or even infinite loops could occur.
1519    
1520    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1521        computing the minimum length on reaching *ACCEPT, and so ended up with the
1522        wrong value of 1 rather than 0. Further investigation indicates that
1523        computing a minimum subject length in the presence of *ACCEPT is difficult
1524        (think back references, subroutine calls), and so I have changed the code
1525        so that no minimum is registered for a pattern that contains *ACCEPT.
1526    
1527    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1528        it was not handled as intended. [But see 16 below.]
1529    
1530    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1531        Sheri Pierce.
1532    
1533    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1534        the first byte in a match must be "a".
1535    
1536    11. Change 17 for 8.13 increased the recursion depth for patterns like
1537        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1538        pattern contains any instances of (*THEN). If it does not, the old
1539        optimizations are restored. It would be nice to do this on a per-group
1540        basis, but at the moment that is not feasible.
1541    
1542    12. In some environments, the output of pcretest -C is CRLF terminated. This
1543        broke RunTest's code that checks for the link size. A single white space
1544        character after the value is now allowed for.
1545    
1546    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1547        For "fr", it uses the Windows-specific input and output files.
1548    
1549    14. If (*THEN) appeared in a group that was called recursively or as a
1550        subroutine, it did not work as intended. [But see next item.]
1551    
1552    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1553        pattern fragments (but not containing any | characters). If A and B are
1554        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1555        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1556        D. In other words, Perl considers parentheses that do not contain any |
1557        characters to be part of a surrounding alternative, whereas PCRE was
1558        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1559        differently. PCRE now behaves in the same way as Perl, except in the case
1560        of subroutine/recursion calls such as (?1) which have in any case always
1561        been different (but PCRE had them first :-).
1562    
1563    16. Related to 15 above: Perl does not treat the | in a conditional group as
1564        creating alternatives. Such a group is treated in the same way as an
1565        ordinary group without any | characters when processing (*THEN). PCRE has
1566        been changed to match Perl's behaviour.
1567    
1568    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1569        RunGrepTest script failed.
1570    
1571    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1572        inevitable for groups that contain captures, but it can lead to a lot of
1573        stack use in large patterns. The old behaviour has been restored for atomic
1574        groups that do not contain any capturing parentheses.
1575    
1576    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1577        suppress the check for a minimum subject length at run time. (If it was
1578        given to pcre_exec() or pcre_dfa_exec() it did work.)
1579    
1580    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1581        fail to work when decoding hex characters in data strings in EBCDIC
1582        environments.
1583    
1584    21. It appears that in at least one Mac OS environment, the isxdigit() function
1585        is implemented as a macro that evaluates to its argument more than once,
1586        contravening the C 90 Standard (I haven't checked a later standard). There
1587        was an instance in pcretest which caused it to go wrong when processing
1588        \x{...} escapes in subject strings. The has been rewritten to avoid using
1589        things like p++ in the argument of isxdigit().
1590    
1591    
1592    Version 8.13 16-Aug-2011
1593    ------------------------
1594    
1595    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1596    
1597    2.  Two minor typos in pcre_internal.h have been fixed.
1598    
1599    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1600        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1601        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1602    
1603    4.  There were a number of related bugs in the code for matching backrefences
1604        caselessly in UTF-8 mode when codes for the characters concerned were
1605        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1606        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1607        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1608        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1609        2-byte code at the end of the subject (it thought there wasn't enough data
1610        left).
1611    
1612    5.  Comprehensive information about what went wrong is now returned by
1613        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1614        as the output vector has at least 2 elements. The offset of the start of
1615        the failing character and a reason code are placed in the vector.
1616    
1617    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1618        now returned is for the first byte of the failing character, instead of the
1619        last byte inspected. This is an incompatible change, but I hope it is small
1620        enough not to be a problem. It makes the returned offset consistent with
1621        pcre_exec() and pcre_dfa_exec().
1622    
1623    7.  pcretest now gives a text phrase as well as the error number when
1624        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1625        failure, the offset and reason code are output.
1626    
1627    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1628        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1629        back over a single character (\n). This seems wrong (because it treated the
1630        two characters as a single entity when going forwards), conflicts with the
1631        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1632        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1633        has been changed.
1634    
1635    9.  Some internal refactoring has changed the processing so that the handling
1636        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1637        time (the PCRE_DOTALL option was changed this way some time ago: version
1638        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1639        which was always a bit of a fudge. It also means that there is one less
1640        argument for the match() function, which reduces its stack requirements
1641        slightly. This change also fixes an incompatibility with Perl: the pattern
1642        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1643    
1644    10. More internal refactoring has drastically reduced the number of recursive
1645        calls to match() for possessively repeated groups such as (abc)++ when
1646        using pcre_exec().
1647    
1648    11. While implementing 10, a number of bugs in the handling of groups were
1649        discovered and fixed:
1650    
1651        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1652        (a|)*(?1) gave a compile-time internal error.
1653        ((a|)+)+  did not notice that the outer group could match an empty string.
1654        (^a|^)+   was not marked as anchored.
1655        (.*a|.*)+ was not marked as matching at start or after a newline.
1656    
1657    12. Yet more internal refactoring has removed another argument from the match()
1658        function. Special calls to this function are now indicated by setting a
1659        value in a variable in the "match data" data block.
1660    
1661    13. Be more explicit in pcre_study() instead of relying on "default" for
1662        opcodes that mean there is no starting character; this means that when new
1663        ones are added and accidentally left out of pcre_study(), testing should
1664        pick them up.
1665    
1666    14. The -s option of pcretest has been documented for ages as being an old
1667        synonym of -m (show memory usage). I have changed it to mean "force study
1668        for every regex", that is, assume /S for every regex. This is similar to -i
1669        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1670        using it. It makes it easier to run collections of tests with and without
1671        study enabled, and thereby test pcre_study() more easily. All the standard
1672        tests are now run with and without -s (but some patterns can be marked as
1673        "never study" - see 20 below).
1674    
1675    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1676        restoration of the capturing data to the outer values was not happening
1677        correctly.
1678    
1679    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1680        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1681        pattern had matched an empty string, and so incorrectly returned a no
1682        match.
1683    
1684    17. There was optimizing code for the last branch of non-capturing parentheses,
1685        and also for the obeyed branch of a conditional subexpression, which used
1686        tail recursion to cut down on stack usage. Unfortunately, now that there is
1687        the possibility of (*THEN) occurring in these branches, tail recursion is
1688        no longer possible because the return has to be checked for (*THEN). These
1689        two optimizations have therefore been removed. [But see 8.20/11 above.]
1690    
1691    18. If a pattern containing \R was studied, it was assumed that \R always
1692        matched two bytes, thus causing the minimum subject length to be
1693        incorrectly computed because \R can also match just one byte.
1694    
1695    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1696        was incorrectly computed.
1697    
1698    20. If /S is present twice on a test pattern in pcretest input, it now
1699        *disables* studying, thereby overriding the use of -s on the command line
1700        (see 14 above). This is necessary for one or two tests to keep the output
1701        identical in both cases.
1702    
1703    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1704        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1705    
1706    22. When an atomic group that contained a capturing parenthesis was
1707        successfully matched, but the branch in which it appeared failed, the
1708        capturing was not being forgotten if a higher numbered group was later
1709        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1710        group 1 to "a", when in fact it should be unset. This applied to multi-
1711        branched capturing and non-capturing groups, repeated or not, and also to
1712        positive assertions (capturing in negative assertions does not happen
1713        in PCRE) and also to nested atomic groups.
1714    
1715    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1716        subject after a captured substring, to make it easier to tell which of a
1717        number of identical substrings has been captured.
1718    
1719    24. The way atomic groups are processed by pcre_exec() has been changed so that
1720        if they are repeated, backtracking one repetition now resets captured
1721        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1722        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1723        "aaa". Previously, it would have been "a". As part of this code
1724        refactoring, the way recursive calls are handled has also been changed.
1725    
1726    25. If an assertion condition captured any substrings, they were not passed
1727        back unless some other capturing happened later. For example, if
1728        (?(?=(a))a) was matched against "a", no capturing was returned.
1729    
1730    26. When studying a pattern that contained subroutine calls or assertions,
1731        the code for finding the minimum length of a possible match was handling
1732        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1733        group 1 called group 2 while simultaneously a separate group 2 called group
1734        1). A stack overflow occurred in this case. I have fixed this by limiting
1735        the recursion depth to 10.
1736    
1737    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1738        Fortmann. This supports explicit test numbers on the command line, and has
1739        argument validation and error reporting.
1740    
1741    28. An instance of \X with an unlimited repeat could fail if at any point the
1742        first character it looked at was a mark character.
1743    
1744    29. Some minor code refactoring concerning Unicode properties and scripts
1745        should reduce the stack requirement of match() slightly.
1746    
1747    30. Added the '=' option to pcretest to check the setting of unused capturing
1748        slots at the end of the pattern, which are documented as being -1, but are
1749        not included in the return count.
