/[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 1628 by ph10, Tue Feb 9 09:46:52 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    
54    Version 8.38 23-November-2015
55    -----------------------------
56    
57    1.  If a group that contained a recursive back reference also contained a
58        forward reference subroutine call followed by a non-forward-reference
59        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
60        compile correct code, leading to undefined behaviour or an internally
61        detected error. This bug was discovered by the LLVM fuzzer.
62    
63    2.  Quantification of certain items (e.g. atomic back references) could cause
64        incorrect code to be compiled when recursive forward references were
65        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
66        This bug was discovered by the LLVM fuzzer.
67    
68    3.  A repeated conditional group whose condition was a reference by name caused
69        a buffer overflow if there was more than one group with the given name.
70        This bug was discovered by the LLVM fuzzer.
71    
72    4.  A recursive back reference by name within a group that had the same name as
73        another group caused a buffer overflow. For example:
74        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
75    
76    5.  A forward reference by name to a group whose number is the same as the
77        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
78        a buffer overflow at compile time. This bug was discovered by the LLVM
79        fuzzer.
80    
81    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
82        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
83    
84    7.  Another buffer overflow bug involved duplicate named groups with a
85        reference between their definition, with a group that reset capture
86        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
87        fixed by always allowing for more memory, even if not needed. (A proper fix
88        is implemented in PCRE2, but it involves more refactoring.)
89    
90    8.  There was no check for integer overflow in subroutine calls such as (?123).
91    
92    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
93        being treated as a literal 'l' instead of causing an error.
94    
95    10. There was a buffer overflow if pcre_exec() was called with an ovector of
96        size 1. This bug was found by american fuzzy lop.
97    
98    11. If a non-capturing group containing a conditional group that could match
99        an empty string was repeated, it was not identified as matching an empty
100        string itself. For example: /^(?:(?(1)x|)+)+$()/.
101    
102    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
103        \a and \e in test subject lines.
104    
105    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
106        instead of the EBCDIC value.
107    
108    14. The handling of \c in an EBCDIC environment has been revised so that it is
109        now compatible with the specification in Perl's perlebcdic page.
110    
111    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
112        ASCII/Unicode. This has now been added to the list of characters that are
113        recognized as white space in EBCDIC.
114    
115    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
116        error (correctly) when used outside a class, but did not give an error
117        within a class.
118    
119    17. \h within a class was incorrectly compiled in EBCDIC environments.
120    
121    18. A pattern with an unmatched closing parenthesis that contained a backward
122        assertion which itself contained a forward reference caused buffer
123        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
124    
125    19. JIT should return with error when the compiled pattern requires more stack
126        space than the maximum.
127    
128    20. A possessively repeated conditional group that could match an empty string,
129        for example, /(?(R))*+/, was incorrectly compiled.
130    
131    21. Fix infinite recursion in the JIT compiler when certain patterns such as
132        /(?:|a|){100}x/ are analysed.
133    
134    22. Some patterns with character classes involving [: and \\ were incorrectly
135        compiled and could cause reading from uninitialized memory or an incorrect
136        error diagnosis.
137    
138    23. Pathological patterns containing many nested occurrences of [: caused
139        pcre_compile() to run for a very long time.
140    
141    24. A conditional group with only one branch has an implicit empty alternative
142        branch and must therefore be treated as potentially matching an empty
143        string.
144    
145    25. If (?R was followed by - or + incorrect behaviour happened instead of a
146        diagnostic.
147    
148    26. Arrange to give up on finding the minimum matching length for overly
149        complex patterns.
150    
151    27. Similar to (4) above: in a pattern with duplicated named groups and an
152        occurrence of (?| it is possible for an apparently non-recursive back
153        reference to become recursive if a later named group with the relevant
154        number is encountered. This could lead to a buffer overflow. Wen Guanxing
155        from Venustech ADLAB discovered this bug.
156    
157    28. If pcregrep was given the -q option with -c or -l, or when handling a
158        binary file, it incorrectly wrote output to stdout.
159    
160    29. The JIT compiler did not restore the control verb head in case of *THEN
161        control verbs. This issue was found by Karl Skomski with a custom LLVM
162        fuzzer.
163    
164    30. Error messages for syntax errors following \g and \k were giving inaccurate
165        offsets in the pattern.
