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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1648 by ph10, Fri May 20 10:33:36 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 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. Migrating appropriate PCRE2 JIT improvements to PCRE.
48    
49    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50        assertion, caused pcretest to generate incorrect output, and also to read
51        uninitialized memory (detected by ASAN or valgrind).
52    
53    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
54        nested set of parentheses of sufficient size caused an overflow of the
55        compiling workspace (which was diagnosed, but of course is not desirable).
56    
57    14. And yet another buffer overflow bug involving duplicate named groups, this
58        time nested, with a nested back reference. Yet again, I have just allowed
59        for more memory, because anything more needs all the refactoring that has
60        been done for PCRE2. An example pattern that provoked this bug is:
61        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
62        registered as CVE-2016-1283.
63    
64    15. pcretest went into a loop if global matching was requested with an ovector
65        size less than 2. It now gives an error message. This bug was found by
66        afl-fuzz.
67    
68    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
69        ("assertion expected") when (?(?C) was not followed by an opening
70        parenthesis.
71    
72    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
73        actually affect anything, by sheer luck.
74    
75    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
76        static compilation.
77    
78    19. Modified the RunTest script to incorporate a valgrind suppressions file so
79        that certain errors, provoked by the SSE2 instruction set when JIT is used,
80        are ignored.
81    
82    20. A racing condition is fixed in JIT reported by Mozilla.
83    
84    21. Minor code refactor to avoid "array subscript is below array bounds"
85    compiler warning.
86    
87    
88    Version 8.38 23-November-2015
89    -----------------------------
90    
91    1.  If a group that contained a recursive back reference also contained a
92        forward reference subroutine call followed by a non-forward-reference
93        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
94        compile correct code, leading to undefined behaviour or an internally
95        detected error. This bug was discovered by the LLVM fuzzer.
96    
97    2.  Quantification of certain items (e.g. atomic back references) could cause
98        incorrect code to be compiled when recursive forward references were
99        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
100        This bug was discovered by the LLVM fuzzer.
101    
102    3.  A repeated conditional group whose condition was a reference by name caused
103        a buffer overflow if there was more than one group with the given name.
104        This bug was discovered by the LLVM fuzzer.
105    
106    4.  A recursive back reference by name within a group that had the same name as
107        another group caused a buffer overflow. For example:
108        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
109    
110    5.  A forward reference by name to a group whose number is the same as the
111        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
112        a buffer overflow at compile time. This bug was discovered by the LLVM
113        fuzzer.
114    
115    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
116        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
117    
118    7.  Another buffer overflow bug involved duplicate named groups with a
119        reference between their definition, with a group that reset capture
120        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
121        fixed by always allowing for more memory, even if not needed. (A proper fix
122        is implemented in PCRE2, but it involves more refactoring.)
123    
124    8.  There was no check for integer overflow in subroutine calls such as (?123).
125    
126    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
127        being treated as a literal 'l' instead of causing an error.
128    
129    10. There was a buffer overflow if pcre_exec() was called with an ovector of
130        size 1. This bug was found by american fuzzy lop.
131    
132    11. If a non-capturing group containing a conditional group that could match
133        an empty string was repeated, it was not identified as matching an empty
134        string itself. For example: /^(?:(?(1)x|)+)+$()/.
135    
136    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
137        \a and \e in test subject lines.
138    
139    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
140        instead of the EBCDIC value.
141    
142    14. The handling of \c in an EBCDIC environment has been revised so that it is
143        now compatible with the specification in Perl's perlebcdic page.
144    
145    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
146        ASCII/Unicode. This has now been added to the list of characters that are
147        recognized as white space in EBCDIC.
148    
149    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
150        error (correctly) when used outside a class, but did not give an error
151        within a class.
152    
153    17. \h within a class was incorrectly compiled in EBCDIC environments.
154    
155    18. A pattern with an unmatched closing parenthesis that contained a backward
156        assertion which itself contained a forward reference caused buffer
157        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
158    
159    19. JIT should return with error when the compiled pattern requires more stack
160        space than the maximum.
161    
162    20. A possessively repeated conditional group that could match an empty string,
163        for example, /(?(R))*+/, was incorrectly compiled.
164    
165    21. Fix infinite recursion in the JIT compiler when certain patterns such as
166        /(?:|a|){100}x/ are analysed.
167    
168    22. Some patterns with character classes involving [: and \\ were incorrectly
169        compiled and could cause reading from uninitialized memory or an incorrect
170        error diagnosis.
171    
172    23. Pathological patterns containing many nested occurrences of [: caused
173        pcre_compile() to run for a very long time.
174    
175    24. A conditional group with only one branch has an implicit empty alternative
176        branch and must therefore be treated as potentially matching an empty
177        string.
178    
179    25. If (?R was followed by - or + incorrect behaviour happened instead of a
180        diagnostic.
181    
182    26. Arrange to give up on finding the minimum matching length for overly
183        complex patterns.
184    
185    27. Similar to (4) above: in a pattern with duplicated named groups and an
186        occurrence of (?| it is possible for an apparently non-recursive back
187        reference to become recursive if a later named group with the relevant
188        number is encountered. This could lead to a buffer overflow. Wen Guanxing
189        from Venustech ADLAB discovered this bug.
190    
191    28. If pcregrep was given the -q option with -c or -l, or when handling a
192        binary file, it incorrectly wrote output to stdout.
193    
194    29. The JIT compiler did not restore the control verb head in case of *THEN
195        control verbs. This issue was found by Karl Skomski with a custom LLVM
196        fuzzer.
197    
198    30. Error messages for syntax errors following \g and \k were giving inaccurate
199        offsets in the pattern.
200    
201    31. Added a check for integer overflow in conditions (?(<digits>) and
202        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
203        fuzzer.
204    
205    32. Handling recursive references such as (?2) when the reference is to a group
206        later in the pattern uses code that is very hacked about and error-prone.
207        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
208        give an internal error if it is obvious that compiling has gone wrong.
209    
210    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
211        This issue was found by Karl Skomski with a custom LLVM fuzzer.
212    
213    34. The JIT compiler should restore the control chain for empty possessive
214        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
215    
216    35. Match limit check added to JIT recursion. This issue was found by Karl
217        Skomski with a custom LLVM fuzzer.
218    
219    36. Yet another case similar to 27 above has been circumvented by an
220        unconditional allocation of extra memory. This issue is fixed "properly" in
221        PCRE2 by refactoring the way references are handled. Wen Guanxing
222        from Venustech ADLAB discovered this bug.
223    
224    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
225        with a custom LLVM fuzzer.
226    
227    38. Fixed a corner case of range optimization in JIT.
228    
229    39. An incorrect error "overran compiling workspace" was given if there were
230        exactly enough group forward references such that the last one extended
231        into the workspace safety margin. The next one would have expanded the
232        workspace. The test for overflow was not including the safety margin.
233    
234    40. A match limit issue is fixed in JIT which was found by Karl Skomski
235        with a custom LLVM fuzzer.
236    
237    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
238        work under Windows. (Why has it taken so long for anyone to notice?)
239    
240    42. In a character class such as [\W\p{Any}] where both a negative-type escape
241        ("not a word character") and a property escape were present, the property
242        escape was being ignored.
243    
244    43. Fix crash caused by very long (*MARK) or (*THEN) names.
245    
246    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
247        by a single ASCII character in a class item, was incorrectly compiled in
248        UCP mode. The POSIX class got lost, but only if the single character
249        followed it.
250    
251    45. [:punct:] in UCP mode was matching some characters in the range 128-255
252        that should not have been matched.
253    
254    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
255        class, all characters with code points greater than 255 are in the class.
256        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
257        such as \w are turned into Unicode properties), wide characters were not
258        correctly handled, and could fail to match.
259    
260    
261    Version 8.37 28-April-2015
262    --------------------------
263    
264    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
265        for those parentheses to be closed with whatever has been captured so far.
266        However, it was failing to mark any other groups between the hightest
267        capture so far and the currrent group as "unset". Thus, the ovector for
268        those groups contained whatever was previously there. An example is the
269        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
270    
271    2.  If an assertion condition was quantified with a minimum of zero (an odd
272        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
273    
274    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
275        unrecognized modifier, a crash could occur.
276    
277    4.  An attempt to do global matching in pcretest with a zero-length ovector
278        caused a crash.
279    
280    5.  Fixed a memory leak during matching that could occur for a subpattern
281        subroutine call (recursive or otherwise) if the number of captured groups
282        that had to be saved was greater than ten.
283    
284    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
285        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
286        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
287    
288    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
289        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
290    
291    8. When a pattern is compiled, it remembers the highest back reference so that
292       when matching, if the ovector is too small, extra memory can be obtained to
293       use instead. A conditional subpattern whose condition is a check on a
294       capture having happened, such as, for example in the pattern
295       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
296       setting the highest backreference number. This mattered only if pcre_exec()
297       was called with an ovector that was too small to hold the capture, and there
298       was no other kind of back reference (a situation which is probably quite
299       rare). The effect of the bug was that the condition was always treated as
300       FALSE when the capture could not be consulted, leading to a incorrect
301       behaviour by pcre_exec(). This bug has been fixed.
302    
303    9. A reference to a duplicated named group (either a back reference or a test
304       for being set in a conditional) that occurred in a part of the pattern where
305       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
306       to be incorrectly calculated, leading to overwriting.
307    
308    10. A mutually recursive set of back references such as (\2)(\1) caused a
309        segfault at study time (while trying to find the minimum matching length).
310        The infinite loop is now broken (with the minimum length unset, that is,
311        zero).
312    
313    11. If an assertion that was used as a condition was quantified with a minimum
314        of zero, matching went wrong. In particular, if the whole group had
315        unlimited repetition and could match an empty string, a segfault was
316        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
317        assertions to be quantified, but not if they are being used as conditions,
318        so the above pattern is faulted by Perl. PCRE has now been changed so that
319        it also rejects such patterns.
320    
321    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
322        failed to allow the zero-repeat case if pcre2_exec() was called with an
323        ovector too small to capture the group.
324    
325    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
326        Red Hat Product Security:
327    
328        (a) A crash if /K and /F were both set with the option to save the compiled
329        pattern.
330    
331        (b) Another crash if the option to print captured substrings in a callout
332        was combined with setting a null ovector, for example \O\C+ as a subject
333        string.
334    
335    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
336        forward reference repeated a large (but limited) number of times within a
337        repeated outer group that has a zero minimum quantifier, caused incorrect
338        code to be compiled, leading to the error "internal error:
339        previously-checked referenced subpattern not found" when an incorrect
340        memory address was read. This bug was reported as "heap overflow",
341        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
342        CVE-2015-2325.
343    
344    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
345        call within a group that also contained a recursive back reference caused
346        incorrect code to be compiled. This bug was reported as "heap overflow",
347        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
348        number CVE-2015-2326.
349    
350    24. Computing the size of the JIT read-only data in advance has been a source
351        of various issues, and new ones are still appear unfortunately. To fix
352        existing and future issues, size computation is eliminated from the code,
353        and replaced by on-demand memory allocation.
354    
355    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
356        adjacent to the end of the range, and the range contained characters with
357        more than one other case, caused incorrect behaviour when compiled in UTF
358        mode. In that example, the range a-j was left out of the class.
359    
360    26. Fix JIT compilation of conditional blocks, which assertion
361        is converted to (*FAIL). E.g: /(?(?!))/.
362    
363    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
364        discovered by the LLVM fuzzer.
365    
366    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
367        when this assertion was used as a condition, for example (?(?!)a|b). In
368        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
369        error about an unsupported item.
370    
371    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
372        possessification code could take exponential time to complete. A recursion
373        depth limit of 1000 has been imposed to limit the resources used by this
374        optimization.
375    
376    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
377        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
378        because \S ensures they are all in the class. The code for doing this was
379        interacting badly with the code for computing the amount of space needed to
380        compile the pattern, leading to a buffer overflow. This bug was discovered
381        by the LLVM fuzzer.
382    
383    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
384        other kinds of group caused stack overflow at compile time. This bug was
385        discovered by the LLVM fuzzer.
386    
387    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
388        between a subroutine call and its quantifier was incorrectly compiled,
389        leading to buffer overflow or other errors. This bug was discovered by the
390        LLVM fuzzer.
391    
392    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
393        assertion after (?(. The code was failing to check the character after
394        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
395        was discovered by the LLVM fuzzer.
396    
397    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
398        a fixed maximum following a group that contains a subroutine reference was
399        incorrectly compiled and could trigger buffer overflow. This bug was
400        discovered by the LLVM fuzzer.
401    
402    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
403        caused a stack overflow instead of the diagnosis of a non-fixed length
404        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
405    
406    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
407        (e.g. /(?<=\Ka)/) could make pcregrep loop.
408    
409    37. There was a similar problem to 36 in pcretest for global matches.
410    
411    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
412        and a subsequent item in the pattern caused a non-match, backtracking over
413        the repeated \X did not stop, but carried on past the start of the subject,
414        causing reference to random memory and/or a segfault. There were also some
415        other cases where backtracking after \C could crash. This set of bugs was
416        discovered by the LLVM fuzzer.
417    
418    39. The function for finding the minimum length of a matching string could take
419        a very long time if mutual recursion was present many times in a pattern,
420        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
421        method has been implemented. This infelicity was discovered by the LLVM
422        fuzzer.
423    
424    40. Static linking against the PCRE library using the pkg-config module was
425        failing on missing pthread symbols.
426    
427    
428    Version 8.36 26-September-2014
429    ------------------------------
430    
431    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
432        -Wmissing-field-initializers and -Wunused-parameter.
433    
434    2.  The tests for quantifiers being too big (greater than 65535) were being
435        applied after reading the number, and stupidly assuming that integer
436        overflow would give a negative number. The tests are now applied as the
437        numbers are read.
438    
439    3.  Tidy code in pcre_exec.c where two branches that used to be different are
440        now the same.