1750    
1751    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1752        compiled something random. Now it gives a compile-time error (as does
1753        Perl).
1754    
1755    32. A *MARK encountered during the processing of a positive assertion is now
1756        recorded and passed back (compatible with Perl).
1757    
1758    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1759        had alternative anchored branches, the search for a second match in a line
1760        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1761        matched the line "0102" twice. The same bug affected patterns that started
1762        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1763        twice.
1764    
1765    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1766        does, and because of capturing effects, quantifying parenthesized
1767        assertions may at times be useful. Quantifiers are now allowed for
1768        parenthesized assertions.
1769    
1770    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1771    
1772    36. \g was being checked for fancy things in a character class, when it should
1773        just be a literal "g".
1774    
1775    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1776        appearance of a nested POSIX class supersedes an apparent external class.
1777        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1778        unescaped square brackets may also appear as part of class names. For
1779        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1780        more like Perl. (But see 8.20/1 above.)
1781    
1782    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1783        was because it thought it was \N{name}, which is not supported).
1784    
1785    39. Add minix to OS list not supporting the -S option in pcretest.
1786    
1787    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1788        cannot analyze patterns in sufficient detail to catch mutual recursions
1789        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1790        subgroup is called recursively as a subpattern for a second time at the
1791        same position in the subject string. In previous releases this might have
1792        been caught by the recursion limit, or it might have run out of stack.
1793    
1794    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1795        happen only once. PCRE was, however incorrectly giving a compile time error
1796        "recursive call could loop indefinitely" because it cannot analyze the
1797        pattern in sufficient detail. The compile time test no longer happens when
1798        PCRE is compiling a conditional subpattern, but actual runaway loops are
1799        now caught at runtime (see 40 above).
1800    
1801    42. It seems that Perl allows any characters other than a closing parenthesis
1802        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1803        has been changed to be the same.
1804    
1805    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1806        as not to get warnings when autogen.sh is called. Also changed
1807        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1808    
1809    44. To help people who use pcregrep to scan files containing exceedingly long
1810        lines, the following changes have been made:
1811    
1812        (a) The default value of the buffer size parameter has been increased from
1813            8K to 20K. (The actual buffer used is three times this size.)
1814    
1815        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1816            PCRE is built.
1817    
1818        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1819            to be set at run time.
1820    
1821        (d) Numerical values in pcregrep options can be followed by K or M, for
1822            example --buffer-size=50K.
1823    
1824        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1825            given and the return code is set to 2.
1826    
1827    45. Add a pointer to the latest mark to the callout data block.
1828    
1829    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1830        partial match of an empty string instead of no match. This was specific to
1831        the use of ".".
1832    
1833    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1834        complete match instead of a partial match. This bug was dependent on both
1835        the PCRE_UTF8 and PCRE_DOTALL options being set.
1836    
1837    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1838        starting byte set, because \b was not being ignored.
1839    
1840    
1841    Version 8.12 15-Jan-2011
1842    ------------------------
1843    
1844    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1845        checks for such things as part of the documentation building process.
1846    
1847    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1848        --match-limit and --recursion-limit options (added for 8.11). In
1849        particular, this made one of the standard tests fail. (The integer value
1850        went into the wrong half of a long int.)
1851    
1852    3.  If the --colour option was given to pcregrep with -v (invert match), it
1853        did strange things, either producing crazy output, or crashing. It should,
1854        of course, ignore a request for colour when reporting lines that do not
1855        match.
1856    
1857    4.  Another pcregrep bug caused similar problems if --colour was specified with
1858        -M (multiline) and the pattern match finished with a line ending.
1859    
1860    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1861        matched in multiline mode, the following line was shown as part of the
1862        match. This seems wrong, so I have changed it.
1863    
1864    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1865        the check for further matches in the same line (so they could be coloured)
1866        to overrun the end of the current line. If another match was found, it was
1867        incorrectly shown (and then shown again when found in the next line).
1868    
1869    7.  If pcregrep was compiled under Windows, there was a reference to the
1870        function pcregrep_exit() before it was defined. I am assuming this was
1871        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1872        reported by a user. I've moved the definition above the reference.
1873    
1874    
1875    Version 8.11 10-Dec-2010
1876    ------------------------
1877    
1878    1.  (*THEN) was not working properly if there were untried alternatives prior
1879        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1880        backtracked to try for "b" instead of moving to the next alternative branch
1881        at the same level (in this case, to look for "c"). The Perl documentation
1882        is clear that when (*THEN) is backtracked onto, it goes to the "next
1883        alternative in the innermost enclosing group".
1884    
1885    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1886        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1887        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1888        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1889        (*THEN).
1890    
1891    3.  If \s appeared in a character class, it removed the VT character from
1892        the class, even if it had been included by some previous item, for example
1893        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1894        of \s, but is part of the POSIX "space" class.)
1895    
1896    4.  A partial match never returns an empty string (because you can always
1897        match an empty string at the end of the subject); however the checking for
1898        an empty string was starting at the "start of match" point. This has been
1899        changed to the "earliest inspected character" point, because the returned
1900        data for a partial match starts at this character. This means that, for
1901        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1902        (previously it gave "no match").
1903    
1904    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1905        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1906        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1907        has an implication that the given string is incomplete (because a partial
1908        match is preferred over a full match). For this reason, these items now
1909        give a partial match in this situation. [Aside: previously, the one case
1910        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1911        match rather than a full match, which was wrong by the old rules, but is
1912        now correct.]
1913    
1914    6.  There was a bug in the handling of #-introduced comments, recognized when
1915        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1916        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1917        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1918        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1919        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1920        places in pcre_compile().
1921    
1922    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1923        comments when looking ahead for named forward references to subpatterns,
1924        the only newline sequence it recognized was NL. It now handles newlines
1925        according to the set newline convention.
1926    
1927    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1928        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1929        cater for a lack of strerror(). These oversights have been fixed.
1930    
1931    9.  Added --match-limit and --recursion-limit to pcregrep.
1932    
1933    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1934    
1935    11. When the -o option was used, pcregrep was setting a return code of 1, even
1936        when matches were found, and --line-buffered was not being honoured.
1937    
1938    12. Added an optional parentheses number to the -o and --only-matching options
1939        of pcregrep.
1940    
1941    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1942        can match an empty string. The code to do it in pcretest and pcredemo
1943        needed fixing:
1944    
1945        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1946            only one byte after an empty string match just before CRLF (this case
1947            just got forgotten; "any" and "anycrlf" were OK).
1948    
1949        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1950            mode when an empty string match preceded an ASCII character followed by
1951            a non-ASCII character. (The code for advancing by one character rather
1952            than one byte was nonsense.)
1953    
1954        (c) The pcredemo.c sample program did not have any code at all to handle
1955            the cases when CRLF is a valid newline sequence.
1956    
1957    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1958        as a starting offset was within the subject string. There is now a new
1959        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1960        negative or greater than the length of the string. In order to test this,
1961        pcretest is extended to allow the setting of negative starting offsets.
1962    
1963    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1964        starting offset points to the beginning of a UTF-8 character was
1965        unnecessarily clumsy. I tidied it up.
1966    
1967    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1968        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1969    
1970    17. Nobody had reported that the --include_dir option, which was added in
1971        release 7.7 should have been called --include-dir (hyphen, not underscore)
1972        for compatibility with GNU grep. I have changed it to --include-dir, but
1973        left --include_dir as an undocumented synonym, and the same for
1974        --exclude-dir, though that is not available in GNU grep, at least as of
1975        release 2.5.4.
1976    
1977    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1978        characters from a string of bytes) have been redefined so as not to use
1979        loops, in order to improve performance in some environments. At the same
1980        time, I abstracted some of the common code into auxiliary macros to save
1981        repetition (this should not affect the compiled code).
1982    
1983    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1984        compile-time error is now given if \c is not followed by an ASCII
1985        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1986        different, and any byte value is allowed.)
1987    
1988    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1989        START_OPTIMIZE option, which is now allowed at compile time - but just
1990        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1991        to pcregrep and other applications that have no direct access to PCRE
1992        options. The new /Y option in pcretest sets this option when calling
1993        pcre_compile().
1994    
1995    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1996        back references. Groups containing recursive back references were forced to
1997        be atomic by that change, but in the case of named groups, the amount of
1998        memory required was incorrectly computed, leading to "Failed: internal
1999        error: code overflow". This has been fixed.
2000    
2001    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2002        pcretest.c, to avoid build problems in some Borland environments.
2003    
2004    
2005    Version 8.10 25-Jun-2010
2006    ------------------------
2007    
2008    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2009        THEN.
2010    
2011    2.  (*ACCEPT) was not working when inside an atomic group.
2012    
2013    3.  Inside a character class, \B is treated as a literal by default, but
2014        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2015        causes the error). The code is unchanged, but I tidied the documentation.
2016    
2017    4.  Inside a character class, PCRE always treated \R and \X as literals,
2018        whereas Perl faults them if its -w option is set. I have changed PCRE so
2019        that it faults them when PCRE_EXTRA is set.