166    
167    31. Added a check for integer overflow in conditions (?(<digits>) and
168        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
169        fuzzer.
170    
171    32. Handling recursive references such as (?2) when the reference is to a group
172        later in the pattern uses code that is very hacked about and error-prone.
173        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
174        give an internal error if it is obvious that compiling has gone wrong.
175    
176    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
177        This issue was found by Karl Skomski with a custom LLVM fuzzer.
178    
179    34. The JIT compiler should restore the control chain for empty possessive
180        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
181    
182    35. Match limit check added to JIT recursion. This issue was found by Karl
183        Skomski with a custom LLVM fuzzer.
184    
185    36. Yet another case similar to 27 above has been circumvented by an
186        unconditional allocation of extra memory. This issue is fixed "properly" in
187        PCRE2 by refactoring the way references are handled. Wen Guanxing
188        from Venustech ADLAB discovered this bug.
189    
190    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
191        with a custom LLVM fuzzer.
192    
193    38. Fixed a corner case of range optimization in JIT.
194    
195    39. An incorrect error "overran compiling workspace" was given if there were
196        exactly enough group forward references such that the last one extended
197        into the workspace safety margin. The next one would have expanded the
198        workspace. The test for overflow was not including the safety margin.
199    
200    40. A match limit issue is fixed in JIT which was found by Karl Skomski
201        with a custom LLVM fuzzer.
202    
203    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
204        work under Windows. (Why has it taken so long for anyone to notice?)
205    
206    42. In a character class such as [\W\p{Any}] where both a negative-type escape
207        ("not a word character") and a property escape were present, the property
208        escape was being ignored.
209    
210    43. Fix crash caused by very long (*MARK) or (*THEN) names.
211    
212    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
213        by a single ASCII character in a class item, was incorrectly compiled in
214        UCP mode. The POSIX class got lost, but only if the single character
215        followed it.
216    
217    45. [:punct:] in UCP mode was matching some characters in the range 128-255
218        that should not have been matched.
219    
220    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
221        class, all characters with code points greater than 255 are in the class.
222        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
223        such as \w are turned into Unicode properties), wide characters were not
224        correctly handled, and could fail to match.
225    
226    
227    Version 8.37 28-April-2015
228    --------------------------
229    
230    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
231        for those parentheses to be closed with whatever has been captured so far.
232        However, it was failing to mark any other groups between the hightest
233        capture so far and the currrent group as "unset". Thus, the ovector for
234        those groups contained whatever was previously there. An example is the
235        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
236    
237    2.  If an assertion condition was quantified with a minimum of zero (an odd
238        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
239    
240    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
241        unrecognized modifier, a crash could occur.
242    
243    4.  An attempt to do global matching in pcretest with a zero-length ovector
244        caused a crash.
245    
246    5.  Fixed a memory leak during matching that could occur for a subpattern
247        subroutine call (recursive or otherwise) if the number of captured groups
248        that had to be saved was greater than ten.
249    
250    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
251        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
252        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
253    
254    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
255        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
256    
257    8. When a pattern is compiled, it remembers the highest back reference so that
258       when matching, if the ovector is too small, extra memory can be obtained to
259       use instead. A conditional subpattern whose condition is a check on a
260       capture having happened, such as, for example in the pattern
261       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
262       setting the highest backreference number. This mattered only if pcre_exec()
263       was called with an ovector that was too small to hold the capture, and there
264       was no other kind of back reference (a situation which is probably quite
265       rare). The effect of the bug was that the condition was always treated as
266       FALSE when the capture could not be consulted, leading to a incorrect
267       behaviour by pcre_exec(). This bug has been fixed.
268    
269    9. A reference to a duplicated named group (either a back reference or a test
270       for being set in a conditional) that occurred in a part of the pattern where
271       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
272       to be incorrectly calculated, leading to overwriting.
273    
274    10. A mutually recursive set of back references such as (\2)(\1) caused a
275        segfault at study time (while trying to find the minimum matching length).
276        The infinite loop is now broken (with the minimum length unset, that is,
277        zero).
278    
279    11. If an assertion that was used as a condition was quantified with a minimum
280        of zero, matching went wrong. In particular, if the whole group had
281        unlimited repetition and could match an empty string, a segfault was
282        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
283        assertions to be quantified, but not if they are being used as conditions,
284        so the above pattern is faulted by Perl. PCRE has now been changed so that
285        it also rejects such patterns.