441    
442    4.  The JIT compiler did not generate match limit checks for certain
443        bracketed expressions with quantifiers. This may lead to exponential
444        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
445        issue should be resolved now.
446    
447    5.  Fixed an issue, which occures when nested alternatives are optimized
448        with table jumps.
449    
450    6.  Inserted two casts and changed some ints to size_t in the light of some
451        reported 64-bit compiler warnings (Bugzilla 1477).
452    
453    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
454        an empty string, which sometimes were behaving incorrectly in the
455        interpreter (though correctly in the JIT matcher). This pcretest input is
456        an example:
457    
458          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
459          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
460    
461        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
462        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
463        for an empty string was breaking the inner loop and carrying on at a lower
464        level, when possessive repeated groups should always return to a higher
465        level as they have no backtrack points in them. The empty string test now
466        occurs at the outer level.
467    
468    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
469        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
470    
471    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
472        doesn't).
473    
474    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
475        to hold a pattern to be incorrectly computed (too small) when there were
476        named back references to duplicated names. This could cause "internal
477        error: code overflow" or "double free or corruption" or other memory
478        handling errors.
479    
480    11. When named subpatterns had the same prefixes, back references could be
481        confused. For example, in this pattern:
482    
483          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
484    
485        the reference to 'Name' was incorrectly treated as a reference to a
486        duplicate name.
487    
488    12. A pattern such as /^s?c/mi8 where the optional character has more than
489        one "other case" was incorrectly compiled such that it would only try to
490        match starting at "c".
491    
492    13. When a pattern starting with \s was studied, VT was not included in the
493        list of possible starting characters; this should have been part of the
494        8.34/18 patch.
495    
496    14. If a character class started [\Qx]... where x is any character, the class
497        was incorrectly terminated at the ].
498    
499    15. If a pattern that started with a caseless match for a character with more
500        than one "other case" was studied, PCRE did not set up the starting code
501        unit bit map for the list of possible characters. Now it does. This is an
502        optimization improvement, not a bug fix.
503    
504    16. The Unicode data tables have been updated to Unicode 7.0.0.
505    
506    17. Fixed a number of memory leaks in pcregrep.
507    
508    18. Avoid a compiler warning (from some compilers) for a function call with
509        a cast that removes "const" from an lvalue by using an intermediate
510        variable (to which the compiler does not object).
511    
512    19. Incorrect code was compiled if a group that contained an internal recursive
513        back reference was optional (had quantifier with a minimum of zero). This
514        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
515        caused segmentation faults because of stack overflows at compile time.
516    
517    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
518        group that is quantified with an indefinite repeat, caused a compile-time
519        loop which used up all the system stack and provoked a segmentation fault.
520        This was not the same bug as 19 above.
521    
522    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
523        Patch by Mike Frysinger.
524    
525    
526    Version 8.35 04-April-2014
527    --------------------------
528    
529    1.  A new flag is set, when property checks are present in an XCLASS.
530        When this flag is not set, PCRE can perform certain optimizations
531        such as studying these XCLASS-es.
532    
533    2.  The auto-possessification of character sets were improved: a normal
534        and an extended character set can be compared now. Furthermore
535        the JIT compiler optimizes more character set checks.
536    
537    3.  Got rid of some compiler warnings for potentially uninitialized variables
538        that show up only when compiled with -O2.
539    
540    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
541        of a match later then the end of the match. The pcretest program was not
542        handling the case sensibly - it was outputting from the start to the next
543        binary zero. It now reports this situation in a message, and outputs the
544        text from the end to the start.
545    
546    5.  Fast forward search is improved in JIT. Instead of the first three
547        characters, any three characters with fixed position can be searched.
548        Search order: first, last, middle.
549    
550    6.  Improve character range checks in JIT. Characters are read by an inprecise
551        function now, which returns with an unknown value if the character code is
552        above a certain threshold (e.g: 256). The only limitation is that the value
553        must be bigger than the threshold as well. This function is useful when
554        the characters above the threshold are handled in the same way.
555    
556    7.  The macros whose names start with RAWUCHAR are placeholders for a future
557        mode in which only the bottom 21 bits of 32-bit data items are used. To
558        make this more memorable for those maintaining the code, the names have
559        been changed to start with UCHAR21, and an extensive comment has been added
560        to their definition.
561    
562    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
563        to the export list in Makefile.am (they were accidentally omitted from the
564        8.34 tarball).
565    
566    9.  The informational output from pcretest used the phrase "starting byte set"
567        which is inappropriate for the 16-bit and 32-bit libraries. As the output
568        for "first char" and "need char" really means "non-UTF-char", I've changed
569        "byte" to "char", and slightly reworded the output. The documentation about
570        these values has also been (I hope) clarified.
571    
572    10. Another JIT related optimization: use table jumps for selecting the correct
573        backtracking path, when more than four alternatives are present inside a
574        bracket.
575    
576    11. Empty match is not possible, when the minimum length is greater than zero,
577        and there is no \K in the pattern. JIT should avoid empty match checks in
578        such cases.
579    
580    12. In a caseless character class with UCP support, when a character with more
581        than one alternative case was not the first character of a range, not all
582        the alternative cases were added to the class. For example, s and \x{17f}
583        are both alternative cases for S: the class [RST] was handled correctly,
584        but [R-T] was not.
585    
586    13. The configure.ac file always checked for pthread support when JIT was
587        enabled. This is not used in Windows, so I have put this test inside a
588        check for the presence of windows.h (which was already tested for).
589    
590    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
591        The algorithm provides a way to skip certain starting offsets, and usually
592        faster than linear prefix searches.
593    
594    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
595        as for 'fr_FR' and 'french'. For some reason, however, it then used the
596        Windows-specific input and output files, which have 'french' screwed in.
597        So this could never have worked. One of the problems with locales is that
598        they aren't always the same. I have now updated RunTest so that it checks
599        the output of the locale test (test 3) against three different output
600        files, and it allows the test to pass if any one of them matches. With luck
601        this should make the test pass on some versions of Solaris where it was
602        failing. Because of the uncertainty, the script did not used to stop if
603        test 3 failed; it now does. If further versions of a French locale ever
604        come to light, they can now easily be added.
605    
606    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
607        there was a message during ./configure, but it did not stop. This now
608        provokes an error. The invalid example in README has been corrected.
609        If a value less than the minimum is given, the minimum value has always
610        been used, but now a warning is given.
611    
612    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
613        was a bug in the test system, which is now fixed. Also, the list of various
614        configurations that are tested for each release did not have one with both
615        16/32 bits and --enable-bar-anycrlf. It now does.
616    
617    18. pcretest was missing "-C bsr" for displaying the \R default setting.
618    
619    19. Little endian PowerPC systems are supported now by the JIT compiler.
620    
621    20. The fast forward newline mechanism could enter to an infinite loop on
622        certain invalid UTF-8 input. Although we don't support these cases
623        this issue can be fixed by a performance optimization.
624    
625    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
626        not take account if existing stack usage. There is now a new global
627        variable called pcre_stack_guard that can be set to point to an external
628        function to check stack availability. It is called at the start of
629        processing every parenthesized group.
630    
631    22. A typo in the code meant that in ungreedy mode the max/min qualifier
632        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
633        not match "ab".
634    
635    23. When UTF was disabled, the JIT program reported some incorrect compile
636        errors. These messages are silenced now.
637    
638    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
639        compiler.
640    
641    25. Change all the temporary files used in RunGrepTest to be different to those
642        used by RunTest so that the tests can be run simultaneously, for example by
643        "make -j check".
644    
645    
646    Version 8.34 15-December-2013
647    -----------------------------
648    
649    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
650        executable memory. Patch inspired by Carsten Klein.
651    
652    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
653        this macro is never tested and has no effect, because the work to support
654        coverage involves only compiling and linking options and special targets in
655        the Makefile. The comment in config.h implied that defining the macro would
656        enable coverage support, which is totally false. There was also support for
657        setting this macro in the CMake files (my fault, I just copied it from
658        configure). SUPPORT_GCOV has now been removed.
659    
660    3.  Make a small performance improvement in strlen16() and strlen32() in
661        pcretest.
662    
663    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
664        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
665        these cases). There was also one in pcretest.c.
666    
667    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
668    
669    6.  In UTF mode, the code for checking whether a group could match an empty
670        string (which is used for indefinitely repeated groups to allow for
671        breaking an infinite loop) was broken when the group contained a repeated
672        negated single-character class with a character that occupied more than one
673        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
674        UTF-8 mode). The effect was undefined: the group might or might not be
675        deemed as matching an empty string, or the program might have crashed.
676    
677    7.  The code for checking whether a group could match an empty string was not
678        recognizing that \h, \H, \v, \V, and \R must match a character.
679    
680    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
681        an empty string. If it can, pcretest shows this in its information output.
682    
683    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
684        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
685        matched by pcre_exec() without using JIT:
686    
687        (a) If the rest of the pattern did not match after a maximal run of
688            grapheme clusters, the code for backing up to try with fewer of them
689            did not always back up over a full grapheme when characters that do not
690            have the modifier quality were involved, e.g. Hangul syllables.
691    
692        (b) If the match point in a subject started with modifier character, and
693            there was no match, the code could incorrectly back up beyond the match
694            point, and potentially beyond the first character in the subject,
695            leading to a segfault or an incorrect match result.
696    
697    10. A conditional group with an assertion condition could lead to PCRE
698        recording an incorrect first data item for a match if no other first data
699        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
700        first data item, and therefore matched "ca" after "c" instead of at the
701        start.
702    
703    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
704        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
705    
706    12. The source of pcregrep now includes z/OS-specific code so that it can be
707        compiled for z/OS as part of the special z/OS distribution.
708    
709    13. Added the -T and -TM options to pcretest.
710    
711    14. The code in pcre_compile.c for creating the table of named capturing groups
712        has been refactored. Instead of creating the table dynamically during the
713        actual compiling pass, the information is remembered during the pre-compile
714        pass (on the stack unless there are more than 20 named groups, in which
715        case malloc() is used) and the whole table is created before the actual
716        compile happens. This has simplified the code (it is now nearly 150 lines
717        shorter) and prepared the way for better handling of references to groups
718        with duplicate names.
719    
720    15. A back reference to a named subpattern when there is more than one of the
721        same name now checks them in the order in which they appear in the pattern.
722        The first one that is set is used for the reference. Previously only the
723        first one was inspected. This change makes PCRE more compatible with Perl.
724    
725    16. Unicode character properties were updated from Unicode 6.3.0.
726    
727    17. The compile-time code for auto-possessification has been refactored, based
728        on a patch by Zoltan Herczeg. It now happens after instead of during
729        compilation. The code is cleaner, and more cases are handled. The option
730        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
731        options in pcretest are provided to set it. It can also be set by
732        (*NO_AUTO_POSSESS) at the start of a pattern.
733    
734    18. The character VT has been added to the default ("C" locale) set of
735        characters that match \s and are generally treated as white space,
736        following this same change in Perl 5.18. There is now no difference between
737        "Perl space" and "POSIX space". Whether VT is treated as white space in
738        other locales depends on the locale.
739    
740    19. The code for checking named groups as conditions, either for being set or
741        for being recursed, has been refactored (this is related to 14 and 15
742        above). Processing unduplicated named groups should now be as fast at
743        numerical groups, and processing duplicated groups should be faster than
744        before.
745    
746    20. Two patches to the CMake build system, by Alexander Barkov:
747    
748          (1) Replace the "source" command by "." in CMakeLists.txt because
749              "source" is a bash-ism.
750    
751          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
752              without these the CMake build does not work on Solaris.
753    
754    21. Perl has changed its handling of \8 and \9. If there is no previously
755        encountered capturing group of those numbers, they are treated as the
756        literal characters 8 and 9 instead of a binary zero followed by the
757        literals. PCRE now does the same.
758    
759    22. Following Perl, added \o{} to specify codepoints in octal, making it
760        possible to specify values greater than 0777 and also making them
761        unambiguous.
762    
763    23. Perl now gives an error for missing closing braces after \x{... instead of
764        treating the string as literal. PCRE now does the same.
765    
766    24. RunTest used to grumble if an inappropriate test was selected explicitly,
767        but just skip it when running all tests. This make it awkward to run ranges
768        of tests when one of them was inappropriate. Now it just skips any
769        inappropriate tests, as it always did when running all tests.
770    
771    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
772        character types such as \d or \w, too many callouts were inserted, and the
773        data that they returned was rubbish.
774    
775    26. In UCP mode, \s was not matching two of the characters that Perl matches,
776        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
777        were matched by \h. The code has now been refactored so that the lists of
778        the horizontal and vertical whitespace characters used for \h and \v (which
779        are defined only in one place) are now also used for \s.
780    
781    27. Add JIT support for the 64 bit TileGX architecture.
782        Patch by Jiong Wang (Tilera Corporation).
783    
784    28. Possessive quantifiers for classes (both explicit and automatically
785        generated) now use special opcodes instead of wrapping in ONCE brackets.
786    
787    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
788        (because it's meaningless), this was not happening when PCRE_CASELESS was
789        set. Not wrong, but inefficient.
790    
791    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
792        properties for \w, \d, etc) is present in a test regex. Otherwise if the
793        test contains no characters greater than 255, Perl doesn't realise it
794        should be using Unicode semantics.
795    
796    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
797        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
798        does in Unicode mode.
799    
800    32. Added the "forbid" facility to pcretest so that putting tests into the
801        wrong test files can sometimes be quickly detected.
802    
803    33. There is now a limit (default 250) on the depth of nesting of parentheses.
804        This limit is imposed to control the amount of system stack used at compile
805        time. It can be changed at build time by --with-parens-nest-limit=xxx or
806        the equivalent in CMake.
807    
808    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
809        errors. Perl warns for these when in warning mode, but PCRE has no facility
810        for giving warnings.
811    
812    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
813        However, this was not working for (?!) because it is optimized to (*FAIL),
814        for which PCRE does not allow quantifiers. The optimization is now disabled
815        when a quantifier follows (?!). I can't see any use for this, but it makes
816        things uniform.