2020    
2021    5.  Added support for \N, which always matches any character other than
2022        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2023    
2024    6.  When compiling pcregrep with newer versions of gcc which may have
2025        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2026        declared with attribute warn_unused_result" were given. Just casting the
2027        result to (void) does not stop the warnings; a more elaborate fudge is
2028        needed. I've used a macro to implement this.
2029    
2030    7.  Minor change to pcretest.c to avoid a compiler warning.
2031    
2032    8.  Added four artifical Unicode properties to help with an option to make
2033        \s etc use properties (see next item). The new properties are: Xan
2034        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2035    
2036    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2037        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2038        this option. Modified pcretest to add /W to test this facility. Added
2039        REG_UCP to make it available via the POSIX interface.
2040    
2041    10. Added --line-buffered to pcregrep.
2042    
2043    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2044        studied, and the match started with a letter with a code point greater than
2045        127 whose first byte was different to the first byte of the other case of
2046        the letter, the other case of this starting letter was not recognized
2047        (#976).
2048    
2049    12. If a pattern that was studied started with a repeated Unicode property
2050        test, for example, \p{Nd}+, there was the theoretical possibility of
2051        setting up an incorrect bitmap of starting bytes, but fortunately it could
2052        not have actually happened in practice until change 8 above was made (it
2053        added property types that matched character-matching opcodes).
2054    
2055    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2056        possible starting bytes for non-anchored patterns.
2057    
2058    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2059        \R, and also a number of cases that involve Unicode properties, both
2060        explicit and implicit when PCRE_UCP is set.
2061    
2062    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2063        input, it could crash or give wrong results if characters with values
2064        greater than 0xc0 were present in the subject string. (Detail: it assumed
2065        UTF-8 input when processing these items.)
2066    
2067    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2068        size_t is 64-bit (#991).
2069    
2070    17. Added a check for running out of memory when PCRE is compiled with
2071        --disable-stack-for-recursion (#990).
2072    
2073    18. If the last data line in a file for pcretest does not have a newline on
2074        the end, a newline was missing in the output.
2075    
2076    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2077        less than 128) in its various bitmaps. However, there is a facility for
2078        generating tables according to the current locale when PCRE is compiled. It
2079        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2080        characters, are recognized by isspace() and therefore were getting set in
2081        these tables, and indeed these tables seem to approximate to ISO 8859. This
2082        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2083        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2084        which of course cannot start UTF-8 characters. I have changed the code so
2085        that only real ASCII characters (less than 128) and the correct starting
2086        bytes for UTF-8 encodings are set for characters greater than 127 when in
2087        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2088        altogether.)
2089    
2090    20. Added the /T option to pcretest so as to be able to run tests with non-
2091        standard character tables, thus making it possible to include the tests
2092        used for 19 above in the standard set of tests.
2093    
2094    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2095        reference to a subpattern the other side of a comment that contains an
2096        opening parenthesis caused either an internal compiling error, or a
2097        reference to the wrong subpattern.
2098    
2099    
2100    Version 8.02 19-Mar-2010
2101    ------------------------
2102    
2103    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2104    
2105    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2106        configured.
2107    
2108    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2109        original author of that file, following a query about its status.
2110    
2111    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2112        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2113    
2114    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2115        quantifier applied to a forward-referencing subroutine call, could compile
2116        incorrect code or give the error "internal error: previously-checked
2117        referenced subpattern not found".
2118    
2119    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2120        variables to point to external functions. For these systems, therefore,
2121        pcre_malloc etc. are now initialized to local functions that call the
2122        relevant global functions.
2123    
2124    7.  There were two entries missing in the vectors called coptable and poptable
2125        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2126        I've fixed the data, and added a kludgy way of testing at compile time that
2127        the lengths are correct (equal to the number of opcodes).
2128    
2129    8.  Following on from 7, I added a similar kludge to check the length of the
2130        eint vector in pcreposix.c.
2131    
2132    9.  Error texts for pcre_compile() are held as one long string to avoid too
2133        much relocation at load time. To find a text, the string is searched,
2134        counting zeros. There was no check for running off the end of the string,
2135        which could happen if a new error number was added without updating the
2136        string.
2137    
2138    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2139    
2140    11. \K was not working if it appeared in an atomic group or in a group that
2141        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2142        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2143        the assertion is positive, but not if it is negative.
2144    
2145    12. Change 11 fortuitously reduced the size of the stack frame used in the
2146        "match()" function of pcre_exec.c by one pointer. Forthcoming
2147        implementation of support for (*MARK) will need an extra pointer on the
2148        stack; I have reserved it now, so that the stack frame size does not
2149        decrease.
2150    
2151    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2152        item in branch that calls a recursion is a subroutine call - as in the
2153        second branch in the above example - was incorrectly given the compile-
2154        time error "recursive call could loop indefinitely" because pcre_compile()
2155        was not correctly checking the subroutine for matching a non-empty string.
2156    
2157    14. The checks for overrunning compiling workspace could trigger after an
2158        overrun had occurred. This is a "should never occur" error, but it can be
2159        triggered by pathological patterns such as hundreds of nested parentheses.
2160        The checks now trigger 100 bytes before the end of the workspace.
2161    
2162    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2163    
2164    
2165    Version 8.01 19-Jan-2010
2166    ------------------------
2167    
2168    1.  If a pattern contained a conditional subpattern with only one branch (in
2169        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2170        computed the wrong minimum data length (which is of course zero for such
2171        subpatterns). This could cause incorrect "no match" results.
2172    
2173    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2174        the pattern is reset in the first branch, pcre_compile() failed with
2175        "internal error: code overflow at offset...". This happened only when
2176        the reset was to the original external option setting. (An optimization
2177        abstracts leading options settings into an external setting, which was the
2178        cause of this.)
2179    
2180    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2181        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2182        assertion pattern did not match (meaning that the assertion was true), it
2183        was incorrectly treated as false if the SKIP had been reached during the
2184        matching. This also applied to assertions used as conditions.
2185    
2186    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2187        assertion subpattern, including such a pattern used as a condition,
2188        unpredictable results occurred, instead of the error return
2189        PCRE_ERROR_DFA_UITEM.
2190    
2191    5.  The C++ GlobalReplace function was not working like Perl for the special
2192        situation when an empty string is matched. It now does the fancy magic
2193        stuff that is necessary.
2194    
2195    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2196        removed. (These were left over from very, very early versions of PCRE.)
2197    
2198    7.  Some cosmetic changes to the code to make life easier when compiling it
2199        as part of something else:
2200    
2201        (a) Change DEBUG to PCRE_DEBUG.
2202    
2203        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2204            called "current" as "current_branch", to prevent a collision with the
2205            Linux macro when compiled as a kernel module.
2206    
2207        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2208            prevent a collision with the Linux macro when compiled as a kernel
2209            module.
2210    
2211    8.  In pcre_compile() there are some checks for integer overflows that used to
2212        cast potentially large values to (double). This has been changed to that
2213        when building, a check for int64_t is made, and if it is found, it is used
2214        instead, thus avoiding the use of floating point arithmetic. (There is no
2215        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2216        double.
2217    
2218    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2219        2005 (difference between two addresses compared to an unsigned value).
2220    
2221    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2222        custom one, because of the following reported problem in Windows:
2223    
2224          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2225              under Win32.
2226          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2227              therefore missing the function definition.
2228          - The compiler thus generates a "C" signature for the test function.
2229          - The linker fails to find the "C" function.
2230          - PCRE fails to configure if asked to do so against libbz2.
2231    
2232    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2233        messages were output:
2234    
2235          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2236          rerunning libtoolize, to keep the correct libtool macros in-tree.
2237          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2238    
2239        I have done both of these things.
2240    
2241    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2242        most of the time, it *can* run out if it is given a pattern that contains a
2243        runaway infinite recursion. I updated the discussion in the pcrestack man
2244        page.
2245    
2246    13. Now that we have gone to the x.xx style of version numbers, the minor
2247        version may start with zero. Using 08 or 09 is a bad idea because users
2248        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2249        interpreted as invalid octal numbers. I've updated the previous comment in
2250        configure.ac, and also added a check that gives an error if 08 or 09 are
2251        used.
2252    
2253    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2254        causing partial matching to fail when the end of the subject matched \W
2255        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2256    
2257    15. There were some discrepancies between the declarations in pcre_internal.h
2258        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2259        their definitions. The declarations used "const uschar *" and the
2260        definitions used USPTR. Even though USPTR is normally defined as "const
2261        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2262        reported that: "This difference in casting confuses some C++ compilers, for
2263        example, SunCC recognizes above declarations as different functions and
2264        generates broken code for hbpcre." I have changed the declarations to use
2265        USPTR.
2266    
2267    16. GNU libtool is named differently on some systems. The autogen.sh script now
2268        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2269        (FreeBSD).
2270    
2271    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2272        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2273        comment: "Figure out how to create a longlong from a string: strtoll and
2274        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2275        instance, but it only takes 2 args instead of 3!"