286    
287    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
288        failed to allow the zero-repeat case if pcre2_exec() was called with an
289        ovector too small to capture the group.
290    
291    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
292        Red Hat Product Security:
293    
294        (a) A crash if /K and /F were both set with the option to save the compiled
295        pattern.
296    
297        (b) Another crash if the option to print captured substrings in a callout
298        was combined with setting a null ovector, for example \O\C+ as a subject
299        string.
300    
301    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
302        forward reference repeated a large (but limited) number of times within a
303        repeated outer group that has a zero minimum quantifier, caused incorrect
304        code to be compiled, leading to the error "internal error:
305        previously-checked referenced subpattern not found" when an incorrect
306        memory address was read. This bug was reported as "heap overflow",
307        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
308        CVE-2015-2325.
309    
310    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
311        call within a group that also contained a recursive back reference caused
312        incorrect code to be compiled. This bug was reported as "heap overflow",
313        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
314        number CVE-2015-2326.
315    
316    24. Computing the size of the JIT read-only data in advance has been a source
317        of various issues, and new ones are still appear unfortunately. To fix
318        existing and future issues, size computation is eliminated from the code,
319        and replaced by on-demand memory allocation.
320    
321    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
322        adjacent to the end of the range, and the range contained characters with
323        more than one other case, caused incorrect behaviour when compiled in UTF
324        mode. In that example, the range a-j was left out of the class.
325    
326    26. Fix JIT compilation of conditional blocks, which assertion
327        is converted to (*FAIL). E.g: /(?(?!))/.
328    
329    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
330        discovered by the LLVM fuzzer.
331    
332    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
333        when this assertion was used as a condition, for example (?(?!)a|b). In
334        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
335        error about an unsupported item.
336    
337    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
338        possessification code could take exponential time to complete. A recursion
339        depth limit of 1000 has been imposed to limit the resources used by this
340        optimization.
341    
342    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
343        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
344        because \S ensures they are all in the class. The code for doing this was
345        interacting badly with the code for computing the amount of space needed to
346        compile the pattern, leading to a buffer overflow. This bug was discovered
347        by the LLVM fuzzer.
348    
349    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
350        other kinds of group caused stack overflow at compile time. This bug was
351        discovered by the LLVM fuzzer.
352    
353    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
354        between a subroutine call and its quantifier was incorrectly compiled,
355        leading to buffer overflow or other errors. This bug was discovered by the
356        LLVM fuzzer.
357    
358    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
359        assertion after (?(. The code was failing to check the character after
360        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
361        was discovered by the LLVM fuzzer.
362    
363    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
364        a fixed maximum following a group that contains a subroutine reference was
365        incorrectly compiled and could trigger buffer overflow. This bug was
366        discovered by the LLVM fuzzer.
367    
368    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
369        caused a stack overflow instead of the diagnosis of a non-fixed length
370        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
371    
372    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
373        (e.g. /(?<=\Ka)/) could make pcregrep loop.
374    
375    37. There was a similar problem to 36 in pcretest for global matches.
376    
377    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
378        and a subsequent item in the pattern caused a non-match, backtracking over
379        the repeated \X did not stop, but carried on past the start of the subject,
380        causing reference to random memory and/or a segfault. There were also some
381        other cases where backtracking after \C could crash. This set of bugs was
382        discovered by the LLVM fuzzer.
383    
384    39. The function for finding the minimum length of a matching string could take
385        a very long time if mutual recursion was present many times in a pattern,
386        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
387        method has been implemented. This infelicity was discovered by the LLVM
388        fuzzer.
389    
390    40. Static linking against the PCRE library using the pkg-config module was
391        failing on missing pthread symbols.
392    
393    
394    Version 8.36 26-September-2014
395    ------------------------------
396    
397    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
398        -Wmissing-field-initializers and -Wunused-parameter.
399    
400    2.  The tests for quantifiers being too big (greater than 65535) were being
401        applied after reading the number, and stupidly assuming that integer
402        overflow would give a negative number. The tests are now applied as the
403        numbers are read.
404    
405    3.  Tidy code in pcre_exec.c where two branches that used to be different are
406        now the same.