817    
818    36. Perl no longer allows group names to start with digits, so I have made this
819        change also in PCRE. It simplifies the code a bit.
820    
821    37. In extended mode, Perl ignores spaces before a + that indicates a
822        possessive quantifier. PCRE allowed a space before the quantifier, but not
823        before the possessive +. It now does.
824    
825    38. The use of \K (reset reported match start) within a repeated possessive
826        group such as (a\Kb)*+ was not working.
827    
828    40. Document that the same character tables must be used at compile time and
829        run time, and that the facility to pass tables to pcre_exec() and
830        pcre_dfa_exec() is for use only with saved/restored patterns.
831    
832    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
833        features for Builds with MSVC:
834    
835        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
836           builds). The .rc files can be used to set FileDescription and many other
837           attributes.
838    
839        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
840           This allows higher-level build scripts which want .pdb files to avoid
841           hard-coding the exact files needed."
842    
843    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
844        mean "start of word" and "end of word", respectively, as a transition aid.
845    
846    43. A minimizing repeat of a class containing codepoints greater than 255 in
847        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
848        compiled to use the heap for recursion.
849    
850    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
851        is configured.
852    
853    
854    Version 8.33 28-May-2013
855    ------------------------
856    
857    1.  Added 'U' to some constants that are compared to unsigned integers, to
858        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
859        variables that are added to signed variables, to ensure the result is
860        signed and can be negated.
861    
862    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
863        CMake config files.
864    
865    3.  Revise the creation of config.h.generic so that all boolean macros are
866        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
867        overriding via -D on the command line possible.
868    
869    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
870        to unsigned int is reported to make a quite noticeable speed difference in
871        a specific Windows environment. Testing on Linux did also appear to show
872        some benefit (and it is clearly not harmful). Also fixed the definition of
873        Xop which should be unsigned.
874    
875    5.  Related to (4), changing the definition of the intermediate variable cc
876        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
877        improvements.
878    
879    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
880        unnecessary spaces.
881    
882    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
883        and later.
884    
885    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
886    
887    9.  Optimizing fast_forward_start_bits in JIT.
888    
889    10. Adding support for callouts in JIT, and fixing some issues revealed
890        during this work. Namely:
891    
892        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
893    
894        (b) Minimum length was not checked before the matching is started.
895    
896    11. The value of capture_last that is passed to callouts was incorrect in some
897        cases when there was a capture on one path that was subsequently abandoned
898        after a backtrack. Also, the capture_last value is now reset after a
899        recursion, since all captures are also reset in this case.
900    
901    12. The interpreter no longer returns the "too many substrings" error in the
902        case when an overflowing capture is in a branch that is subsequently
903        abandoned after a backtrack.
904    
905    13. In the pathological case when an offset vector of size 2 is used, pcretest
906        now prints out the matched string after a yield of 0 or 1.
907    
908    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
909        Only supported by the JIT compiler at the moment.
910    
911    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
912    
913    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
914        offset of the starting point of the matching process, provided the offsets
915        vector is large enough.
916    
917    17. The \A escape now records a lookbehind value of 1, though its execution
918        does not actually inspect the previous character. This is to ensure that,
919        in partial multi-segment matching, at least one character from the old
920        segment is retained when a new segment is processed. Otherwise, if there
921        are no lookbehinds in the pattern, \A might match incorrectly at the start
922        of a new segment.
923    
924    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
925    
926    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
927        gives some modest performance improvement in 8-bit mode.
928    
929    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
930        be expressed in certain programming languages using Universal Character
931        Names.
932    
933    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
934        which points out that "non characters" are not "characters that may not
935        appear in Unicode strings" but rather "characters that are reserved for
936        internal use and have only local meaning".
937    
938    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
939        there was a conditional group that depended on an assertion, if the
940        assertion was false, the callout that immediately followed the alternation
941        in the condition was skipped when pcre_exec() was used for matching.
942    
943    23. Allow an explicit callout to be inserted before an assertion that is the
944        condition for a conditional group, for compatibility with automatic
945        callouts, which always insert a callout at this point.
946    
947    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
948        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
949    
950    25. (*PRUNE) is now supported by the JIT compiler.
951    
952    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
953    
954    27. Fix the case where there are two or more SKIPs with arguments that may be
955        ignored.
956    
957    28. (*SKIP) is now supported by the JIT compiler.
958    
959    29. (*THEN) is now supported by the JIT compiler.
960    
961    30. Update RunTest with additional test selector options.
962    
963    31. The way PCRE handles backtracking verbs has been changed in two ways.
964    
965        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
966        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
967        backtracking. In some cases this makes it more Perl-compatible, but Perl's
968        rather obscure rules do not always do the same thing.
969    
970        (2) Previously, backtracking verbs were confined within assertions. This is
971        no longer the case for positive assertions, except for (*ACCEPT). Again,
972        this sometimes improves Perl compatibility, and sometimes does not.
973    
974    32. A number of tests that were in test 2 because Perl did things differently
975        have been moved to test 1, because either Perl or PCRE has changed, and
976        these tests are now compatible.
977    
978    32. Backtracking control verbs are now handled in the same way in JIT and
979        interpreter.
980    
981    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
982        contained a forward subroutine reference caused a compile error.
983    
984    34. Auto-detect and optimize limited repetitions in JIT.
985    
986    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
987        blocking (*UTF) etc.
988    
989    36. In the interpreter, maximizing pattern repetitions for characters and
990        character types now use tail recursion, which reduces stack usage.
991    
992    37. The value of the max lookbehind was not correctly preserved if a compiled
993        and saved regex was reloaded on a host of different endianness.
994    
995    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
996        of the compiled pattern block, expand the flags field from 16 to 32 bits
997        because it was almost full.
998    
999    39. Try madvise first before posix_madvise.
1000    
1001    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1002        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1003        original change locked out all matches of empty strings. This has been
1004        changed so that one match of an empty string per line is recognized.
1005        Subsequent searches on the same line (for colouring or for --only-matching,
1006        for example) do not recognize empty strings.
1007    
1008    41. Applied a user patch to fix a number of spelling mistakes in comments.
1009    
1010    42. Data lines longer than 65536 caused pcretest to crash.
1011    
1012    43. Clarified the data type for length and startoffset arguments for pcre_exec
1013        and pcre_dfa_exec in the function-specific man pages, where they were
1014        explicitly stated to be in bytes, never having been updated. I also added
1015        some clarification to the pcreapi man page.
1016    
1017    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1018        a segmentation fault.
1019    
1020    
1021    Version 8.32 30-November-2012
1022    -----------------------------
1023    
1024    1.  Improved JIT compiler optimizations for first character search and single
1025        character iterators.
1026    
1027    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1028        Patch by Daniel Richard G.
1029    
1030    3.  Single character iterator optimizations in the JIT compiler.
1031    
1032    4.  Improved JIT compiler optimizations for character ranges.
1033    
1034    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1035        Reported by Giuseppe D'Angelo.
1036    
1037    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1038        of a line, was being set incorrectly in cases where .* appeared inside
1039        atomic brackets at the start of a pattern, or where there was a subsequent
1040        *PRUNE or *SKIP.
1041    
1042    7.  Improved instruction cache flush for POWER/PowerPC.
1043        Patch by Daniel Richard G.
1044    
1045    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1046        grep:
1047    
1048        (a) There is now no limit to the number of patterns to be matched.
1049    
1050        (b) An error is given if a pattern is too long.
1051    
1052        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1053            are now supported.
1054    
1055        (d) --exclude-from and --include-from (multiple use) have been added.
1056    
1057        (e) Exclusions and inclusions now apply to all files and directories, not
1058            just to those obtained from scanning a directory recursively.
1059    
1060        (f) Multiple uses of -f and --file-list are now supported.
1061    
1062        (g) In a Windows environment, the default for -d has been changed from
1063            "read" (the GNU grep default) to "skip", because otherwise the presence
1064            of a directory in the file list provokes an error.
1065    
1066        (h) The documentation has been revised and clarified in places.
1067    
1068    9.  Improve the matching speed of capturing brackets.
1069    
1070    10. Changed the meaning of \X so that it now matches a Unicode extended
1071        grapheme cluster.
1072    
1073    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1074        out POSIX threads when JIT support is configured.
1075    
1076    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1077    
1078    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1079        structure could go wrong in environments where size_t is not the same size
1080        as int.
1081    
1082    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1083        set.
1084    
1085    15. The EBCDIC support had decayed; later updates to the code had included
1086        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1087        general tidy up of EBCDIC-related issues, and the documentation was also
1088        not quite right. There is now a test that can be run on ASCII systems to
1089        check some of the EBCDIC-related things (but is it not a full test).
1090    
1091    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1092        in a small tidy to the code.
1093    
1094    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1095    
1096    18. If the --only-matching (-o) option in pcregrep is specified multiple
1097        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1098        substrings matched by the 1st and 2nd capturing parentheses. A separating
1099        string can be specified by --om-separator (default empty).
1100    
1101    19. Improving the first n character searches.
1102    
1103    20. Turn case lists for horizontal and vertical white space into macros so that
1104        they are defined only once.
1105    
1106    21. This set of changes together give more compatible Unicode case-folding
1107        behaviour for characters that have more than one other case when UCP
1108        support is available.
1109    
1110        (a) The Unicode property table now has offsets into a new table of sets of
1111            three or more characters that are case-equivalent. The MultiStage2.py
1112            script that generates these tables (the pcre_ucd.c file) now scans
1113            CaseFolding.txt instead of UnicodeData.txt for character case
1114            information.
1115    
1116        (b) The code for adding characters or ranges of characters to a character
1117            class has been abstracted into a generalized function that also handles
1118            case-independence. In UTF-mode with UCP support, this uses the new data
1119            to handle characters with more than one other case.
1120    
1121        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1122            whose other case is greater than 256 are now correctly matched
1123            caselessly. Previously, the high codepoint matched the low one, but not
1124            vice versa.
1125    
1126        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1127            of the new class addition function, using character lists defined as
1128            macros alongside the case definitions of 20 above.
1129    
1130        (e) Caseless back references now work with characters that have more than
1131            one other case.
1132    
1133        (f) General caseless matching of characters with more than one other case
1134            is supported.
1135    
1136    22. Unicode character properties were updated from Unicode 6.2.0
1137    
1138    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1139    
1140    24. Add support for 32-bit character strings, and UTF-32
1141    
1142    25. Major JIT compiler update (code refactoring and bugfixing).
1143        Experimental Sparc 32 support is added.
1144    
1145    26. Applied a modified version of Daniel Richard G's patch to create
1146        pcre.h.generic and config.h.generic by "make" instead of in the
1147        PrepareRelease script.
1148    
1149    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1150        pcre_compile.c when checking for a zero character.
1151    
1152    28. Introducing a native interface for JIT. Through this interface, the compiled
1153        machine code can be directly executed. The purpose of this interface is to
1154        provide fast pattern matching, so several sanity checks are not performed.
1155        However, feature tests are still performed. The new interface provides
1156        1.4x speedup compared to the old one.
1157    
1158    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1159        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1160        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1161    
1162    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1163        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1164        the "old" RFC 2279). Instead, it ended up passing a negative length to
1165        pcre_exec().
1166    
1167    31. Add support for GCC's visibility feature to hide internal functions.
1168    
1169    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1170        "unknown -C option" after outputting 0 or 1.
1171    
1172    33. There is now support for generating a code coverage report for the test
1173        suite in environments where gcc is the compiler and lcov is installed. This
1174        is mainly for the benefit of the developers.
1175    
1176    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1177        unaddressable using valgrind annotations, allowing valgrind to detect
1178        invalid memory accesses. This is mainly for the benefit of the developers.
1179    
1180    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1181    
1182    26. Give configure error if --enable-cpp but no C++ compiler found.
1183    
1184    
1185    Version 8.31 06-July-2012
1186    -------------------------
1187    
1188    1.  Fixing a wrong JIT test case and some compiler warnings.
1189    
1190    2.  Removed a bashism from the RunTest script.
1191    
1192    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1193        to unsigned type, result still unsigned" that was given by an MS compiler
1194        on encountering the code "-sizeof(xxx)".
1195    
1196    4.  Partial matching support is added to the JIT compiler.
1197    
1198    5.  Fixed several bugs concerned with partial matching of items that consist
1199        of more than one character:
1200    
1201        (a) /^(..)\1/ did not partially match "aba" because checking references was
1202            done on an "all or nothing" basis. This also applied to repeated
1203            references.
1204    
1205        (b) \R did not give a hard partial match if \r was found at the end of the
1206            subject.
1207    
1208        (c) \X did not give a hard partial match after matching one or more
1209            characters at the end of the subject.
1210    
1211        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1212            a partial match for the string "\r".
1213    
1214        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1215            a partial match for a CR character at the end of the subject string.
1216    
1217    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1218        running pcretest, the text "(JIT)" added to the output whenever JIT is
1219        actually used to run the match.
1220    
1221    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1222        or /S+[+] with a digit between 1 and 7.
1223    
1224    8.  OP_NOT now supports any UTF character not just single-byte ones.
1225    
1226    9.  (*MARK) control verb is now supported by the JIT compiler.
1227    
1228    10. The command "./RunTest list" lists the available tests without actually
1229        running any of them. (Because I keep forgetting what they all are.)
1230    
1231    11. Add PCRE_INFO_MAXLOOKBEHIND.
1232    
1233    12. Applied a (slightly modified) user-supplied patch that improves performance
1234        when the heap is used for recursion (compiled with --disable-stack-for-
1235        recursion). Instead of malloc and free for each heap frame each time a
1236        logical recursion happens, frames are retained on a chain and re-used where
1237        possible. This sometimes gives as much as 30% improvement.
1238    
1239    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1240        call.
1241    
1242    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1243    
1244    15. It is now possible to link pcretest with libedit as an alternative to
1245        libreadline.