2276    
2277    18. A subtle bug concerned with back references has been fixed by a change of
2278        specification, with a corresponding code fix. A pattern such as
2279        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2280        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2281        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2282        same bug. Such groups have to be quantified to be useful, or contained
2283        inside another quantified group. (If there's no repetition, the reference
2284        can never match.) The problem arises because, having left the group and
2285        moved on to the rest of the pattern, a later failure that backtracks into
2286        the group uses the captured value from the final iteration of the group
2287        rather than the correct earlier one. I have fixed this in PCRE by forcing
2288        any group that contains a reference to itself to be an atomic group; that
2289        is, there cannot be any backtracking into it once it has completed. This is
2290        similar to recursive and subroutine calls.
2291    
2292    
2293    Version 8.00 19-Oct-09
2294    ----------------------
2295    
2296    1.  The table for translating pcre_compile() error codes into POSIX error codes
2297        was out-of-date, and there was no check on the pcre_compile() error code
2298        being within the table. This could lead to an OK return being given in
2299        error.
2300    
2301    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2302        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2303        in a Windows environment.
2304    
2305    3.  The pcregrep --count option prints the count for each file even when it is
2306        zero, as does GNU grep. However, pcregrep was also printing all files when
2307        --files-with-matches was added. Now, when both options are given, it prints
2308        counts only for those files that have at least one match. (GNU grep just
2309        prints the file name in this circumstance, but including the count seems
2310        more useful - otherwise, why use --count?) Also ensured that the
2311        combination -clh just lists non-zero counts, with no names.
2312    
2313    4.  The long form of the pcregrep -F option was incorrectly implemented as
2314        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2315        but it seems right to fix it, and I didn't think it was worth preserving
2316        the old behaviour.
2317    
2318    5.  The command line items --regex=pattern and --regexp=pattern were not
2319        recognized by pcregrep, which required --regex pattern or --regexp pattern
2320        (with a space rather than an '='). The man page documented the '=' forms,
2321        which are compatible with GNU grep; these now work.
2322    
2323    6.  No libpcreposix.pc file was created for pkg-config; there was just
2324        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2325    
2326    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2327        when UCP support is not needed, by modifying the Python script that
2328        generates it from Unicode data files. This should not matter if the module
2329        is correctly used as a library, but I received one complaint about 50K of
2330        unwanted data. My guess is that the person linked everything into his
2331        program rather than using a library. Anyway, it does no harm.
2332    
2333    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2334        was a minimum greater than 1 for a wide character in a possessive
2335        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2336        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2337        character. Chaos in the form of incorrect output or a compiling loop could
2338        result.
2339    
2340    9.  The restrictions on what a pattern can contain when partial matching is
2341        requested for pcre_exec() have been removed. All patterns can now be
2342        partially matched by this function. In addition, if there are at least two
2343        slots in the offset vector, the offset of the earliest inspected character
2344        for the match and the offset of the end of the subject are set in them when
2345        PCRE_ERROR_PARTIAL is returned.
2346    
2347    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2348        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2349        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2350        and may be more useful for multi-segment matching.
2351    
2352    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2353        used to be given if ever the end of the subject was reached; now it is
2354        given only if matching could not proceed because another character was
2355        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2356        string "Z", which now yields "no match" instead of "partial match". In the
2357        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2358        final character ended with (*FAIL).
2359    
2360    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2361        if the pattern had a "must contain" character that was already found in the
2362        earlier partial match, unless partial matching was again requested. For
2363        example, with the pattern /dog.(body)?/, the "must contain" character is
2364        "g". If the first part-match was for the string "dog", restarting with
2365        "sbody" failed. This bug has been fixed.
2366    
2367    13. The string returned by pcre_dfa_exec() after a partial match has been
2368        changed so that it starts at the first inspected character rather than the
2369        first character of the match. This makes a difference only if the pattern
2370        starts with a lookbehind assertion or \b or \B (\K is not supported by
2371        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2372        matching functions compatible, and I think it's the right thing to do.
2373    
2374    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2375        so that the demonstration program is easily available in environments where
2376        PCRE has not been installed from source.
2377    
2378    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2379        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2380        library.
2381    
2382    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2383        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2384        is not the first non-POSIX option to be added. Clearly some people find
2385        these options useful.
2386    
2387    17. If a caller to the POSIX matching function regexec() passes a non-zero
2388        value for nmatch with a NULL value for pmatch, the value of
2389        nmatch is forced to zero.
2390    
2391    18. RunGrepTest did not have a test for the availability of the -u option of
2392        the diff command, as RunTest does. It now checks in the same way as
2393        RunTest, and also checks for the -b option.
2394    
2395    19. If an odd number of negated classes containing just a single character
2396        interposed, within parentheses, between a forward reference to a named
2397        subpattern and the definition of the subpattern, compilation crashed with
2398        an internal error, complaining that it could not find the referenced
2399        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2400        [The bug was that it was starting one character too far in when skipping
2401        over the character class, thus treating the ] as data rather than
2402        terminating the class. This meant it could skip too much.]
2403    
2404    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2405        /g option in pcretest when the pattern contains \K, which makes it possible
2406        to have an empty string match not at the start, even when the pattern is
2407        anchored. Updated pcretest and pcredemo to use this option.
2408    
2409    21. If the maximum number of capturing subpatterns in a recursion was greater
2410        than the maximum at the outer level, the higher number was returned, but
2411        with unset values at the outer level. The correct (outer level) value is
2412        now given.
2413    
2414    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2415        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2416        make it do so. The string so far is captured, making this feature
2417        compatible with Perl.
2418    
2419    23. The tests have been re-organized, adding tests 11 and 12, to make it
2420        possible to check the Perl 5.10 features against Perl 5.10.
2421    
2422    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2423        pattern matches a fixed length string. PCRE did not allow this; now it
2424        does. Neither allows recursion.
2425    
2426    25. I finally figured out how to implement a request to provide the minimum
2427        length of subject string that was needed in order to match a given pattern.
2428        (It was back references and recursion that I had previously got hung up
2429        on.) This code has now been added to pcre_study(); it finds a lower bound
2430        to the length of subject needed. It is not necessarily the greatest lower
2431        bound, but using it to avoid searching strings that are too short does give
2432        some useful speed-ups. The value is available to calling programs via
2433        pcre_fullinfo().
2434    
2435    26. While implementing 25, I discovered to my embarrassment that pcretest had
2436        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2437        study optimizations had never been tested with that matching function.
2438        Oops. What is worse, even when it was passed study data, there was a bug in
2439        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2440        were also very few tests of studied patterns with pcre_dfa_exec().
2441    
2442    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2443        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2444        on the other side of the coin, they are no longer allowed to have different
2445        names, because these cannot be distinguished in PCRE, and this has caused
2446        confusion. (This is a difference from Perl.)
2447    
2448    28. When duplicate subpattern names are present (necessarily with different
2449        numbers, as required by 27 above), and a test is made by name in a
2450        conditional pattern, either for a subpattern having been matched, or for
2451        recursion in such a pattern, all the associated numbered subpatterns are
2452        tested, and the overall condition is true if the condition is true for any
2453        one of them. This is the way Perl works, and is also more like the way
2454        testing by number works.
2455    
2456    
2457    Version 7.9 11-Apr-09
2458    ---------------------
2459    
2460    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2461        (pcretest), all targets were linked against these libraries. This included
2462        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2463        libraries. This caused unwanted dependencies to be created. This problem
2464        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2465        pcretest is linked with readline.
2466    
2467    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2468        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2469        moved outside it again, because FALSE and TRUE are already defined in AIX,
2470        but BOOL is not.
2471    
2472    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2473        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2474    
2475    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2476        hyphens) following file names and line numbers when outputting matching
2477        lines. This is not true; no spaces are inserted. I have also clarified the
2478        wording for the --colour (or --color) option.
2479    
2480    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2481        was not coloured; this is different to GNU grep, so I have changed it to be
2482        the same.
2483    
2484    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2485        each matching line was coloured. Now it goes on to look for further matches
2486        of any of the test patterns, which is the same behaviour as GNU grep.
2487    
2488    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2489        doesn't make sense to accept an empty string match in pcregrep, so I have
2490        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2491        seems to be how GNU grep behaves. [But see later change 40 for release
2492        8.33.]
2493    
2494    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2495        start or after a newline", because the conditional assertion was not being
2496        correctly handled. The rule now is that both the assertion and what follows
2497        in the first alternative must satisfy the test.
2498    
2499    9.  If auto-callout was enabled in a pattern with a conditional group whose
2500        condition was an assertion, PCRE could crash during matching, both with
2501        pcre_exec() and pcre_dfa_exec().
2502    
2503    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2504        used for matching.
2505    
2506    11. Unicode property support in character classes was not working for
2507        characters (bytes) greater than 127 when not in UTF-8 mode.
2508    
2509    12. Added the -M command line option to pcretest.