407    
408    4.  The JIT compiler did not generate match limit checks for certain
409        bracketed expressions with quantifiers. This may lead to exponential
410        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
411        issue should be resolved now.
412    
413    5.  Fixed an issue, which occures when nested alternatives are optimized
414        with table jumps.
415    
416    6.  Inserted two casts and changed some ints to size_t in the light of some
417        reported 64-bit compiler warnings (Bugzilla 1477).
418    
419    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
420        an empty string, which sometimes were behaving incorrectly in the
421        interpreter (though correctly in the JIT matcher). This pcretest input is
422        an example:
423    
424          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
425          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
426    
427        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
428        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
429        for an empty string was breaking the inner loop and carrying on at a lower
430        level, when possessive repeated groups should always return to a higher
431        level as they have no backtrack points in them. The empty string test now
432        occurs at the outer level.
433    
434    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
435        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
436    
437    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
438        doesn't).
439    
440    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
441        to hold a pattern to be incorrectly computed (too small) when there were
442        named back references to duplicated names. This could cause "internal
443        error: code overflow" or "double free or corruption" or other memory
444        handling errors.
445    
446    11. When named subpatterns had the same prefixes, back references could be
447        confused. For example, in this pattern:
448    
449          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
450    
451        the reference to 'Name' was incorrectly treated as a reference to a
452        duplicate name.
453    
454    12. A pattern such as /^s?c/mi8 where the optional character has more than
455        one "other case" was incorrectly compiled such that it would only try to
456        match starting at "c".
457    
458    13. When a pattern starting with \s was studied, VT was not included in the
459        list of possible starting characters; this should have been part of the
460        8.34/18 patch.
461    
462    14. If a character class started [\Qx]... where x is any character, the class
463        was incorrectly terminated at the ].
464    
465    15. If a pattern that started with a caseless match for a character with more
466        than one "other case" was studied, PCRE did not set up the starting code
467        unit bit map for the list of possible characters. Now it does. This is an
468        optimization improvement, not a bug fix.
469    
470    16. The Unicode data tables have been updated to Unicode 7.0.0.
471    
472    17. Fixed a number of memory leaks in pcregrep.
473    
474    18. Avoid a compiler warning (from some compilers) for a function call with
475        a cast that removes "const" from an lvalue by using an intermediate
476        variable (to which the compiler does not object).
477    
478    19. Incorrect code was compiled if a group that contained an internal recursive
479        back reference was optional (had quantifier with a minimum of zero). This
480        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
481        caused segmentation faults because of stack overflows at compile time.
482    
483    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
484        group that is quantified with an indefinite repeat, caused a compile-time
485        loop which used up all the system stack and provoked a segmentation fault.
486        This was not the same bug as 19 above.
487    
488    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
489        Patch by Mike Frysinger.
490    
491    
492    Version 8.35 04-April-2014
493    --------------------------
494    
495    1.  A new flag is set, when property checks are present in an XCLASS.
496        When this flag is not set, PCRE can perform certain optimizations
497        such as studying these XCLASS-es.
498    
499    2.  The auto-possessification of character sets were improved: a normal
500        and an extended character set can be compared now. Furthermore
501        the JIT compiler optimizes more character set checks.
502    
503    3.  Got rid of some compiler warnings for potentially uninitialized variables
504        that show up only when compiled with -O2.
505    
506    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
507        of a match later then the end of the match. The pcretest program was not
508        handling the case sensibly - it was outputting from the start to the next
509        binary zero. It now reports this situation in a message, and outputs the
510        text from the end to the start.
511    
512    5.  Fast forward search is improved in JIT. Instead of the first three
513        characters, any three characters with fixed position can be searched.
514        Search order: first, last, middle.
515    
516    6.  Improve character range checks in JIT. Characters are read by an inprecise
517        function now, which returns with an unknown value if the character code is
518        above a certain threshold (e.g: 256). The only limitation is that the value
519        must be bigger than the threshold as well. This function is useful when
520        the characters above the threshold are handled in the same way.
521    
522    7.  The macros whose names start with RAWUCHAR are placeholders for a future
523        mode in which only the bottom 21 bits of 32-bit data items are used. To
524        make this more memorable for those maintaining the code, the names have
525        been changed to start with UCHAR21, and an extensive comment has been added
526        to their definition.