1246    
1247    16. (*COMMIT) control verb is now supported by the JIT compiler.
1248    
1249    17. The Unicode data tables have been updated to Unicode 6.1.0.
1250    
1251    18. Added --file-list option to pcregrep.
1252    
1253    19. Added binary file support to pcregrep, including the -a, --binary-files,
1254        -I, and --text options.
1255    
1256    20. The madvise function is renamed for posix_madvise for QNX compatibility
1257        reasons. Fixed by Giuseppe D'Angelo.
1258    
1259    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1260    
1261    22. Changed the option for creating symbolic links for 16-bit man pages from
1262        -s to -sf so that re-installing does not cause issues.
1263    
1264    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1265    
1266    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1267        matches in certain environments (the workspace was not being correctly
1268        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1269        some of the workspace data at the beginning of a restart.
1270    
1271    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1272        was not doing so when it should - probably a typo introduced by SVN 528
1273        (change 8.10/14).
1274    
1275    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1276        \w+ when the character tables indicated that \x{c4} was a word character.
1277        There were several related cases, all because the tests for doing a table
1278        lookup were testing for characters less than 127 instead of 255.
1279    
1280    27. If a pattern contains capturing parentheses that are not used in a match,
1281        their slots in the ovector are set to -1. For those that are higher than
1282        any matched groups, this happens at the end of processing. In the case when
1283        there were back references that the ovector was too small to contain
1284        (causing temporary malloc'd memory to be used during matching), and the
1285        highest capturing number was not used, memory off the end of the ovector
1286        was incorrectly being set to -1. (It was using the size of the temporary
1287        memory instead of the true size.)
1288    
1289    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1290        ovector size, it uses memory at the end of the block that it has got.
1291    
1292    29. Check for an overlong MARK name and give an error at compile time. The
1293        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1294    
1295    30. JIT compiler update.
1296    
1297    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1298        Rill for the patch.
1299    
1300    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1301    
1302    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1303    
1304    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1305        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1306        was enabled.
1307    
1308    35. Improve JIT code generation for greedy plus quantifier.
1309    
1310    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1311        1 to "aa" instead of to an empty string. The bug affected repeated groups
1312        that could potentially match an empty string.
1313    
1314    37. Optimizing single character iterators in JIT.
1315    
1316    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1317        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1318        codepoints that are too big for the mode are faulted, and in a UTF mode,
1319        disallowed codepoints are also faulted.
1320    
1321    39. If PCRE was compiled with UTF support, in three places in the DFA
1322        matcher there was code that should only have been obeyed in UTF mode, but
1323        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1324        processing when bytes with values greater than 127 were present. In 16-bit
1325        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1326        both cases the values are those that cannot be the first data item in a UTF
1327        character. The three items that might have provoked this were recursions,
1328        possessively repeated groups, and atomic groups.
1329    
1330    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1331        and pcregrep, because some OS require shared objects to be explicitly
1332        passed to ld, causing the link step to fail if they are not.
1333    
1334    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1335        mode, patterns that started with \h* or \R* might be incorrectly matched.
1336    
1337    
1338    Version 8.30 04-February-2012
1339    -----------------------------
1340    
1341    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1342        name is defined in ctype.h.
1343    
1344    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1345        only in quite long subpatterns.
1346    
1347    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1348        since it was replaced by pcre_fullinfo() in February 2000.
1349    
1350    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1351        match a (*MARK), and the match failed at the start of the subject, a
1352        reference to memory before the start of the subject could occur. This bug
1353        was introduced by fix 17 of release 8.21.
1354    
1355    5.  A reference to an unset group with zero minimum repetition was giving
1356        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1357        /(another)?(\1?)test/ matched against "hello world test". This bug was
1358        introduced in release 8.13.
1359    
1360    6.  Add support for 16-bit character strings (a large amount of work involving
1361        many changes and refactorings).
1362    
1363    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1364        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1365        from a file.
1366    
1367    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1368        rounding is not applied in this particular case).
1369    
1370    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1371        if they appear, or are escaped, in patterns.
1372    
1373    10. Get rid of a number of -Wunused-but-set-variable warnings.
1374    
1375    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1376        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1377        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1378        also returns the mark "x". This bug applied to capturing parentheses,
1379        non-capturing parentheses, and atomic parentheses. It also applied to some
1380        assertions.
1381    
1382    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1383        information out of configure.ac instead of relying on pcre.h.generic, which
1384        is not stored in the repository.
1385    
1386    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1387        -lreadline.
1388    
1389    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1390    
1391    15. Applied Graycode's patch to put the top-level frame on the stack rather
1392        than the heap when not using the stack for recursion. This gives a
1393        performance improvement in many cases when recursion is not deep.
1394    
1395    16. Experimental code added to "pcretest -C" to output the stack frame size.
1396    
1397    
1398    Version 8.21 12-Dec-2011
1399    ------------------------
1400    
1401    1.  Updating the JIT compiler.
1402    
1403    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1404        are added as well.
1405    
1406    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1407        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1408        calling _pcre_jit_exec. Some extra comments are added.
1409    
1410    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1411        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1412        was introduced by change 18 for 8.20.
1413    
1414    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1415        ECMA-262 standard.
1416    
1417    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1418        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1419        This bug was probably introduced by change 9 of 8.13.
1420    
1421    7.  While fixing 6 above, I noticed that a number of other items were being
1422        incorrectly rejected as "not fixed length". This arose partly because newer
1423        opcodes had not been added to the fixed-length checking code. I have (a)
1424        corrected the bug and added tests for these items, and (b) arranged for an
1425        error to occur if an unknown opcode is encountered while checking for fixed
1426        length instead of just assuming "not fixed length". The items that were
1427        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1428        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1429        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1430    
1431    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1432        being incorrectly compiled and would have given unpredicatble results.
1433    
1434    9.  A possessively repeated subpattern with minimum repeat count greater than
1435        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1436        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1437        the first (A) could occur when it should not.
1438    
1439    10. Add a cast and remove a redundant test from the code.
1440    
1441    11. JIT should use pcre_malloc/pcre_free for allocation.
1442    
1443    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1444        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1445        is anything other than /usr, -L is still shown).
1446    
1447    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1448    
1449    14. Perl does not support \N without a following name in a [] class; PCRE now
1450        also gives an error.
1451    
1452    15. If a forward reference was repeated with an upper limit of around 2000,
1453        it caused the error "internal error: overran compiling workspace". The
1454        maximum number of forward references (including repeats) was limited by the
1455        internal workspace, and dependent on the LINK_SIZE. The code has been
1456        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1457        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1458        of around 200,000 forward references. While doing this, I also speeded up
1459        the filling in of repeated forward references.
1460    
1461    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1462        incorrectly expecting the subject to contain another "a" after the start.
1463    
1464    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1465        in the match, the SKIP should be ignored. This was not happening; instead
1466        the SKIP was being treated as NOMATCH. For patterns such as
1467        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1468        tested.
1469    
1470    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1471        now much more compatible with Perl, in particular in cases where the result
1472        is a non-match for a non-anchored pattern. For example, if
1473        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1474        "m", where previously it did not return a name. A side effect of this
1475        change is that for partial matches, the last encountered mark name is
1476        returned, as for non matches. A number of tests that were previously not
1477        Perl-compatible have been moved into the Perl-compatible test files. The
1478        refactoring has had the pleasing side effect of removing one argument from
1479        the match() function, thus reducing its stack requirements.
1480    
1481    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1482        subsequent uses of /S (without +) incorrectly behaved like /S+.
1483    
1484    21. Retrieve executable code size support for the JIT compiler and fixing
1485        some warnings.
1486    
1487    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1488        not work when the shorter character appeared right at the end of the
1489        subject string.
1490    
1491    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1492        systems.
1493    
1494    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1495        output it when the /M option is used in pcretest.
1496    
1497    25. The CheckMan script was not being included in the distribution. Also, added
1498        an explicit "perl" to run Perl scripts from the PrepareRelease script
1499        because this is reportedly needed in Windows.
1500    
1501    26. If study data was being save in a file and studying had not found a set of
1502        "starts with" bytes for the pattern, the data written to the file (though
1503        never used) was taken from uninitialized memory and so caused valgrind to
1504        complain.
1505    
1506    27. Updated RunTest.bat as provided by Sheri Pierce.
1507    
1508    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1509    
1510    29. Computation of memory usage for the table of capturing group names was
1511        giving an unnecessarily large value.
1512    
1513    
1514    Version 8.20 21-Oct-2011
1515    ------------------------
1516    
1517    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1518        a POSIX class. After further experiments with Perl, which convinced me that
1519        Perl has bugs and confusions, a closing square bracket is no longer allowed
1520        in a POSIX name. This bug also affected patterns with classes that started
1521        with full stops.
1522    
1523    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1524        captured substring, but while checking the failing first alternative,
1525        substring 1 is temporarily captured. If the output vector supplied to
1526        pcre_exec() was not big enough for this capture, the yield of the function
1527        was still zero ("insufficient space for captured substrings"). This cannot
1528        be totally fixed without adding another stack variable, which seems a lot
1529        of expense for a edge case. However, I have improved the situation in cases
1530        such as /(a)(b)x|abc/ matched against "abc", where the return code
1531        indicates that fewer than the maximum number of slots in the ovector have
1532        been set.
1533    
1534    3.  Related to (2) above: when there are more back references in a pattern than
1535        slots in the output vector, pcre_exec() uses temporary memory during
1536        matching, and copies in the captures as far as possible afterwards. It was
1537        using the entire output vector, but this conflicts with the specification
1538        that only 2/3 is used for passing back captured substrings. Now it uses
1539        only the first 2/3, for compatibility. This is, of course, another edge
1540        case.
1541    
1542    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1543        main code base, and can be used by building with --enable-jit. When this is
1544        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1545        runtime --no-jit option is given.
1546    
1547    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1548        ovector, the return from the function was zero, implying that there were
1549        other matches that did not fit. The correct "exactly full" value is now
1550        returned.
1551    
1552    6.  If a subpattern that was called recursively or as a subroutine contained
1553        (*PRUNE) or any other control that caused it to give a non-standard return,
1554        invalid errors such as "Error -26 (nested recursion at the same subject
1555        position)" or even infinite loops could occur.
1556    
1557    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1558        computing the minimum length on reaching *ACCEPT, and so ended up with the
1559        wrong value of 1 rather than 0. Further investigation indicates that
1560        computing a minimum subject length in the presence of *ACCEPT is difficult
1561        (think back references, subroutine calls), and so I have changed the code
1562        so that no minimum is registered for a pattern that contains *ACCEPT.
1563    
1564    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1565        it was not handled as intended. [But see 16 below.]
1566    
1567    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1568        Sheri Pierce.
1569    
1570    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1571        the first byte in a match must be "a".
1572    
1573    11. Change 17 for 8.13 increased the recursion depth for patterns like
1574        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1575        pattern contains any instances of (*THEN). If it does not, the old
1576        optimizations are restored. It would be nice to do this on a per-group
1577        basis, but at the moment that is not feasible.
1578    
1579    12. In some environments, the output of pcretest -C is CRLF terminated. This
1580        broke RunTest's code that checks for the link size. A single white space
1581        character after the value is now allowed for.
1582    
1583    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1584        For "fr", it uses the Windows-specific input and output files.
1585    
1586    14. If (*THEN) appeared in a group that was called recursively or as a
1587        subroutine, it did not work as intended. [But see next item.]
1588    
1589    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1590        pattern fragments (but not containing any | characters). If A and B are
1591        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1592        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1593        D. In other words, Perl considers parentheses that do not contain any |
1594        characters to be part of a surrounding alternative, whereas PCRE was
1595        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1596        differently. PCRE now behaves in the same way as Perl, except in the case
1597        of subroutine/recursion calls such as (?1) which have in any case always
1598        been different (but PCRE had them first :-).
1599    
1600    16. Related to 15 above: Perl does not treat the | in a conditional group as
1601        creating alternatives. Such a group is treated in the same way as an
1602        ordinary group without any | characters when processing (*THEN). PCRE has
1603        been changed to match Perl's behaviour.
1604    
1605    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1606        RunGrepTest script failed.
1607    
1608    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1609        inevitable for groups that contain captures, but it can lead to a lot of
1610        stack use in large patterns. The old behaviour has been restored for atomic
1611        groups that do not contain any capturing parentheses.
1612    
1613    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1614        suppress the check for a minimum subject length at run time. (If it was
1615        given to pcre_exec() or pcre_dfa_exec() it did work.)
1616    
1617    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1618        fail to work when decoding hex characters in data strings in EBCDIC
1619        environments.
1620    
1621    21. It appears that in at least one Mac OS environment, the isxdigit() function
1622        is implemented as a macro that evaluates to its argument more than once,
1623        contravening the C 90 Standard (I haven't checked a later standard). There
1624        was an instance in pcretest which caused it to go wrong when processing
1625        \x{...} escapes in subject strings. The has been rewritten to avoid using
1626        things like p++ in the argument of isxdigit().
1627    
1628    
1629    Version 8.13 16-Aug-2011
1630    ------------------------
1631    
1632    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1633    
1634    2.  Two minor typos in pcre_internal.h have been fixed.
1635    
1636    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1637        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1638        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1639    
1640    4.  There were a number of related bugs in the code for matching backrefences
1641        caselessly in UTF-8 mode when codes for the characters concerned were
1642        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1643        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1644        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1645        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1646        2-byte code at the end of the subject (it thought there wasn't enough data
1647        left).
1648    
1649    5.  Comprehensive information about what went wrong is now returned by
1650        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1651        as the output vector has at least 2 elements. The offset of the start of
1652        the failing character and a reason code are placed in the vector.
1653    
1654    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1655        now returned is for the first byte of the failing character, instead of the
1656        last byte inspected. This is an incompatible change, but I hope it is small
1657        enough not to be a problem. It makes the returned offset consistent with
1658        pcre_exec() and pcre_dfa_exec().