2510    
2511    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2512    
2513    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2514    
2515    16. Added comments and documentation about mis-use of no_arg in the C++
2516        wrapper.
2517    
2518    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2519        from Martin Jerabek that uses macro names for all relevant character and
2520        string constants.
2521    
2522    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2523        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2524        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2525        these, but not everybody uses configure.
2526    
2527    19. A conditional group that had only one branch was not being correctly
2528        recognized as an item that could match an empty string. This meant that an
2529        enclosing group might also not be so recognized, causing infinite looping
2530        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2531        with the subject "ab", where knowledge that the repeated group can match
2532        nothing is needed in order to break the loop.
2533    
2534    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2535        exec(), but without supplying a callout function, matching went wrong.
2536    
2537    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2538        leak if the size of the offset vector was greater than 30. When the vector
2539        is smaller, the saved offsets during recursion go onto a local stack
2540        vector, but for larger vectors malloc() is used. It was failing to free
2541        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2542        error, in fact).
2543    
2544    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2545        heapframe that is used only when UTF-8 support is enabled. This caused no
2546        problem, but was untidy.
2547    
2548    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2549        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2550        included within another project.
2551    
2552    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2553        slightly modified by me:
2554    
2555          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2556              not building pcregrep.
2557    
2558          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2559              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2560    
2561    25. Forward references, both numeric and by name, in patterns that made use of
2562        duplicate group numbers, could behave incorrectly or give incorrect errors,
2563        because when scanning forward to find the reference group, PCRE was not
2564        taking into account the duplicate group numbers. A pattern such as
2565        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2566    
2567    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2568        the feature of a custom pointer more persuasive (as requested by a user).
2569    
2570    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2571        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2572        pre-defined.
2573    
2574    28. Added support for (*UTF8) at the start of a pattern.
2575    
2576    29. Arrange for flags added by the "release type" setting in CMake to be shown
2577        in the configuration summary.
2578    
2579    
2580    Version 7.8 05-Sep-08
2581    ---------------------
2582    
2583    1.  Replaced UCP searching code with optimized version as implemented for Ad
2584        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2585        stage table and inline lookup instead of a function, giving speed ups of 2
2586        to 5 times on some simple patterns that I tested. Permission was given to
2587        distribute the MultiStage2.py script that generates the tables (it's not in
2588        the tarball, but is in the Subversion repository).
2589    
2590    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2591        scripts.
2592    
2593    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2594        a group with a zero qualifier. The result of the study could be incorrect,
2595        or the function might crash, depending on the pattern.
2596    
2597    4.  Caseless matching was not working for non-ASCII characters in back
2598        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2599        It now works when Unicode Property Support is available.
2600    
2601    5.  In pcretest, an escape such as \x{de} in the data was always generating
2602        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2603        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2604        truncation.
2605    
2606    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2607    
2608    7.  Added two (int) casts to pcregrep when printing the difference of two
2609        pointers, in case they are 64-bit values.
2610    
2611    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2612        test 2 if it fails.
2613    
2614    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2615        and a #define of that name to empty if it is not externally set. This is to
2616        allow users of MSVC to set it if necessary.
2617    
2618    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2619        the convenience functions in the pcre_get.c source file.
2620    
2621    11. An option change at the start of a pattern that had top-level alternatives
2622        could cause overwriting and/or a crash. This command provoked a crash in
2623        some environments:
2624    
2625          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2626    
2627        This potential security problem was recorded as CVE-2008-2371.
2628    
2629    12. For a pattern where the match had to start at the beginning or immediately
2630        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2631        pcre_dfa_exec() could read past the end of the passed subject if there was
2632        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2633        pcretest so that it places the subject at the end of its malloc-ed buffer.
2634    
2635    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2636        exec() might read past the end of the data buffer in UTF-8 mode.
2637    
2638    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2639        the data contained the byte 0x85 as part of a UTF-8 character within its
2640        first line. This applied both to normal and DFA matching.
2641    
2642    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2643        /^[^d]*?$/8 failed to match "abc".
2644    
2645    16. Added a missing copyright notice to pcrecpp_internal.h.
2646    
2647    17. Make it more clear in the documentation that values returned from
2648        pcre_exec() in ovector are byte offsets, not character counts.
2649    
2650    18. Tidied a few places to stop certain compilers from issuing warnings.
2651    
2652    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2653        supplied by Stefan Weber. I made a further small update for 7.8 because
2654        there is a change of source arrangements: the pcre_searchfuncs.c module is
2655        replaced by pcre_ucd.c.
2656    
2657    
2658    Version 7.7 07-May-08
2659    ---------------------
2660    
2661    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2662        a string to a long long, pretend we don't even have a long long." This is
2663        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2664    
2665    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2666        pre-7.6 versions, which defined a global no_arg variable instead of putting
2667        it in the RE class. (See also #8 below.)
2668    
2669    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2670        Lopes.
2671    
2672    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2673    
2674        (1) The include/exclude patterns were being applied to the whole pathnames
2675            of files, instead of just to the final components.
2676    
2677        (2) If there was more than one level of directory, the subdirectories were
2678            skipped unless they satisfied the include/exclude conditions. This is
2679            inconsistent with GNU grep (and could even be seen as contrary to the
2680            pcregrep specification - which I improved to make it absolutely clear).
2681            The action now is always to scan all levels of directory, and just
2682            apply the include/exclude patterns to regular files.
2683    
2684    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2685        --exclude_dir in the tests to avoid scanning .svn directories.
2686    
2687    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2688        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2689        doesn't support NULs in patterns.
2690    
2691    7.  Added some missing "const"s to declarations of static tables in
2692        pcre_compile.c and pcre_dfa_exec.c.
2693    
2694    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2695        caused by fix #2  above. (Subsequently also a second patch to fix the
2696        first patch. And a third patch - this was a messy problem.)
2697    
2698    9.  Applied Craig's patch to remove the use of push_back().
2699    
2700    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2701        matching function regexec().
2702    
2703    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2704        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2705        references. PCRE supports relative numbers with this syntax (I don't think
2706        Oniguruma does).
2707    
2708    12. Previously, a group with a zero repeat such as (...){0} was completely
2709        omitted from the compiled regex. However, this means that if the group
2710        was called as a subroutine from elsewhere in the pattern, things went wrong
2711        (an internal error was given). Such groups are now left in the compiled
2712        pattern, with a new opcode that causes them to be skipped at execution
2713        time.
2714    
2715    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2716        to the way PCRE behaves:
2717    
2718        (a) A lone ] character is dis-allowed (Perl treats it as data).
2719    
2720        (b) A back reference to an unmatched subpattern matches an empty string
2721            (Perl fails the current match path).
2722    
2723        (c) A data ] in a character class must be notated as \] because if the
2724            first data character in a class is ], it defines an empty class. (In
2725            Perl it is not possible to have an empty class.) The empty class []
2726            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2727            The negative empty class [^] matches any one character, independently
2728            of the DOTALL setting.
2729    
2730    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2731        non-existent subpattern following a character class starting with ']' and
2732        containing () gave an internal compiling error instead of "reference to
2733        non-existent subpattern". Fortunately, when the pattern did exist, the
2734        compiled code was correct. (When scanning forwards to check for the
2735        existence of the subpattern, it was treating the data ']' as terminating
2736        the class, so got the count wrong. When actually compiling, the reference
2737        was subsequently set up correctly.)
2738    
2739    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2740        it was being rejected as not supported by pcre_dfa_exec(), even though
2741        other assertions are supported. I have made pcre_dfa_exec() support
2742        (*FAIL).
2743    
2744    16. The implementation of 13c above involved the invention of a new opcode,
2745        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2746        cannot be changed at match time, I realized I could make a small
2747        improvement to matching performance by compiling OP_ALLANY instead of
2748        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2749        on the OP_ANY path.
2750    
2751    17. Compiling pcretest on Windows with readline support failed without the
2752        following two fixes: (1) Make the unistd.h include conditional on
2753        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2754    
2755    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2756        ncurses library to be included for pcretest when ReadLine support is
2757        requested, but also to allow for it to be overridden. This patch came from
2758        Daniel Bergström.
2759    
2760    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2761        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2762        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2763        spotting this.
2764    
2765    
2766    Version 7.6 28-Jan-08
2767    ---------------------
2768    
2769    1.  A character class containing a very large number of characters with
2770        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2771        overflow.
2772    
2773    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2774        HAVE_LONG_LONG is not defined.
2775    
2776    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2777        bring them up to date and include new features. This patch includes:
2778    
2779        - Fixed PH's badly added libz and libbz2 support.
2780        - Fixed a problem with static linking.
2781        - Added pcredemo. [But later removed - see 7 below.]
2782        - Fixed dftables problem and added an option.
2783        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2784            HAVE_LONG_LONG.
2785        - Added readline support for pcretest.
2786        - Added an listing of the option settings after cmake has run.