527    
528    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
529        to the export list in Makefile.am (they were accidentally omitted from the
530        8.34 tarball).
531    
532    9.  The informational output from pcretest used the phrase "starting byte set"
533        which is inappropriate for the 16-bit and 32-bit libraries. As the output
534        for "first char" and "need char" really means "non-UTF-char", I've changed
535        "byte" to "char", and slightly reworded the output. The documentation about
536        these values has also been (I hope) clarified.
537    
538    10. Another JIT related optimization: use table jumps for selecting the correct
539        backtracking path, when more than four alternatives are present inside a
540        bracket.
541    
542    11. Empty match is not possible, when the minimum length is greater than zero,
543        and there is no \K in the pattern. JIT should avoid empty match checks in
544        such cases.
545    
546    12. In a caseless character class with UCP support, when a character with more
547        than one alternative case was not the first character of a range, not all
548        the alternative cases were added to the class. For example, s and \x{17f}
549        are both alternative cases for S: the class [RST] was handled correctly,
550        but [R-T] was not.
551    
552    13. The configure.ac file always checked for pthread support when JIT was
553        enabled. This is not used in Windows, so I have put this test inside a
554        check for the presence of windows.h (which was already tested for).
555    
556    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
557        The algorithm provides a way to skip certain starting offsets, and usually
558        faster than linear prefix searches.
559    
560    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
561        as for 'fr_FR' and 'french'. For some reason, however, it then used the
562        Windows-specific input and output files, which have 'french' screwed in.
563        So this could never have worked. One of the problems with locales is that
564        they aren't always the same. I have now updated RunTest so that it checks
565        the output of the locale test (test 3) against three different output
566        files, and it allows the test to pass if any one of them matches. With luck
567        this should make the test pass on some versions of Solaris where it was
568        failing. Because of the uncertainty, the script did not used to stop if
569        test 3 failed; it now does. If further versions of a French locale ever
570        come to light, they can now easily be added.
571    
572    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
573        there was a message during ./configure, but it did not stop. This now
574        provokes an error. The invalid example in README has been corrected.
575        If a value less than the minimum is given, the minimum value has always
576        been used, but now a warning is given.
577    
578    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
579        was a bug in the test system, which is now fixed. Also, the list of various
580        configurations that are tested for each release did not have one with both
581        16/32 bits and --enable-bar-anycrlf. It now does.
582    
583    18. pcretest was missing "-C bsr" for displaying the \R default setting.
584    
585    19. Little endian PowerPC systems are supported now by the JIT compiler.
586    
587    20. The fast forward newline mechanism could enter to an infinite loop on
588        certain invalid UTF-8 input. Although we don't support these cases
589        this issue can be fixed by a performance optimization.
590    
591    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
592        not take account if existing stack usage. There is now a new global
593        variable called pcre_stack_guard that can be set to point to an external
594        function to check stack availability. It is called at the start of
595        processing every parenthesized group.
596    
597    22. A typo in the code meant that in ungreedy mode the max/min qualifier
598        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
599        not match "ab".
600    
601    23. When UTF was disabled, the JIT program reported some incorrect compile
602        errors. These messages are silenced now.
603    
604    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
605        compiler.
606    
607    25. Change all the temporary files used in RunGrepTest to be different to those
608        used by RunTest so that the tests can be run simultaneously, for example by
609        "make -j check".
610    
611    
612    Version 8.34 15-December-2013
613    -----------------------------
614    
615    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
616        executable memory. Patch inspired by Carsten Klein.
617    
618    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
619        this macro is never tested and has no effect, because the work to support
620        coverage involves only compiling and linking options and special targets in
621        the Makefile. The comment in config.h implied that defining the macro would
622        enable coverage support, which is totally false. There was also support for
623        setting this macro in the CMake files (my fault, I just copied it from
624        configure). SUPPORT_GCOV has now been removed.
625    
626    3.  Make a small performance improvement in strlen16() and strlen32() in
627        pcretest.
628    
629    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
630        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
631        these cases). There was also one in pcretest.c.
632    
633    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
634    
635    6.  In UTF mode, the code for checking whether a group could match an empty
636        string (which is used for indefinitely repeated groups to allow for
637        breaking an infinite loop) was broken when the group contained a repeated
638        negated single-character class with a character that occupied more than one
639        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
640        UTF-8 mode). The effect was undefined: the group might or might not be
641        deemed as matching an empty string, or the program might have crashed.