1659    
1660    7.  pcretest now gives a text phrase as well as the error number when
1661        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1662        failure, the offset and reason code are output.
1663    
1664    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1665        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1666        back over a single character (\n). This seems wrong (because it treated the
1667        two characters as a single entity when going forwards), conflicts with the
1668        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1669        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1670        has been changed.
1671    
1672    9.  Some internal refactoring has changed the processing so that the handling
1673        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1674        time (the PCRE_DOTALL option was changed this way some time ago: version
1675        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1676        which was always a bit of a fudge. It also means that there is one less
1677        argument for the match() function, which reduces its stack requirements
1678        slightly. This change also fixes an incompatibility with Perl: the pattern
1679        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1680    
1681    10. More internal refactoring has drastically reduced the number of recursive
1682        calls to match() for possessively repeated groups such as (abc)++ when
1683        using pcre_exec().
1684    
1685    11. While implementing 10, a number of bugs in the handling of groups were
1686        discovered and fixed:
1687    
1688        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1689        (a|)*(?1) gave a compile-time internal error.
1690        ((a|)+)+  did not notice that the outer group could match an empty string.
1691        (^a|^)+   was not marked as anchored.
1692        (.*a|.*)+ was not marked as matching at start or after a newline.
1693    
1694    12. Yet more internal refactoring has removed another argument from the match()
1695        function. Special calls to this function are now indicated by setting a
1696        value in a variable in the "match data" data block.
1697    
1698    13. Be more explicit in pcre_study() instead of relying on "default" for
1699        opcodes that mean there is no starting character; this means that when new
1700        ones are added and accidentally left out of pcre_study(), testing should
1701        pick them up.
1702    
1703    14. The -s option of pcretest has been documented for ages as being an old
1704        synonym of -m (show memory usage). I have changed it to mean "force study
1705        for every regex", that is, assume /S for every regex. This is similar to -i
1706        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1707        using it. It makes it easier to run collections of tests with and without
1708        study enabled, and thereby test pcre_study() more easily. All the standard
1709        tests are now run with and without -s (but some patterns can be marked as
1710        "never study" - see 20 below).
1711    
1712    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1713        restoration of the capturing data to the outer values was not happening
1714        correctly.
1715    
1716    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1717        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1718        pattern had matched an empty string, and so incorrectly returned a no
1719        match.
1720    
1721    17. There was optimizing code for the last branch of non-capturing parentheses,
1722        and also for the obeyed branch of a conditional subexpression, which used
1723        tail recursion to cut down on stack usage. Unfortunately, now that there is
1724        the possibility of (*THEN) occurring in these branches, tail recursion is
1725        no longer possible because the return has to be checked for (*THEN). These
1726        two optimizations have therefore been removed. [But see 8.20/11 above.]
1727    
1728    18. If a pattern containing \R was studied, it was assumed that \R always
1729        matched two bytes, thus causing the minimum subject length to be
1730        incorrectly computed because \R can also match just one byte.
1731    
1732    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1733        was incorrectly computed.
1734    
1735    20. If /S is present twice on a test pattern in pcretest input, it now
1736        *disables* studying, thereby overriding the use of -s on the command line
1737        (see 14 above). This is necessary for one or two tests to keep the output
1738        identical in both cases.
1739    
1740    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1741        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1742    
1743    22. When an atomic group that contained a capturing parenthesis was
1744        successfully matched, but the branch in which it appeared failed, the
1745        capturing was not being forgotten if a higher numbered group was later
1746        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1747        group 1 to "a", when in fact it should be unset. This applied to multi-
1748        branched capturing and non-capturing groups, repeated or not, and also to
1749        positive assertions (capturing in negative assertions does not happen
1750        in PCRE) and also to nested atomic groups.
1751    
1752    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1753        subject after a captured substring, to make it easier to tell which of a
1754        number of identical substrings has been captured.
1755    
1756    24. The way atomic groups are processed by pcre_exec() has been changed so that
1757        if they are repeated, backtracking one repetition now resets captured
1758        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1759        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1760        "aaa". Previously, it would have been "a". As part of this code
1761        refactoring, the way recursive calls are handled has also been changed.
1762    
1763    25. If an assertion condition captured any substrings, they were not passed
1764        back unless some other capturing happened later. For example, if
1765        (?(?=(a))a) was matched against "a", no capturing was returned.
1766    
1767    26. When studying a pattern that contained subroutine calls or assertions,
1768        the code for finding the minimum length of a possible match was handling
1769        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1770        group 1 called group 2 while simultaneously a separate group 2 called group
1771        1). A stack overflow occurred in this case. I have fixed this by limiting
1772        the recursion depth to 10.
1773    
1774    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1775        Fortmann. This supports explicit test numbers on the command line, and has
1776        argument validation and error reporting.
1777    
1778    28. An instance of \X with an unlimited repeat could fail if at any point the
1779        first character it looked at was a mark character.
1780    
1781    29. Some minor code refactoring concerning Unicode properties and scripts
1782        should reduce the stack requirement of match() slightly.
1783    
1784    30. Added the '=' option to pcretest to check the setting of unused capturing
1785        slots at the end of the pattern, which are documented as being -1, but are
1786        not included in the return count.
1787    
1788    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1789        compiled something random. Now it gives a compile-time error (as does
1790        Perl).
1791    
1792    32. A *MARK encountered during the processing of a positive assertion is now
1793        recorded and passed back (compatible with Perl).
1794    
1795    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1796        had alternative anchored branches, the search for a second match in a line
1797        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1798        matched the line "0102" twice. The same bug affected patterns that started
1799        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1800        twice.
1801    
1802    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1803        does, and because of capturing effects, quantifying parenthesized
1804        assertions may at times be useful. Quantifiers are now allowed for
1805        parenthesized assertions.
1806    
1807    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1808    
1809    36. \g was being checked for fancy things in a character class, when it should
1810        just be a literal "g".
1811    
1812    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1813        appearance of a nested POSIX class supersedes an apparent external class.
1814        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1815        unescaped square brackets may also appear as part of class names. For
1816        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1817        more like Perl. (But see 8.20/1 above.)
1818    
1819    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1820        was because it thought it was \N{name}, which is not supported).
1821    
1822    39. Add minix to OS list not supporting the -S option in pcretest.
1823    
1824    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1825        cannot analyze patterns in sufficient detail to catch mutual recursions
1826        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1827        subgroup is called recursively as a subpattern for a second time at the
1828        same position in the subject string. In previous releases this might have
1829        been caught by the recursion limit, or it might have run out of stack.
1830    
1831    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1832        happen only once. PCRE was, however incorrectly giving a compile time error
1833        "recursive call could loop indefinitely" because it cannot analyze the
1834        pattern in sufficient detail. The compile time test no longer happens when
1835        PCRE is compiling a conditional subpattern, but actual runaway loops are
1836        now caught at runtime (see 40 above).
1837    
1838    42. It seems that Perl allows any characters other than a closing parenthesis
1839        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1840        has been changed to be the same.
1841    
1842    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1843        as not to get warnings when autogen.sh is called. Also changed
1844        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1845    
1846    44. To help people who use pcregrep to scan files containing exceedingly long
1847        lines, the following changes have been made:
1848    
1849        (a) The default value of the buffer size parameter has been increased from
1850            8K to 20K. (The actual buffer used is three times this size.)
1851    
1852        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1853            PCRE is built.
1854    
1855        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1856            to be set at run time.
1857    
1858        (d) Numerical values in pcregrep options can be followed by K or M, for
1859            example --buffer-size=50K.
1860    
1861        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1862            given and the return code is set to 2.
1863    
1864    45. Add a pointer to the latest mark to the callout data block.
1865    
1866    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1867        partial match of an empty string instead of no match. This was specific to
1868        the use of ".".
1869    
1870    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1871        complete match instead of a partial match. This bug was dependent on both
1872        the PCRE_UTF8 and PCRE_DOTALL options being set.
1873    
1874    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1875        starting byte set, because \b was not being ignored.
1876    
1877    
1878    Version 8.12 15-Jan-2011
1879    ------------------------
1880    
1881    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1882        checks for such things as part of the documentation building process.
1883    
1884    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1885        --match-limit and --recursion-limit options (added for 8.11). In
1886        particular, this made one of the standard tests fail. (The integer value
1887        went into the wrong half of a long int.)
1888    
1889    3.  If the --colour option was given to pcregrep with -v (invert match), it
1890        did strange things, either producing crazy output, or crashing. It should,
1891        of course, ignore a request for colour when reporting lines that do not
1892        match.
1893    
1894    4.  Another pcregrep bug caused similar problems if --colour was specified with
1895        -M (multiline) and the pattern match finished with a line ending.
1896    
1897    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1898        matched in multiline mode, the following line was shown as part of the
1899        match. This seems wrong, so I have changed it.
1900    
1901    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1902        the check for further matches in the same line (so they could be coloured)
1903        to overrun the end of the current line. If another match was found, it was
1904        incorrectly shown (and then shown again when found in the next line).
1905    
1906    7.  If pcregrep was compiled under Windows, there was a reference to the
1907        function pcregrep_exit() before it was defined. I am assuming this was
1908        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1909        reported by a user. I've moved the definition above the reference.
1910    
1911    
1912    Version 8.11 10-Dec-2010
1913    ------------------------
1914    
1915    1.  (*THEN) was not working properly if there were untried alternatives prior
1916        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1917        backtracked to try for "b" instead of moving to the next alternative branch
1918        at the same level (in this case, to look for "c"). The Perl documentation
1919        is clear that when (*THEN) is backtracked onto, it goes to the "next
1920        alternative in the innermost enclosing group".
1921    
1922    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1923        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1924        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1925        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1926        (*THEN).
1927    
1928    3.  If \s appeared in a character class, it removed the VT character from
1929        the class, even if it had been included by some previous item, for example
1930        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1931        of \s, but is part of the POSIX "space" class.)
1932    
1933    4.  A partial match never returns an empty string (because you can always
1934        match an empty string at the end of the subject); however the checking for
1935        an empty string was starting at the "start of match" point. This has been
1936        changed to the "earliest inspected character" point, because the returned
1937        data for a partial match starts at this character. This means that, for
1938        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1939        (previously it gave "no match").
1940    
1941    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1942        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1943        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1944        has an implication that the given string is incomplete (because a partial
1945        match is preferred over a full match). For this reason, these items now
1946        give a partial match in this situation. [Aside: previously, the one case
1947        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1948        match rather than a full match, which was wrong by the old rules, but is
1949        now correct.]
1950    
1951    6.  There was a bug in the handling of #-introduced comments, recognized when
1952        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1953        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1954        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1955        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1956        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1957        places in pcre_compile().
1958    
1959    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1960        comments when looking ahead for named forward references to subpatterns,
1961        the only newline sequence it recognized was NL. It now handles newlines
1962        according to the set newline convention.
1963    
1964    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1965        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1966        cater for a lack of strerror(). These oversights have been fixed.
1967    
1968    9.  Added --match-limit and --recursion-limit to pcregrep.
1969    
1970    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1971    
1972    11. When the -o option was used, pcregrep was setting a return code of 1, even
1973        when matches were found, and --line-buffered was not being honoured.
1974    
1975    12. Added an optional parentheses number to the -o and --only-matching options
1976        of pcregrep.
1977    
1978    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1979        can match an empty string. The code to do it in pcretest and pcredemo
1980        needed fixing:
1981    
1982        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1983            only one byte after an empty string match just before CRLF (this case
1984            just got forgotten; "any" and "anycrlf" were OK).
1985    
1986        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1987            mode when an empty string match preceded an ASCII character followed by
1988            a non-ASCII character. (The code for advancing by one character rather
1989            than one byte was nonsense.)
1990    
1991        (c) The pcredemo.c sample program did not have any code at all to handle
1992            the cases when CRLF is a valid newline sequence.
1993    
1994    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1995        as a starting offset was within the subject string. There is now a new
1996        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1997        negative or greater than the length of the string. In order to test this,
1998        pcretest is extended to allow the setting of negative starting offsets.
1999    
2000    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2001        starting offset points to the beginning of a UTF-8 character was
2002        unnecessarily clumsy. I tidied it up.
2003    
2004    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2005        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2006    
2007    17. Nobody had reported that the --include_dir option, which was added in
2008        release 7.7 should have been called --include-dir (hyphen, not underscore)
2009        for compatibility with GNU grep. I have changed it to --include-dir, but
2010        left --include_dir as an undocumented synonym, and the same for
2011        --exclude-dir, though that is not available in GNU grep, at least as of
2012        release 2.5.4.
2013    
2014    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2015        characters from a string of bytes) have been redefined so as not to use
2016        loops, in order to improve performance in some environments. At the same
2017        time, I abstracted some of the common code into auxiliary macros to save
2018        repetition (this should not affect the compiled code).
2019    
2020    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2021        compile-time error is now given if \c is not followed by an ASCII
2022        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2023        different, and any byte value is allowed.)
2024    
2025    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2026        START_OPTIMIZE option, which is now allowed at compile time - but just
2027        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2028        to pcregrep and other applications that have no direct access to PCRE
2029        options. The new /Y option in pcretest sets this option when calling
2030        pcre_compile().
2031    
2032    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2033        back references. Groups containing recursive back references were forced to
2034        be atomic by that change, but in the case of named groups, the amount of
2035        memory required was incorrectly computed, leading to "Failed: internal
2036        error: code overflow". This has been fixed.
2037    
2038    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2039        pcretest.c, to avoid build problems in some Borland environments.
2040    
2041    
2042    Version 8.10 25-Jun-2010
2043    ------------------------
2044    
2045    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2046        THEN.
2047    
2048    2.  (*ACCEPT) was not working when inside an atomic group.
2049    
2050    3.  Inside a character class, \B is treated as a literal by default, but
2051        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2052        causes the error). The code is unchanged, but I tidied the documentation.