2787    
2788    4.  A user submitted a patch to Makefile that makes it easy to create
2789        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2790        Makefile.am that cause it to include this special target, without
2791        affecting anything else. Note that the same mingw target plus all
2792        the other distribution libraries and programs are now supported
2793        when configuring with CMake (see 6 below) instead of with
2794        Configure/Make.
2795    
2796    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2797        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2798        exported in the Windows port". It has not yet been confirmed that the patch
2799        solves the problem, but it does no harm.
2800    
2801    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2802        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2803        with CMake, and also correct the comment about stack recursion.
2804    
2805    7.  Remove the automatic building of pcredemo from the ./configure system and
2806        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2807        of a program that users should build themselves after PCRE is installed, so
2808        building it automatically is not really right. What is more, it gave
2809        trouble in some build environments.
2810    
2811    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2812    
2813    
2814    Version 7.5 10-Jan-08
2815    ---------------------
2816    
2817    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2818        values in parens when parsing an RE using the C++ wrapper."
2819    
2820    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2821        Characters greater than 255 were excluded from the class instead of being
2822        included.
2823    
2824    3.  The same bug as (2) above applied to negated POSIX classes such as
2825        [:^space:].
2826    
2827    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2828        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2829        I have changed it.
2830    
2831    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2832        first named subpattern) and a construct such as (?&a) would reference the
2833        first named subpattern whose name started with "a" (in other words, the
2834        length check was missing). Both these problems are fixed. "Subpattern name
2835        expected" is now given for (?&) (a zero-length name), and this patch also
2836        makes it give the same error for \k'' (previously it complained that that
2837        was a reference to a non-existent subpattern).
2838    
2839    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2840        this is right because (?- can be followed by option settings as well as by
2841        digits. I have, however, made the messages clearer.
2842    
2843    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2844        than the number used in the conditional) now cause a compile-time error.
2845        This is actually not compatible with Perl, which accepts such patterns, but
2846        treats the conditional as always being FALSE (as PCRE used to), but it
2847        seems to me that giving a diagnostic is better.
2848    
2849    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2850        and messages.
2851    
2852    9.  Fix two occurrences of "backslash" in comments that should have been
2853        "backspace".
2854    
2855    10. Remove two redundant lines of code that can never be obeyed (their function
2856        was moved elsewhere).
2857    
2858    11. The program that makes PCRE's Unicode character property table had a bug
2859        which caused it to generate incorrect table entries for sequences of
2860        characters that have the same character type, but are in different scripts.
2861        It amalgamated them into a single range, with the script of the first of
2862        them. In other words, some characters were in the wrong script. There were
2863        thirteen such cases, affecting characters in the following ranges:
2864    
2865          U+002b0 - U+002c1
2866          U+0060c - U+0060d
2867          U+0061e - U+00612
2868          U+0064b - U+0065e
2869          U+0074d - U+0076d
2870          U+01800 - U+01805
2871          U+01d00 - U+01d77
2872          U+01d9b - U+01dbf
2873          U+0200b - U+0200f
2874          U+030fc - U+030fe
2875          U+03260 - U+0327f
2876          U+0fb46 - U+0fbb1
2877          U+10450 - U+1049d
2878    
2879    12. The -o option (show only the matching part of a line) for pcregrep was not
2880        compatible with GNU grep in that, if there was more than one match in a
2881        line, it showed only the first of them. It now behaves in the same way as
2882        GNU grep.
2883    
2884    13. If the -o and -v options were combined for pcregrep, it printed a blank
2885        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2886        does the same. The return code can be used to tell if there were any
2887        non-matching lines.
2888    
2889    14. Added --file-offsets and --line-offsets to pcregrep.
2890    
2891    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2892        infinitely looping recursion. The bug was that positive lookaheads were not
2893        being skipped when checking for a possible empty match (negative lookaheads
2894        and both kinds of lookbehind were skipped).
2895    
2896    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2897        inclusion of <windows.h> to before rather than after the definition of
2898        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2899    
2900    17. Specifying a possessive quantifier with a specific limit for a Unicode
2901        character property caused pcre_compile() to compile bad code, which led at
2902        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2903        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2904        caused the error; without that there was no problem.
2905    
2906    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2907    
2908    19. Added --enable-pcretest-libreadline.
2909    
2910    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2911        RE::GlobalReplace(). As a result, the number of replacements returned was
2912        double what it should be. I removed one of the increments, but Craig sent a
2913        later patch that removed the other one (the right fix) and added unit tests
2914        that check the return values (which was not done before).
2915    
2916    21. Several CMake things:
2917    
2918        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2919            the names libpcre and libpcreposix, not just pcre and pcreposix.
2920    
2921        (2) The above change means that pcretest and pcregrep are now correctly
2922            linked with the newly-built libraries, not previously installed ones.
2923    
2924        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2925    
2926    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2927        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2928        UTF-8 newline character). The key issue is that the pattern starts .*;
2929        this means that the match must be either at the beginning, or after a
2930        newline. The bug was in the code for advancing after a failed match and
2931        checking that the new position followed a newline. It was not taking
2932        account of UTF-8 characters correctly.
2933    
2934    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2935        character classes. PCRE was not treating the sequence [:...:] as a
2936        character class unless the ... were all letters. Perl, however, seems to
2937        allow any characters between [: and :], though of course it rejects as
2938        unknown any "names" that contain non-letters, because all the known class
2939        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2940        for example, whereas PCRE did not - it did not recognize a POSIX character
2941        class. This seemed a bit dangerous, so the code has been changed to be
2942        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2943        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2944        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2945        Perl does, and where it didn't before.
2946    
2947    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2948        Windows environments %n is disabled by default.
2949    
2950    
2951    Version 7.4 21-Sep-07
2952    ---------------------
2953    
2954    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2955        means that a class such as [\s] counted as "explicit reference to CR or
2956        LF". That isn't really right - the whole point of the change was to try to
2957        help when there was an actual mention of one of the two characters. So now
2958        the change happens only if \r or \n (or a literal CR or LF) character is
2959        encountered.
2960    
2961    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2962        of both had grown to the point where there were only 3 bits left.
2963        Fortunately, there was spare space in the data structure, and so I have
2964        moved the internal flags into a new 16-bit field to free up more option
2965        bits.
2966    
2967    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2968        but did not set the internal JCHANGED flag - either of these is enough to
2969        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2970        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2971        start sets both bits.
2972    
2973    4.  Added options (at build time, compile time, exec time) to change \R from
2974        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2975    
2976    5.  doc/pcresyntax.html was missing from the distribution.
2977    
2978    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2979        compatibility, even though it is no longer used.
2980    
2981    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2982        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2983        windows.h file is present (where different names are used). [This was
2984        reversed later after testing - see 16 below.]
2985    
2986    8.  Changed all #include <config.h> to #include "config.h". There were also
2987        some further <pcre.h> cases that I changed to "pcre.h".
2988    
2989    9.  When pcregrep was used with the --colour option, it missed the line ending
2990        sequence off the lines that it output.
2991    
2992    10. It was pointed out to me that arrays of string pointers cause lots of
2993        relocations when a shared library is dynamically loaded. A technique of
2994        using a single long string with a table of offsets can drastically reduce
2995        these. I have refactored PCRE in four places to do this. The result is
2996        dramatic:
2997    
2998          Originally:                          290
2999          After changing UCP table:            187
3000          After changing error message table:   43
3001          After changing table of "verbs"       36
3002          After changing table of Posix names   22
3003    
3004        Thanks to the folks working on Gregex for glib for this insight.
3005    
3006    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3007        unicode-properties was also set.
3008    
3009    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3010    
3011    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3012        checked only for CRLF.
3013    
3014    14. Added casts to pcretest.c to avoid compiler warnings.
3015    
3016    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3017    
3018    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3019        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3020        entirely. This removes changes made in 7 above.
3021    
3022    17. The CMake files have been updated, and there is now more information about
3023        building with CMake in the NON-UNIX-USE document.
3024    
3025    
3026    Version 7.3 28-Aug-07
3027    ---------------------
3028    
3029     1. In the rejigging of the build system that eventually resulted in 7.1, the
3030        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3031        brackets there is not right, since it causes compilers to look for an
3032        installed pcre.h, not the version that is in the source that is being
3033        compiled (which of course may be different). I have changed it back to:
3034    
3035          #include "pcre.h"
3036    
3037        I have a vague recollection that the change was concerned with compiling in
3038        different directories, but in the new build system, that is taken care of
3039        by the VPATH setting the Makefile.
3040    
3041     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3042        when the subject happened to end in the byte 0x85 (e.g. if the last
3043        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3044        characters but of course it shouldn't be taken as a newline when it is part
3045        of another character. The bug was that, for an unlimited repeat of . in
3046        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3047        characters when looking for a newline.
3048    
3049     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3050    
3051     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3052        in debug output.
3053    
3054     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3055        long printing in the pcrecpp unittest when running under MinGW.
3056    
3057     6. ESC_K was left out of the EBCDIC table.
3058    
3059     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3060        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3061        limit also applies to "virtual nesting" when a pattern is recursive, and in
3062        this case 1000 isn't so big. I have been able to remove this limit at the
3063        expense of backing off one optimization in certain circumstances. Normally,
3064        when pcre_exec() would call its internal match() function recursively and
3065        immediately return the result unconditionally, it uses a "tail recursion"
3066        feature to save stack. However, when a subpattern that can match an empty
3067        string has an unlimited repetition quantifier, it no longer makes this
3068        optimization. That gives it a stack frame in which to save the data for
3069        checking that an empty string has been matched. Previously this was taken
3070        from the 1000-entry workspace that had been reserved. So now there is no
3071        explicit limit, but more stack is used.
3072    
3073     8. Applied Daniel's patches to solve problems with the import/export magic
3074        syntax that is required for Windows, and which was going wrong for the
3075        pcreposix and pcrecpp parts of the library. These were overlooked when this
3076        problem was solved for the main library.
3077    
3078     9. There were some crude static tests to avoid integer overflow when computing
3079        the size of patterns that contain repeated groups with explicit upper
3080        limits. As the maximum quantifier is 65535, the maximum group length was
3081        set at 30,000 so that the product of these two numbers did not overflow a
3082        32-bit integer. However, it turns out that people want to use groups that
3083        are longer than 30,000 bytes (though not repeat them that many times).
3084        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3085        made it possible to implement the integer overflow checks in a much more
3086        dynamic way, which I have now done. The artificial limitation on group
3087        length has been removed - we now have only the limit on the total length of
3088        the compiled pattern, which depends on the LINK_SIZE setting.
3089    
3090    10. Fixed a bug in the documentation for get/copy named substring when
3091        duplicate names are permitted. If none of the named substrings are set, the
3092        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3093        empty string.
3094    
3095    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3096        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3097        because the ] is interpreted as the first data character and the
3098        terminating ] is not found. PCRE has been made compatible with Perl in this
3099        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3100        cause memory overwriting.
3101    
3102    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3103        string has been matched (to stop an infinite loop). It was not recognizing
3104        a conditional subpattern that could match an empty string if that
3105        subpattern was within another subpattern. For example, it looped when
3106        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3107        condition was not nested. This bug has been fixed.
3108    
3109    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3110        past the start of the subject in the presence of bytes with the top bit
3111        set, for example "\x8aBCD".
3112    
3113    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3114        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3115    
3116    14. Optimized (?!) to (*FAIL).
3117    
3118    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3119        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3120        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3121        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3122        does: it's just the validity check that is more restrictive.
3123    
3124    16. Inserted checks for integer overflows during escape sequence (backslash)
3125        processing, and also fixed erroneous offset values for syntax errors during
3126        backslash processing.
3127    
3128    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3129        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3130    
3131    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3132        caused an overrun.
3133    
3134    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3135        something other than just ASCII characters) inside a group that had an
3136        unlimited repeat caused a loop at compile time (while checking to see
3137        whether the group could match an empty string).
3138    
3139    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3140        [\P{Any}] did so. (Error in the code for printing property names.)
3141    
3142    21. An orphan \E inside a character class could cause a crash.
3143    
3144    22. A repeated capturing bracket such as (A)? could cause a wild memory
3145        reference during compilation.
3146    
3147    23. There are several functions in pcre_compile() that scan along a compiled
3148        expression for various reasons (e.g. to see if it's fixed length for look
3149        behind). There were bugs in these functions when a repeated \p or \P was
3150        present in the pattern. These operators have additional parameters compared
3151        with \d, etc, and these were not being taken into account when moving along
3152        the compiled data. Specifically:
3153    
3154        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3155            length.
3156    
3157        (b) An item such as \pL+ within a repeated group could cause crashes or
3158            loops.
3159    
3160        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3161            "reference to non-existent subpattern" error.
3162    
3163        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3164    
3165    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3166        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3167    
3168    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3169    
3170    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3171        character were causing crashes (broken optimization).
3172    
3173    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3174        \p or \P) caused a compile-time loop.
3175    
3176    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3177        break. For example, the unstudied pattern [\r\n]A does not match the string
3178        "\r\nA" because change 7.0/46 below moves the current point on by two
3179        characters after failing to match at the start. However, the pattern \nA
3180        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3181        the same is true. There doesn't seem any very clean way out of this, but
3182        what I have chosen to do makes the common cases work: PCRE now takes note
3183        of whether there can be an explicit match for \r or \n anywhere in the
3184        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3185        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3186        pattern has explicit CR or LF references.
3187    
3188    29. Added (*CR) etc for changing newline setting at start of pattern.
3189    
3190    
3191    Version 7.2 19-Jun-07
3192    ---------------------
3193    
3194     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3195        which is apparently normally available under Windows.
3196    
3197     2. Re-jig the pcregrep tests with different newline settings in an attempt
3198        to make them independent of the local environment's newline setting.
3199    
3200     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3201    
3202     4. Some of the "internals" tests were previously cut out when the link size
3203        was not 2, because the output contained actual offsets. The recent new
3204        "Z" feature of pcretest means that these can be cut out, making the tests
3205        usable with all link sizes.
3206    
3207     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3208        stack recursion. This gives a massive performance boost under BSD, but just
3209        a small improvement under Linux. However, it saves one field in the frame
3210        in all cases.
3211    
3212     6. Added more features from the forthcoming Perl 5.10:
3213    
3214        (a) (?-n) (where n is a string of digits) is a relative subroutine or
3215            recursion call. It refers to the nth most recently opened parentheses.
3216    
3217        (b) (?+n) is also a relative subroutine call; it refers to the nth next
3218            to be opened parentheses.
3219    
3220        (c) Conditions that refer to capturing parentheses can be specified
3221            relatively, for example, (?(-2)... or (?(+3)...
3222    
3223        (d) \K resets the start of the current match so that everything before
3224            is not part of it.
3225    
3226        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3227    
3228        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3229            reference syntax.
3230    
3231        (g) (?| introduces a group in which the numbering of parentheses in each
3232            alternative starts with the same number.
3233    
3234        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3235    
3236     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3237        PCRE_INFO_JCHANGED.
3238    
3239     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3240        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3241        for detecting groups that can match an empty string.
3242    
3243     9. A pattern with a very large number of alternatives (more than several
3244        hundred) was running out of internal workspace during the pre-compile
3245        phase, where pcre_compile() figures out how much memory will be needed. A
3246        bit of new cunning has reduced the workspace needed for groups with
3247        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3248        workspace instead of running out of the 4096 that are available.
3249    
3250    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3251    
3252    11. Applied patch from Google to remove an optimization that didn't quite work.
3253        The report of the bug said:
3254    
3255          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3256          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3257          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3258    
3259    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3260        it matched the wrong number of bytes.
3261    
3262    
3263    Version 7.1 24-Apr-07
3264    ---------------------
3265    
3266     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3267        that is more "standard", making use of automake and other Autotools. There
3268        is some re-arrangement of the files and adjustment of comments consequent
3269        on this.
3270    
3271     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3272        for recursive directory scanning broke on some systems because the files
3273        are not scanned in any specific order and on different systems the order
3274        was different. A call to "sort" has been inserted into RunGrepTest for the
3275        approprate test as a short-term fix. In the longer term there may be an
3276        alternative.
3277    
3278     3. I had an email from Eric Raymond about problems translating some of PCRE's
3279        man pages to HTML (despite the fact that I distribute HTML pages, some
3280        people do their own conversions for various reasons). The problems
3281        concerned the use of low-level troff macros .br and .in. I have therefore
3282        removed all such uses from the man pages (some were redundant, some could
3283        be replaced by .nf/.fi pairs). The 132html script that I use to generate
3284        HTML has been updated to handle .nf/.fi and to complain if it encounters
3285        .br or .in.
3286    
3287     4. Updated comments in configure.ac that get placed in config.h.in and also
3288        arranged for config.h to be included in the distribution, with the name
3289        config.h.generic, for the benefit of those who have to compile without
3290        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3291    
3292     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3293        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3294        makevp.bat for the current PCRE, using the additional files
3295        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3296    
3297     6. A Windows user reported a minor discrepancy with test 2, which turned out
3298        to be caused by a trailing space on an input line that had got lost in his
3299        copy. The trailing space was an accident, so I've just removed it.
3300    
3301     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3302        that is needed.
3303    
3304     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3305        as "const" (a) because they are and (b) because it helps the PHP
3306        maintainers who have recently made a script to detect big data structures
3307        in the php code that should be moved to the .rodata section. I remembered
3308        to update Builducptable as well, so it won't revert if ucptable.h is ever
3309        re-created.
3310    
3311     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3312        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3313        order to be able to cut out the UTF-8 tables in the latter when UTF-8
3314        support is not required. This saves 1.5-2K of code, which is important in
3315        some applications.
3316    
3317        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3318        so as not to refer to the tables, even though these functions will never be
3319        called when UTF-8 support is disabled. Otherwise there are problems with a
3320        shared library.
3321    
3322    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3323    
3324        (a) It was defining its arguments as char * instead of void *.