2053    
2054    4.  Inside a character class, PCRE always treated \R and \X as literals,
2055        whereas Perl faults them if its -w option is set. I have changed PCRE so
2056        that it faults them when PCRE_EXTRA is set.
2057    
2058    5.  Added support for \N, which always matches any character other than
2059        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2060    
2061    6.  When compiling pcregrep with newer versions of gcc which may have
2062        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2063        declared with attribute warn_unused_result" were given. Just casting the
2064        result to (void) does not stop the warnings; a more elaborate fudge is
2065        needed. I've used a macro to implement this.
2066    
2067    7.  Minor change to pcretest.c to avoid a compiler warning.
2068    
2069    8.  Added four artifical Unicode properties to help with an option to make
2070        \s etc use properties (see next item). The new properties are: Xan
2071        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2072    
2073    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2074        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2075        this option. Modified pcretest to add /W to test this facility. Added
2076        REG_UCP to make it available via the POSIX interface.
2077    
2078    10. Added --line-buffered to pcregrep.
2079    
2080    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2081        studied, and the match started with a letter with a code point greater than
2082        127 whose first byte was different to the first byte of the other case of
2083        the letter, the other case of this starting letter was not recognized
2084        (#976).
2085    
2086    12. If a pattern that was studied started with a repeated Unicode property
2087        test, for example, \p{Nd}+, there was the theoretical possibility of
2088        setting up an incorrect bitmap of starting bytes, but fortunately it could
2089        not have actually happened in practice until change 8 above was made (it
2090        added property types that matched character-matching opcodes).
2091    
2092    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2093        possible starting bytes for non-anchored patterns.
2094    
2095    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2096        \R, and also a number of cases that involve Unicode properties, both
2097        explicit and implicit when PCRE_UCP is set.
2098    
2099    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2100        input, it could crash or give wrong results if characters with values
2101        greater than 0xc0 were present in the subject string. (Detail: it assumed
2102        UTF-8 input when processing these items.)
2103    
2104    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2105        size_t is 64-bit (#991).
2106    
2107    17. Added a check for running out of memory when PCRE is compiled with
2108        --disable-stack-for-recursion (#990).
2109    
2110    18. If the last data line in a file for pcretest does not have a newline on
2111        the end, a newline was missing in the output.
2112    
2113    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2114        less than 128) in its various bitmaps. However, there is a facility for
2115        generating tables according to the current locale when PCRE is compiled. It
2116        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2117        characters, are recognized by isspace() and therefore were getting set in
2118        these tables, and indeed these tables seem to approximate to ISO 8859. This
2119        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2120        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2121        which of course cannot start UTF-8 characters. I have changed the code so
2122        that only real ASCII characters (less than 128) and the correct starting
2123        bytes for UTF-8 encodings are set for characters greater than 127 when in
2124        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2125        altogether.)
2126    
2127    20. Added the /T option to pcretest so as to be able to run tests with non-
2128        standard character tables, thus making it possible to include the tests
2129        used for 19 above in the standard set of tests.
2130    
2131    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2132        reference to a subpattern the other side of a comment that contains an
2133        opening parenthesis caused either an internal compiling error, or a
2134        reference to the wrong subpattern.
2135    
2136    
2137    Version 8.02 19-Mar-2010
2138    ------------------------
2139    
2140    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2141    
2142    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2143        configured.
2144    
2145    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2146        original author of that file, following a query about its status.
2147    
2148    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2149        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2150    
2151    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2152        quantifier applied to a forward-referencing subroutine call, could compile
2153        incorrect code or give the error "internal error: previously-checked
2154        referenced subpattern not found".
2155    
2156    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2157        variables to point to external functions. For these systems, therefore,
2158        pcre_malloc etc. are now initialized to local functions that call the
2159        relevant global functions.
2160    
2161    7.  There were two entries missing in the vectors called coptable and poptable
2162        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2163        I've fixed the data, and added a kludgy way of testing at compile time that
2164        the lengths are correct (equal to the number of opcodes).
2165    
2166    8.  Following on from 7, I added a similar kludge to check the length of the
2167        eint vector in pcreposix.c.
2168    
2169    9.  Error texts for pcre_compile() are held as one long string to avoid too
2170        much relocation at load time. To find a text, the string is searched,
2171        counting zeros. There was no check for running off the end of the string,
2172        which could happen if a new error number was added without updating the
2173        string.
2174    
2175    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2176    
2177    11. \K was not working if it appeared in an atomic group or in a group that
2178        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2179        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2180        the assertion is positive, but not if it is negative.
2181    
2182    12. Change 11 fortuitously reduced the size of the stack frame used in the
2183        "match()" function of pcre_exec.c by one pointer. Forthcoming
2184        implementation of support for (*MARK) will need an extra pointer on the
2185        stack; I have reserved it now, so that the stack frame size does not
2186        decrease.
2187    
2188    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2189        item in branch that calls a recursion is a subroutine call - as in the
2190        second branch in the above example - was incorrectly given the compile-
2191        time error "recursive call could loop indefinitely" because pcre_compile()
2192        was not correctly checking the subroutine for matching a non-empty string.
2193    
2194    14. The checks for overrunning compiling workspace could trigger after an
2195        overrun had occurred. This is a "should never occur" error, but it can be
2196        triggered by pathological patterns such as hundreds of nested parentheses.
2197        The checks now trigger 100 bytes before the end of the workspace.
2198    
2199    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2200    
2201    
2202    Version 8.01 19-Jan-2010
2203    ------------------------
2204    
2205    1.  If a pattern contained a conditional subpattern with only one branch (in
2206        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2207        computed the wrong minimum data length (which is of course zero for such
2208        subpatterns). This could cause incorrect "no match" results.
2209    
2210    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2211        the pattern is reset in the first branch, pcre_compile() failed with
2212        "internal error: code overflow at offset...". This happened only when
2213        the reset was to the original external option setting. (An optimization
2214        abstracts leading options settings into an external setting, which was the
2215        cause of this.)
2216    
2217    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2218        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2219        assertion pattern did not match (meaning that the assertion was true), it
2220        was incorrectly treated as false if the SKIP had been reached during the
2221        matching. This also applied to assertions used as conditions.
2222    
2223    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2224        assertion subpattern, including such a pattern used as a condition,
2225        unpredictable results occurred, instead of the error return
2226        PCRE_ERROR_DFA_UITEM.
2227    
2228    5.  The C++ GlobalReplace function was not working like Perl for the special
2229        situation when an empty string is matched. It now does the fancy magic
2230        stuff that is necessary.
2231    
2232    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2233        removed. (These were left over from very, very early versions of PCRE.)
2234    
2235    7.  Some cosmetic changes to the code to make life easier when compiling it
2236        as part of something else:
2237    
2238        (a) Change DEBUG to PCRE_DEBUG.
2239    
2240        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2241            called "current" as "current_branch", to prevent a collision with the
2242            Linux macro when compiled as a kernel module.
2243    
2244        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2245            prevent a collision with the Linux macro when compiled as a kernel
2246            module.
2247    
2248    8.  In pcre_compile() there are some checks for integer overflows that used to
2249        cast potentially large values to (double). This has been changed to that
2250        when building, a check for int64_t is made, and if it is found, it is used
2251        instead, thus avoiding the use of floating point arithmetic. (There is no
2252        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2253        double.
2254    
2255    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2256        2005 (difference between two addresses compared to an unsigned value).
2257    
2258    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2259        custom one, because of the following reported problem in Windows:
2260    
2261          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2262              under Win32.
2263          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2264              therefore missing the function definition.
2265          - The compiler thus generates a "C" signature for the test function.
2266          - The linker fails to find the "C" function.
2267          - PCRE fails to configure if asked to do so against libbz2.
2268    
2269    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2270        messages were output:
2271    
2272          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2273          rerunning libtoolize, to keep the correct libtool macros in-tree.
2274          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2275    
2276        I have done both of these things.
2277    
2278    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2279        most of the time, it *can* run out if it is given a pattern that contains a
2280        runaway infinite recursion. I updated the discussion in the pcrestack man
2281        page.
2282    
2283    13. Now that we have gone to the x.xx style of version numbers, the minor
2284        version may start with zero. Using 08 or 09 is a bad idea because users
2285        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2286        interpreted as invalid octal numbers. I've updated the previous comment in
2287        configure.ac, and also added a check that gives an error if 08 or 09 are
2288        used.
2289    
2290    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2291        causing partial matching to fail when the end of the subject matched \W
2292        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2293    
2294    15. There were some discrepancies between the declarations in pcre_internal.h
2295        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2296        their definitions. The declarations used "const uschar *" and the
2297        definitions used USPTR. Even though USPTR is normally defined as "const
2298        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2299        reported that: "This difference in casting confuses some C++ compilers, for
2300        example, SunCC recognizes above declarations as different functions and
2301        generates broken code for hbpcre." I have changed the declarations to use
2302        USPTR.
2303    
2304    16. GNU libtool is named differently on some systems. The autogen.sh script now
2305        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2306        (FreeBSD).
2307    
2308    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2309        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2310        comment: "Figure out how to create a longlong from a string: strtoll and
2311        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2312        instance, but it only takes 2 args instead of 3!"
2313    
2314    18. A subtle bug concerned with back references has been fixed by a change of
2315        specification, with a corresponding code fix. A pattern such as
2316        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2317        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2318        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2319        same bug. Such groups have to be quantified to be useful, or contained
2320        inside another quantified group. (If there's no repetition, the reference
2321        can never match.) The problem arises because, having left the group and
2322        moved on to the rest of the pattern, a later failure that backtracks into
2323        the group uses the captured value from the final iteration of the group
2324        rather than the correct earlier one. I have fixed this in PCRE by forcing
2325        any group that contains a reference to itself to be an atomic group; that
2326        is, there cannot be any backtracking into it once it has completed. This is
2327        similar to recursive and subroutine calls.
2328    
2329    
2330    Version 8.00 19-Oct-09
2331    ----------------------
2332    
2333    1.  The table for translating pcre_compile() error codes into POSIX error codes
2334        was out-of-date, and there was no check on the pcre_compile() error code
2335        being within the table. This could lead to an OK return being given in
2336        error.
2337    
2338    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2339        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2340        in a Windows environment.
2341    
2342    3.  The pcregrep --count option prints the count for each file even when it is
2343        zero, as does GNU grep. However, pcregrep was also printing all files when
2344        --files-with-matches was added. Now, when both options are given, it prints
2345        counts only for those files that have at least one match. (GNU grep just
2346        prints the file name in this circumstance, but including the count seems
2347        more useful - otherwise, why use --count?) Also ensured that the
2348        combination -clh just lists non-zero counts, with no names.
2349    
2350    4.  The long form of the pcregrep -F option was incorrectly implemented as
2351        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2352        but it seems right to fix it, and I didn't think it was worth preserving
2353        the old behaviour.
2354    
2355    5.  The command line items --regex=pattern and --regexp=pattern were not
2356        recognized by pcregrep, which required --regex pattern or --regexp pattern
2357        (with a space rather than an '='). The man page documented the '=' forms,
2358        which are compatible with GNU grep; these now work.
2359    
2360    6.  No libpcreposix.pc file was created for pkg-config; there was just
2361        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2362    
2363    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2364        when UCP support is not needed, by modifying the Python script that
2365        generates it from Unicode data files. This should not matter if the module
2366        is correctly used as a library, but I received one complaint about 50K of
2367        unwanted data. My guess is that the person linked everything into his
2368        program rather than using a library. Anyway, it does no harm.
2369    
2370    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2371        was a minimum greater than 1 for a wide character in a possessive
2372        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2373        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2374        character. Chaos in the form of incorrect output or a compiling loop could
2375        result.
2376    
2377    9.  The restrictions on what a pattern can contain when partial matching is
2378        requested for pcre_exec() have been removed. All patterns can now be
2379        partially matched by this function. In addition, if there are at least two
2380        slots in the offset vector, the offset of the earliest inspected character
2381        for the match and the offset of the end of the subject are set in them when
2382        PCRE_ERROR_PARTIAL is returned.
2383    
2384    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2385        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2386        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2387        and may be more useful for multi-segment matching.
2388    
2389    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2390        used to be given if ever the end of the subject was reached; now it is
2391        given only if matching could not proceed because another character was
2392        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2393        string "Z", which now yields "no match" instead of "partial match". In the
2394        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2395        final character ended with (*FAIL).
2396    
2397    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2398        if the pattern had a "must contain" character that was already found in the
2399        earlier partial match, unless partial matching was again requested. For
2400        example, with the pattern /dog.(body)?/, the "must contain" character is
2401        "g". If the first part-match was for the string "dog", restarting with
2402        "sbody" failed. This bug has been fixed.
2403    
2404    13. The string returned by pcre_dfa_exec() after a partial match has been
2405        changed so that it starts at the first inspected character rather than the
2406        first character of the match. This makes a difference only if the pattern
2407        starts with a lookbehind assertion or \b or \B (\K is not supported by
2408        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2409        matching functions compatible, and I think it's the right thing to do.
2410    
2411    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2412        so that the demonstration program is easily available in environments where
2413        PCRE has not been installed from source.
2414    
2415    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2416        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2417        library.
2418    
2419    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2420        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2421        is not the first non-POSIX option to be added. Clearly some people find
2422        these options useful.
2423    
2424    17. If a caller to the POSIX matching function regexec() passes a non-zero
2425        value for nmatch with a NULL value for pmatch, the value of
2426        nmatch is forced to zero.
2427    
2428    18. RunGrepTest did not have a test for the availability of the -u option of
2429        the diff command, as RunTest does. It now checks in the same way as
2430        RunTest, and also checks for the -b option.
2431    
2432    19. If an odd number of negated classes containing just a single character
2433        interposed, within parentheses, between a forward reference to a named
2434        subpattern and the definition of the subpattern, compilation crashed with
2435        an internal error, complaining that it could not find the referenced
2436        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2437        [The bug was that it was starting one character too far in when skipping
2438        over the character class, thus treating the ] as data rather than
2439        terminating the class. This meant it could skip too much.]
2440    
2441    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2442        /g option in pcretest when the pattern contains \K, which makes it possible
2443        to have an empty string match not at the start, even when the pattern is
2444        anchored. Updated pcretest and pcredemo to use this option.
2445    
2446    21. If the maximum number of capturing subpatterns in a recursion was greater
2447        than the maximum at the outer level, the higher number was returned, but
2448        with unset values at the outer level. The correct (outer level) value is
2449        now given.
2450    
2451    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2452        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2453        make it do so. The string so far is captured, making this feature
2454        compatible with Perl.
2455    
2456    23. The tests have been re-organized, adding tests 11 and 12, to make it
2457        possible to check the Perl 5.10 features against Perl 5.10.
2458    
2459    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2460        pattern matches a fixed length string. PCRE did not allow this; now it
2461        does. Neither allows recursion.
2462    
2463    25. I finally figured out how to implement a request to provide the minimum
2464        length of subject string that was needed in order to match a given pattern.
2465        (It was back references and recursion that I had previously got hung up
2466        on.) This code has now been added to pcre_study(); it finds a lower bound
2467        to the length of subject needed. It is not necessarily the greatest lower
2468        bound, but using it to avoid searching strings that are too short does give
2469        some useful speed-ups. The value is available to calling programs via
2470        pcre_fullinfo().
2471    
2472    26. While implementing 25, I discovered to my embarrassment that pcretest had
2473        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2474        study optimizations had never been tested with that matching function.
2475        Oops. What is worse, even when it was passed study data, there was a bug in
2476        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2477        were also very few tests of studied patterns with pcre_dfa_exec().
2478    
2479    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2480        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2481        on the other side of the coin, they are no longer allowed to have different
2482        names, because these cannot be distinguished in PCRE, and this has caused
2483        confusion. (This is a difference from Perl.)
2484    
2485    28. When duplicate subpattern names are present (necessarily with different
2486        numbers, as required by 27 above), and a test is made by name in a
2487        conditional pattern, either for a subpattern having been matched, or for
2488        recursion in such a pattern, all the associated numbered subpatterns are
2489        tested, and the overall condition is true if the condition is true for any
2490        one of them. This is the way Perl works, and is also more like the way
2491        testing by number works.
2492    
2493    
2494    Version 7.9 11-Apr-09
2495    ---------------------
2496    
2497    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2498        (pcretest), all targets were linked against these libraries. This included
2499        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2500        libraries. This caused unwanted dependencies to be created. This problem
2501        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2502        pcretest is linked with readline.
2503    
2504    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2505        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2506        moved outside it again, because FALSE and TRUE are already defined in AIX,
2507        but BOOL is not.
2508    
2509    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2510        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2511    
2512    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2513        hyphens) following file names and line numbers when outputting matching
2514        lines. This is not true; no spaces are inserted. I have also clarified the
2515        wording for the --colour (or --color) option.
2516    
2517    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2518        was not coloured; this is different to GNU grep, so I have changed it to be
2519        the same.
2520    
2521    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2522        each matching line was coloured. Now it goes on to look for further matches
2523        of any of the test patterns, which is the same behaviour as GNU grep.
2524    
2525    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2526        doesn't make sense to accept an empty string match in pcregrep, so I have
2527        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2528        seems to be how GNU grep behaves. [But see later change 40 for release
2529        8.33.]
2530    
2531    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2532        start or after a newline", because the conditional assertion was not being
2533        correctly handled. The rule now is that both the assertion and what follows
2534        in the first alternative must satisfy the test.
2535    
2536    9.  If auto-callout was enabled in a pattern with a conditional group whose
2537        condition was an assertion, PCRE could crash during matching, both with
2538        pcre_exec() and pcre_dfa_exec().
2539    
2540    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2541        used for matching.
2542    
2543    11. Unicode property support in character classes was not working for
2544        characters (bytes) greater than 127 when not in UTF-8 mode.
2545    
2546    12. Added the -M command line option to pcretest.
2547    
2548    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2549    
2550    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2551    
2552    16. Added comments and documentation about mis-use of no_arg in the C++
2553        wrapper.
2554    
2555    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2556        from Martin Jerabek that uses macro names for all relevant character and
2557        string constants.
2558    
2559    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2560        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2561        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2562        these, but not everybody uses configure.
2563    
2564    19. A conditional group that had only one branch was not being correctly
2565        recognized as an item that could match an empty string. This meant that an
2566        enclosing group might also not be so recognized, causing infinite looping
2567        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2568        with the subject "ab", where knowledge that the repeated group can match
2569        nothing is needed in order to break the loop.
2570    
2571    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2572        exec(), but without supplying a callout function, matching went wrong.
2573    
2574    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2575        leak if the size of the offset vector was greater than 30. When the vector
2576        is smaller, the saved offsets during recursion go onto a local stack
2577        vector, but for larger vectors malloc() is used. It was failing to free
2578        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2579        error, in fact).
2580    
2581    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2582        heapframe that is used only when UTF-8 support is enabled. This caused no
2583        problem, but was untidy.
2584    
2585    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2586        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2587        included within another project.
2588    
2589    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2590        slightly modified by me:
2591    
2592          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2593              not building pcregrep.
2594    
2595          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2596              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2597    
2598    25. Forward references, both numeric and by name, in patterns that made use of
2599        duplicate group numbers, could behave incorrectly or give incorrect errors,
2600        because when scanning forward to find the reference group, PCRE was not
2601        taking into account the duplicate group numbers. A pattern such as
2602        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2603    
2604    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2605        the feature of a custom pointer more persuasive (as requested by a user).
2606    
2607    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2608        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2609        pre-defined.
2610    
2611    28. Added support for (*UTF8) at the start of a pattern.
2612    
2613    29. Arrange for flags added by the "release type" setting in CMake to be shown
2614        in the configuration summary.
2615    
2616    
2617    Version 7.8 05-Sep-08
2618    ---------------------
2619    
2620    1.  Replaced UCP searching code with optimized version as implemented for Ad
2621        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2622        stage table and inline lookup instead of a function, giving speed ups of 2
2623        to 5 times on some simple patterns that I tested. Permission was given to
2624        distribute the MultiStage2.py script that generates the tables (it's not in
2625        the tarball, but is in the Subversion repository).
2626    
2627    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2628        scripts.
2629    
2630    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2631        a group with a zero qualifier. The result of the study could be incorrect,
2632        or the function might crash, depending on the pattern.
2633    
2634    4.  Caseless matching was not working for non-ASCII characters in back
2635        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2636        It now works when Unicode Property Support is available.
2637    
2638    5.  In pcretest, an escape such as \x{de} in the data was always generating
2639        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2640        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2641        truncation.
2642    
2643    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2644    
2645    7.  Added two (int) casts to pcregrep when printing the difference of two
2646        pointers, in case they are 64-bit values.
2647    
2648    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2649        test 2 if it fails.
2650    
2651    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2652        and a #define of that name to empty if it is not externally set. This is to
2653        allow users of MSVC to set it if necessary.
2654    
2655    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2656        the convenience functions in the pcre_get.c source file.
2657    
2658    11. An option change at the start of a pattern that had top-level alternatives
2659        could cause overwriting and/or a crash. This command provoked a crash in
2660        some environments:
2661    
2662          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2663    
2664        This potential security problem was recorded as CVE-2008-2371.
2665    
2666    12. For a pattern where the match had to start at the beginning or immediately
2667        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2668        pcre_dfa_exec() could read past the end of the passed subject if there was
2669        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2670        pcretest so that it places the subject at the end of its malloc-ed buffer.
2671    
2672    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2673        exec() might read past the end of the data buffer in UTF-8 mode.
2674    
2675    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2676        the data contained the byte 0x85 as part of a UTF-8 character within its
2677        first line. This applied both to normal and DFA matching.
2678    
2679    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2680        /^[^d]*?$/8 failed to match "abc".
2681    
2682    16. Added a missing copyright notice to pcrecpp_internal.h.
2683    
2684    17. Make it more clear in the documentation that values returned from
2685        pcre_exec() in ovector are byte offsets, not character counts.
2686    
2687    18. Tidied a few places to stop certain compilers from issuing warnings.
2688    
2689    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2690        supplied by Stefan Weber. I made a further small update for 7.8 because
2691        there is a change of source arrangements: the pcre_searchfuncs.c module is
2692        replaced by pcre_ucd.c.
2693    
2694    
2695    Version 7.7 07-May-08
2696    ---------------------
2697    
2698    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2699        a string to a long long, pretend we don't even have a long long." This is
2700        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2701    
2702    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2703        pre-7.6 versions, which defined a global no_arg variable instead of putting
2704        it in the RE class. (See also #8 below.)
2705    
2706    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2707        Lopes.
2708    
2709    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2710    
2711        (1) The include/exclude patterns were being applied to the whole pathnames
2712            of files, instead of just to the final components.
2713    
2714        (2) If there was more than one level of directory, the subdirectories were
2715            skipped unless they satisfied the include/exclude conditions. This is
2716            inconsistent with GNU grep (and could even be seen as contrary to the
2717            pcregrep specification - which I improved to make it absolutely clear).
2718            The action now is always to scan all levels of directory, and just
2719            apply the include/exclude patterns to regular files.
2720    
2721    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2722        --exclude_dir in the tests to avoid scanning .svn directories.
2723    
2724    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2725        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2726        doesn't support NULs in patterns.
2727    
2728    7.  Added some missing "const"s to declarations of static tables in
2729        pcre_compile.c and pcre_dfa_exec.c.
2730    
2731    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2732        caused by fix #2  above. (Subsequently also a second patch to fix the
2733        first patch. And a third patch - this was a messy problem.)
2734    
2735    9.  Applied Craig's patch to remove the use of push_back().
2736    
2737    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2738        matching function regexec().
2739    
2740    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2741        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2742        references. PCRE supports relative numbers with this syntax (I don't think
2743        Oniguruma does).
2744    
2745    12. Previously, a group with a zero repeat such as (...){0} was completely
2746        omitted from the compiled regex. However, this means that if the group
2747        was called as a subroutine from elsewhere in the pattern, things went wrong
2748        (an internal error was given). Such groups are now left in the compiled
2749        pattern, with a new opcode that causes them to be skipped at execution
2750        time.
2751    
2752    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2753        to the way PCRE behaves:
2754    
2755        (a) A lone ] character is dis-allowed (Perl treats it as data).
2756    
2757        (b) A back reference to an unmatched subpattern matches an empty string
2758            (Perl fails the current match path).
2759    
2760        (c) A data ] in a character class must be notated as \] because if the
2761            first data character in a class is ], it defines an empty class. (In
2762            Perl it is not possible to have an empty class.) The empty class []
2763            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2764            The negative empty class [^] matches any one character, independently
2765            of the DOTALL setting.
2766    
2767    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2768        non-existent subpattern following a character class starting with ']' and
2769        containing () gave an internal compiling error instead of "reference to
2770        non-existent subpattern". Fortunately, when the pattern did exist, the
2771        compiled code was correct. (When scanning forwards to check for the
2772        existence of the subpattern, it was treating the data ']' as terminating
2773        the class, so got the count wrong. When actually compiling, the reference
2774        was subsequently set up correctly.)
2775    
2776    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2777        it was being rejected as not supported by pcre_dfa_exec(), even though
2778        other assertions are supported. I have made pcre_dfa_exec() support
2779        (*FAIL).
2780    
2781    16. The implementation of 13c above involved the invention of a new opcode,
2782        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2783        cannot be changed at match time, I realized I could make a small
2784        improvement to matching performance by compiling OP_ALLANY instead of
2785        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2786        on the OP_ANY path.
2787    
2788    17. Compiling pcretest on Windows with readline support failed without the
2789        following two fixes: (1) Make the unistd.h include conditional on
2790        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2791    
2792    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2793        ncurses library to be included for pcretest when ReadLine support is
2794        requested, but also to allow for it to be overridden. This patch came from
2795        Daniel Bergström.
2796    
2797    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2798        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2799        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2800        spotting this.
2801    
2802    
2803    Version 7.6 28-Jan-08
2804    ---------------------
2805    
2806    1.  A character class containing a very large number of characters with
2807        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2808        overflow.
2809    
2810    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2811        HAVE_LONG_LONG is not defined.
2812    
2813    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2814        bring them up to date and include new features. This patch includes:
2815    
2816        - Fixed PH's badly added libz and libbz2 support.
2817        - Fixed a problem with static linking.
2818        - Added pcredemo. [But later removed - see 7 below.]
2819        - Fixed dftables problem and added an option.
2820        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2821            HAVE_LONG_LONG.
2822        - Added readline support for pcretest.
2823        - Added an listing of the option settings after cmake has run.
2824    
2825    4.  A user submitted a patch to Makefile that makes it easy to create
2826        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2827        Makefile.am that cause it to include this special target, without
2828        affecting anything else. Note that the same mingw target plus all
2829        the other distribution libraries and programs are now supported
2830        when configuring with CMake (see 6 below) instead of with
2831        Configure/Make.
2832    
2833    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2834        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2835        exported in the Windows port". It has not yet been confirmed that the patch
2836        solves the problem, but it does no harm.
2837    
2838    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2839        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2840        with CMake, and also correct the comment about stack recursion.
2841    
2842    7.  Remove the automatic building of pcredemo from the ./configure system and
2843        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2844        of a program that users should build themselves after PCRE is installed, so
2845        building it automatically is not really right. What is more, it gave
2846        trouble in some build environments.
2847    
2848    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2849    
2850    
2851    Version 7.5 10-Jan-08
2852    ---------------------
2853    
2854    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2855        values in parens when parsing an RE using the C++ wrapper."
2856    
2857    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2858        Characters greater than 255 were excluded from the class instead of being
2859        included.
2860    
2861    3.  The same bug as (2) above applied to negated POSIX classes such as
2862        [:^space:].