3325    
3326        (b) It was assuming that all moves were upwards in memory; this was true
3327            a long time ago when I wrote it, but is no longer the case.
3328    
3329        The emulated memove() is provided for those environments that have neither
3330        memmove() nor bcopy(). I didn't think anyone used it these days, but that
3331        is clearly not the case, as these two bugs were recently reported.
3332    
3333    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3334        and Detrail to create the HTML documentation, the .txt form of the man
3335        pages, and it removes trailing spaces from listed files. It also creates
3336        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3337        case, it wraps all the #defines with #ifndefs. This script should be run
3338        before "make dist".
3339    
3340    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3341        with Unicode property support.
3342    
3343        (a) For a maximizing quantifier, if the two different cases of the
3344            character were of different lengths in their UTF-8 codings (there are
3345            some cases like this - I found 11), and the matching function had to
3346            back up over a mixture of the two cases, it incorrectly assumed they
3347            were both the same length.
3348    
3349        (b) When PCRE was configured to use the heap rather than the stack for
3350            recursion during matching, it was not correctly preserving the data for
3351            the other case of a UTF-8 character when checking ahead for a match
3352            while processing a minimizing repeat. If the check also involved
3353            matching a wide character, but failed, corruption could cause an
3354            erroneous result when trying to check for a repeat of the original
3355            character.
3356    
3357    13. Some tidying changes to the testing mechanism:
3358    
3359        (a) The RunTest script now detects the internal link size and whether there
3360            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3361            values substituted by "configure". (The RunGrepTest script already did
3362            this for UTF-8.) The configure.ac script no longer substitutes the
3363            relevant variables.
3364    
3365        (b) The debugging options /B and /D in pcretest show the compiled bytecode
3366            with length and offset values. This means that the output is different
3367            for different internal link sizes. Test 2 is skipped for link sizes
3368            other than 2 because of this, bypassing the problem. Unfortunately,
3369            there was also a test in test 3 (the locale tests) that used /B and
3370            failed for link sizes other than 2. Rather than cut the whole test out,
3371            I have added a new /Z option to pcretest that replaces the length and
3372            offset values with spaces. This is now used to make test 3 independent
3373            of link size. (Test 2 will be tidied up later.)
3374    
3375    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3376        segmentation fault instead of returning the appropriate error message.
3377    
3378    15. In multiline mode when the newline sequence was set to "any", the pattern
3379        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3380        This doesn't seem right; it now treats the CRLF combination as the line
3381        ending, and so does not match in that case. It's only a pattern such as ^$
3382        that would hit this one: something like ^ABC$ would have failed after \r
3383        and then tried again after \r\n.
3384    
3385    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3386        in an attempt to make files that differ only in their line terminators
3387        compare equal. This works on Linux.
3388    
3389    17. Under certain error circumstances pcregrep might try to free random memory
3390        as it exited. This is now fixed, thanks to valgrind.
3391    
3392    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3393        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3394        was because its rules for how to advance for /g after matching an empty
3395        string at the end of a line did not allow for this case. They now check for
3396        it specially.
3397    
3398    20. pcretest is supposed to handle patterns and data of any length, by
3399        extending its buffers when necessary. It was getting this wrong when the
3400        buffer for a data line had to be extended.
3401    
3402    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3403        CRLF as a newline sequence.
3404    
3405    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3406        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3407        I have nevertheless tidied it up.
3408    
3409    23. Added some casts to kill warnings from HP-UX ia64 compiler.
3410    
3411    24. Added a man page for pcre-config.
3412    
3413    
3414    Version 7.0 19-Dec-06
3415    ---------------------
3416    
3417     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3418        moving to gcc 4.1.1.
3419    
3420     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3421        sys/time.h, which is documented as needed for this function. It doesn't
3422        seem to matter on Linux, but it showed up on some releases of OS X.
3423    
3424     3. It seems that there are systems where bytes whose values are greater than
3425        127 match isprint() in the "C" locale. The "C" locale should be the
3426        default when a C program starts up. In most systems, only ASCII printing
3427        characters match isprint(). This difference caused the output from pcretest
3428        to vary, making some of the tests fail. I have changed pcretest so that:
3429    
3430        (a) When it is outputting text in the compiled version of a pattern, bytes
3431            other than 32-126 are always shown as hex escapes.
3432    
3433        (b) When it is outputting text that is a matched part of a subject string,
3434            it does the same, unless a different locale has been set for the match
3435            (using the /L modifier). In this case, it uses isprint() to decide.
3436    
3437     4. Fixed a major bug that caused incorrect computation of the amount of memory
3438        required for a compiled pattern when options that changed within the
3439        pattern affected the logic of the preliminary scan that determines the
3440        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3441        that the computed length was too small. The symptoms of this bug were
3442        either the PCRE error "internal error: code overflow" from pcre_compile(),
3443        or a glibc crash with a message such as "pcretest: free(): invalid next
3444        size (fast)". Examples of patterns that provoked this bug (shown in
3445        pcretest format) are:
3446    
3447          /(?-x: )/x
3448          /(?x)(?-x: \s*#\s*)/
3449          /((?i)[\x{c0}])/8
3450          /(?i:[\x{c0}])/8
3451    
3452        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3453        is now done differently.
3454    
3455     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3456        wrapper classes; (b) implement a new function in the C++ scanner that is
3457        more efficient than the old way of doing things because it avoids levels of
3458        recursion in the regex matching; (c) add a paragraph to the documentation
3459        for the FullMatch() function.
3460    
3461     6. The escape sequence \n was being treated as whatever was defined as
3462        "newline". Not only was this contrary to the documentation, which states
3463        that \n is character 10 (hex 0A), but it also went horribly wrong when
3464        "newline" was defined as CRLF. This has been fixed.
3465    
3466     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3467        was being set to -1 for the "end of line" case (supposedly a value that no
3468        character can have). Though this value is never used (the check for end of
3469        line is "zero bytes in current character"), it caused compiler complaints.
3470        I've changed it to 0xffffffff.
3471    
3472     8. In pcre_version.c, the version string was being built by a sequence of
3473        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3474        string (as it is for production releases) called a macro with an empty
3475        argument. The C standard says the result of this is undefined. The gcc
3476        compiler treats it as an empty string (which was what was wanted) but it is
3477        reported that Visual C gives an error. The source has been hacked around to
3478        avoid this problem.
3479    
3480     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3481        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3482        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3483        of them did).
3484    
3485    10. Originally, pcretest opened its input and output without "b"; then I was
3486        told that "b" was needed in some environments, so it was added for release
3487        5.0 to both the input and output. (It makes no difference on Unix-like
3488        systems.) Later I was told that it is wrong for the input on Windows. I've
3489        now abstracted the modes into two macros, to make it easier to fiddle with
3490        them, and removed "b" from the input mode under Windows.
3491    
3492    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3493    
3494    12. Added -help and --help to pcretest as an official way of being reminded
3495        of the options.
3496    
3497    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3498        and pcrecpp.cc because they annoy compilers at high warning levels.
3499    
3500    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3501    
3502    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3503        scripts are not C programs :-) and which was not noticed because it works
3504        on Linux.
3505    
3506    16. pcretest is supposed to handle any length of pattern and data line (as one
3507        line or as a continued sequence of lines) by extending its input buffer if
3508        necessary. This feature was broken for very long pattern lines, leading to
3509        a string of junk being passed to pcre_compile() if the pattern was longer
3510        than about 50K.
3511    
3512    17. I have done a major re-factoring of the way pcre_compile() computes the
3513        amount of memory needed for a compiled pattern. Previously, there was code
3514        that made a preliminary scan of the pattern in order to do this. That was
3515        OK when PCRE was new, but as the facilities have expanded, it has become
3516        harder and harder to keep it in step with the real compile phase, and there
3517        have been a number of bugs (see for example, 4 above). I have now found a
3518        cunning way of running the real compile function in a "fake" mode that
3519        enables it to compute how much memory it would need, while actually only
3520        ever using a few hundred bytes of working memory and without too many
3521        tests of the mode. This should make future maintenance and development
3522        easier. A side effect of this work is that the limit of 200 on the nesting
3523        depth of parentheses has been removed (though this was never a serious
3524        limitation, I suspect). However, there is a downside: pcre_compile() now
3525        runs more slowly than before (30% or more, depending on the pattern). I
3526        hope this isn't a big issue. There is no effect on runtime performance.
3527    
3528    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3529        newline (only possible for the last line of a file) and it was a
3530        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3531    
3532    19. Added additional timing features to pcretest. (1) The -tm option now times
3533        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3534        separate command line item, by a number that specifies the number of
3535        repeats to use when timing. The default is 50000; this gives better
3536        precision, but takes uncomfortably long for very large patterns.
3537    
3538    20. Extended pcre_study() to be more clever in cases where a branch of a
3539        subpattern has no definite first character. For example, (a*|b*)[cd] would
3540        previously give no result from pcre_study(). Now it recognizes that the