2863    
2864    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2865        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2866        I have changed it.
2867    
2868    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2869        first named subpattern) and a construct such as (?&a) would reference the
2870        first named subpattern whose name started with "a" (in other words, the
2871        length check was missing). Both these problems are fixed. "Subpattern name
2872        expected" is now given for (?&) (a zero-length name), and this patch also
2873        makes it give the same error for \k'' (previously it complained that that
2874        was a reference to a non-existent subpattern).
2875    
2876    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2877        this is right because (?- can be followed by option settings as well as by
2878        digits. I have, however, made the messages clearer.
2879    
2880    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2881        than the number used in the conditional) now cause a compile-time error.
2882        This is actually not compatible with Perl, which accepts such patterns, but
2883        treats the conditional as always being FALSE (as PCRE used to), but it
2884        seems to me that giving a diagnostic is better.
2885    
2886    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2887        and messages.
2888    
2889    9.  Fix two occurrences of "backslash" in comments that should have been
2890        "backspace".
2891    
2892    10. Remove two redundant lines of code that can never be obeyed (their function
2893        was moved elsewhere).
2894    
2895    11. The program that makes PCRE's Unicode character property table had a bug
2896        which caused it to generate incorrect table entries for sequences of
2897        characters that have the same character type, but are in different scripts.
2898        It amalgamated them into a single range, with the script of the first of
2899        them. In other words, some characters were in the wrong script. There were
2900        thirteen such cases, affecting characters in the following ranges:
2901    
2902          U+002b0 - U+002c1
2903          U+0060c - U+0060d
2904          U+0061e - U+00612
2905          U+0064b - U+0065e
2906          U+0074d - U+0076d
2907          U+01800 - U+01805
2908          U+01d00 - U+01d77
2909          U+01d9b - U+01dbf
2910          U+0200b - U+0200f
2911          U+030fc - U+030fe
2912          U+03260 - U+0327f
2913          U+0fb46 - U+0fbb1
2914          U+10450 - U+1049d
2915    
2916    12. The -o option (show only the matching part of a line) for pcregrep was not
2917        compatible with GNU grep in that, if there was more than one match in a
2918        line, it showed only the first of them. It now behaves in the same way as
2919        GNU grep.
2920    
2921    13. If the -o and -v options were combined for pcregrep, it printed a blank
2922        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2923        does the same. The return code can be used to tell if there were any
2924        non-matching lines.
2925    
2926    14. Added --file-offsets and --line-offsets to pcregrep.
2927    
2928    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2929        infinitely looping recursion. The bug was that positive lookaheads were not
2930        being skipped when checking for a possible empty match (negative lookaheads
2931        and both kinds of lookbehind were skipped).
2932    
2933    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2934        inclusion of <windows.h> to before rather than after the definition of
2935        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2936    
2937    17. Specifying a possessive quantifier with a specific limit for a Unicode
2938        character property caused pcre_compile() to compile bad code, which led at
2939        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2940        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2941        caused the error; without that there was no problem.
2942    
2943    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2944    
2945    19. Added --enable-pcretest-libreadline.
2946    
2947    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2948        RE::GlobalReplace(). As a result, the number of replacements returned was
2949        double what it should be. I removed one of the increments, but Craig sent a
2950        later patch that removed the other one (the right fix) and added unit tests
2951        that check the return values (which was not done before).
2952    
2953    21. Several CMake things:
2954    
2955        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2956            the names libpcre and libpcreposix, not just pcre and pcreposix.
2957    
2958        (2) The above change means that pcretest and pcregrep are now correctly
2959            linked with the newly-built libraries, not previously installed ones.
2960    
2961        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2962    
2963    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2964        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2965        UTF-8 newline character). The key issue is that the pattern starts .*;
2966        this means that the match must be either at the beginning, or after a
2967        newline. The bug was in the code for advancing after a failed match and
2968        checking that the new position followed a newline. It was not taking
2969        account of UTF-8 characters correctly.
2970    
2971    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2972        character classes. PCRE was not treating the sequence [:...:] as a
2973        character class unless the ... were all letters. Perl, however, seems to
2974        allow any characters between [: and :], though of course it rejects as
2975        unknown any "names" that contain non-letters, because all the known class
2976        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2977        for example, whereas PCRE did not - it did not recognize a POSIX character
2978        class. This seemed a bit dangerous, so the code has been changed to be
2979        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2980        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2981        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2982        Perl does, and where it didn't before.
2983    
2984    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2985        Windows environments %n is disabled by default.
2986    
2987    
2988    Version 7.4 21-Sep-07
2989    ---------------------
2990    
2991    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2992        means that a class such as [\s] counted as "explicit reference to CR or
2993        LF". That isn't really right - the whole point of the change was to try to
2994        help when there was an actual mention of one of the two characters. So now
2995        the change happens only if \r or \n (or a literal CR or LF) character is
2996        encountered.
2997    
2998    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2999        of both had grown to the point where there were only 3 bits left.
3000        Fortunately, there was spare space in the data structure, and so I have
3001        moved the internal flags into a new 16-bit field to free up more option
3002        bits.
3003    
3004    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3005        but did not set the internal JCHANGED flag - either of these is enough to
3006        control the way the "get" function works - but the PCRE_INFO_JCHANGED
3007        facility is supposed to tell if (?J) was ever used, so now (?J) at the
3008        start sets both bits.
3009    
3010    4.  Added options (at build time, compile time, exec time) to change \R from
3011        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3012    
3013    5.  doc/pcresyntax.html was missing from the distribution.
3014    
3015    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3016        compatibility, even though it is no longer used.
3017    
3018    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3019        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3020        windows.h file is present (where different names are used). [This was
3021        reversed later after testing - see 16 below.]
3022    
3023    8.  Changed all #include <config.h> to #include "config.h". There were also
3024        some further <pcre.h> cases that I changed to "pcre.h".
3025    
3026    9.  When pcregrep was used with the --colour option, it missed the line ending
3027        sequence off the lines that it output.
3028    
3029    10. It was pointed out to me that arrays of string pointers cause lots of
3030        relocations when a shared library is dynamically loaded. A technique of
3031        using a single long string with a table of offsets can drastically reduce
3032        these. I have refactored PCRE in four places to do this. The result is
3033        dramatic:
3034    
3035          Originally:                          290
3036          After changing UCP table:            187
3037          After changing error message table:   43
3038          After changing table of "verbs"       36
3039          After changing table of Posix names   22
3040    
3041        Thanks to the folks working on Gregex for glib for this insight.
3042    
3043    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3044        unicode-properties was also set.
3045    
3046    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3047    
3048    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3049        checked only for CRLF.
3050    
3051    14. Added casts to pcretest.c to avoid compiler warnings.
3052    
3053    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3054    
3055    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3056        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3057        entirely. This removes changes made in 7 above.
3058    
3059    17. The CMake files have been updated, and there is now more information about
3060        building with CMake in the NON-UNIX-USE document.
3061    
3062    
3063    Version 7.3 28-Aug-07
3064    ---------------------
3065    
3066     1. In the rejigging of the build system that eventually resulted in 7.1, the
3067        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3068        brackets there is not right, since it causes compilers to look for an
3069        installed pcre.h, not the version that is in the source that is being
3070        compiled (which of course may be different). I have changed it back to:
3071    
3072          #include "pcre.h"
3073    
3074        I have a vague recollection that the change was concerned with compiling in
3075        different directories, but in the new build system, that is taken care of
3076        by the VPATH setting the Makefile.
3077    
3078     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3079        when the subject happened to end in the byte 0x85 (e.g. if the last
3080        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3081        characters but of course it shouldn't be taken as a newline when it is part
3082        of another character. The bug was that, for an unlimited repeat of . in
3083        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3084        characters when looking for a newline.
3085    
3086     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3087    
3088     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3089        in debug output.
3090    
3091     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3092        long printing in the pcrecpp unittest when running under MinGW.
3093    
3094     6. ESC_K was left out of the EBCDIC table.
3095    
3096     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3097        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3098        limit also applies to "virtual nesting" when a pattern is recursive, and in
3099        this case 1000 isn't so big. I have been able to remove this limit at the
3100        expense of backing off one optimization in certain circumstances. Normally,
3101        when pcre_exec() would call its internal match() function recursively and
3102        immediately return the result unconditionally, it uses a "tail recursion"
3103        feature to save stack. However, when a subpattern that can match an empty
3104        string has an unlimited repetition quantifier, it no longer makes this
3105        optimization. That gives it a stack frame in which to save the data for
3106        checking that an empty string has been matched. Previously this was taken
3107        from the 1000-entry workspace that had been reserved. So now there is no
3108        explicit limit, but more stack is used.
3109    
3110     8. Applied Daniel's patches to solve problems with the import/export magic
3111        syntax that is required for Windows, and which was going wrong for the
3112        pcreposix and pcrecpp parts of the library. These were overlooked when this
3113        problem was solved for the main library.
3114    
3115     9. There were some crude static tests to avoid integer overflow when computing
3116        the size of patterns that contain repeated groups with explicit upper
3117        limits. As the maximum quantifier is 65535, the maximum group length was
3118        set at 30,000 so that the product of these two numbers did not overflow a
3119        32-bit integer. However, it turns out that people want to use groups that
3120        are longer than 30,000 bytes (though not repeat them that many times).
3121        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3122        made it possible to implement the integer overflow checks in a much more
3123        dynamic way, which I have now done. The artificial limitation on group
3124        length has been removed - we now have only the limit on the total length of
3125        the compiled pattern, which depends on the LINK_SIZE setting.
3126    
3127    10. Fixed a bug in the documentation for get/copy named substring when
3128        duplicate names are permitted. If none of the named substrings are set, the
3129        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3130        empty string.
3131    
3132    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3133        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3134        because the ] is interpreted as the first data character and the
3135        terminating ] is not found. PCRE has been made compatible with Perl in this
3136        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3137        cause memory overwriting.
3138    
3139    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3140        string has been matched (to stop an infinite loop). It was not recognizing
3141        a conditional subpattern that could match an empty string if that
3142        subpattern was within another subpattern. For example, it looped when
3143        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3144        condition was not nested. This bug has been fixed.
3145    
3146    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3147        past the start of the subject in the presence of bytes with the top bit
3148        set, for example "\x8aBCD".
3149    
3150    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3151        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3152    
3153    14. Optimized (?!) to (*FAIL).
3154    
3155    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3156        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3157        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3158        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3159        does: it's just the validity check that is more restrictive.
3160    
3161    16. Inserted checks for integer overflows during escape sequence (backslash)
3162        processing, and also fixed erroneous offset values for syntax errors during
3163        backslash processing.
3164    
3165    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3166        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3167    
3168    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3169        caused an overrun.
3170    
3171    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3172        something other than just ASCII characters) inside a group that had an
3173        unlimited repeat caused a loop at compile time (while checking to see
3174        whether the group could match an empty string).
3175    
3176    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3177        [\P{Any}] did so. (Error in the code for printing property names.)
3178    
3179    21. An orphan \E inside a character class could cause a crash.
3180    
3181    22. A repeated capturing bracket such as (A)? could cause a wild memory
3182        reference during compilation.
3183    
3184    23. There are several functions in pcre_compile() that scan along a compiled
3185        expression for various reasons (e.g. to see if it's fixed length for look
3186        behind). There were bugs in these functions when a repeated \p or \P was
3187        present in the pattern. These operators have additional parameters compared
3188        with \d, etc, and these were not being taken into account when moving along
3189        the compiled data. Specifically:
3190    
3191        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3192            length.
3193    
3194        (b) An item such as \pL+ within a repeated group could cause crashes or
3195            loops.
3196    
3197        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3198            "reference to non-existent subpattern" error.
3199    
3200        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3201    
3202    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3203        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3204    
3205    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3206    
3207    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3208        character were causing crashes (broken optimization).
3209    
3210    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3211        \p or \P) caused a compile-time loop.
3212    
3213    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3214        break. For example, the unstudied pattern [\r\n]A does not match the string
3215        "\r\nA" because change 7.0/46 below moves the current point on by two
3216        characters after failing to match at the start. However, the pattern \nA
3217        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3218        the same is true. There doesn't seem any very clean way out of this, but
3219        what I have chosen to do makes the common cases work: PCRE now takes note
3220        of whether there can be an explicit match for \r or \n anywhere in the
3221        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3222        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3223        pattern has explicit CR or LF references.
3224    
3225    29. Added (*CR) etc for changing newline setting at start of pattern.
3226    
3227    
3228    Version 7.2 19-Jun-07
3229  ---------------------  ---------------------
3230    
3231   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3245  Version 7.2 01-May-07
3245      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3246      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3247      in all cases.      in all cases.
3248    
3249   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3250    
3251      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3252          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3253    
3254      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3255          to be opened parentheses.          to be opened parentheses.
3256    
3257      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3258          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3259    
3260        (d) \K resets the start of the current match so that everything before
3261            is not part of it.
3262    
3263        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3264    
3265        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3266            reference syntax.
3267    
3268        (g) (?| introduces a group in which the numbering of parentheses in each
3269            alternative starts with the same number.
3270    
3271        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3272    
3273     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3274        PCRE_INFO_JCHANGED.
3275    
3276     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3277        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3278        for detecting groups that can match an empty string.
3279    
3280     9. A pattern with a very large number of alternatives (more than several
3281        hundred) was running out of internal workspace during the pre-compile
3282        phase, where pcre_compile() figures out how much memory will be needed. A
3283        bit of new cunning has reduced the workspace needed for groups with
3284        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3285        workspace instead of running out of the 4096 that are available.
3286    
3287    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3288    
3289    11. Applied patch from Google to remove an optimization that didn't quite work.
3290        The report of the bug said:
3291    
3292          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3293          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3294          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3295    
3296    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3297        it matched the wrong number of bytes.
3298    
3299    
3300  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3301  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1648

  ViewVC Help
Powered by ViewVC 1.1.5