642    
643    7.  The code for checking whether a group could match an empty string was not
644        recognizing that \h, \H, \v, \V, and \R must match a character.
645    
646    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
647        an empty string. If it can, pcretest shows this in its information output.
648    
649    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
650        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
651        matched by pcre_exec() without using JIT:
652    
653        (a) If the rest of the pattern did not match after a maximal run of
654            grapheme clusters, the code for backing up to try with fewer of them
655            did not always back up over a full grapheme when characters that do not
656            have the modifier quality were involved, e.g. Hangul syllables.
657    
658        (b) If the match point in a subject started with modifier character, and
659            there was no match, the code could incorrectly back up beyond the match
660            point, and potentially beyond the first character in the subject,
661            leading to a segfault or an incorrect match result.
662    
663    10. A conditional group with an assertion condition could lead to PCRE
664        recording an incorrect first data item for a match if no other first data
665        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
666        first data item, and therefore matched "ca" after "c" instead of at the
667        start.
668    
669    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
670        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
671    
672    12. The source of pcregrep now includes z/OS-specific code so that it can be
673        compiled for z/OS as part of the special z/OS distribution.
674    
675    13. Added the -T and -TM options to pcretest.
676    
677    14. The code in pcre_compile.c for creating the table of named capturing groups
678        has been refactored. Instead of creating the table dynamically during the
679        actual compiling pass, the information is remembered during the pre-compile
680        pass (on the stack unless there are more than 20 named groups, in which
681        case malloc() is used) and the whole table is created before the actual
682        compile happens. This has simplified the code (it is now nearly 150 lines
683        shorter) and prepared the way for better handling of references to groups
684        with duplicate names.
685    
686    15. A back reference to a named subpattern when there is more than one of the
687        same name now checks them in the order in which they appear in the pattern.
688        The first one that is set is used for the reference. Previously only the
689        first one was inspected. This change makes PCRE more compatible with Perl.
690    
691    16. Unicode character properties were updated from Unicode 6.3.0.
692    
693    17. The compile-time code for auto-possessification has been refactored, based
694        on a patch by Zoltan Herczeg. It now happens after instead of during
695        compilation. The code is cleaner, and more cases are handled. The option
696        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
697        options in pcretest are provided to set it. It can also be set by
698        (*NO_AUTO_POSSESS) at the start of a pattern.
699    
700    18. The character VT has been added to the default ("C" locale) set of
701        characters that match \s and are generally treated as white space,
702        following this same change in Perl 5.18. There is now no difference between
703        "Perl space" and "POSIX space". Whether VT is treated as white space in
704        other locales depends on the locale.
705    
706    19. The code for checking named groups as conditions, either for being set or
707        for being recursed, has been refactored (this is related to 14 and 15
708        above). Processing unduplicated named groups should now be as fast at
709        numerical groups, and processing duplicated groups should be faster than
710        before.
711    
712    20. Two patches to the CMake build system, by Alexander Barkov:
713    
714          (1) Replace the "source" command by "." in CMakeLists.txt because
715              "source" is a bash-ism.
716    
717          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
718              without these the CMake build does not work on Solaris.
719    
720    21. Perl has changed its handling of \8 and \9. If there is no previously
721        encountered capturing group of those numbers, they are treated as the
722        literal characters 8 and 9 instead of a binary zero followed by the
723        literals. PCRE now does the same.
724    
725    22. Following Perl, added \o{} to specify codepoints in octal, making it
726        possible to specify values greater than 0777 and also making them
727        unambiguous.
728    
729    23. Perl now gives an error for missing closing braces after \x{... instead of
730        treating the string as literal. PCRE now does the same.
731    
732    24. RunTest used to grumble if an inappropriate test was selected explicitly,
733        but just skip it when running all tests. This make it awkward to run ranges
734        of tests when one of them was inappropriate. Now it just skips any
735        inappropriate tests, as it always did when running all tests.
736    
737    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
738        character types such as \d or \w, too many callouts were inserted, and the
739        data that they returned was rubbish.
740    
741    26. In UCP mode, \s was not matching two of the characters that Perl matches,
742        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
743        were matched by \h. The code has now been refactored so that the lists of
744        the horizontal and vertical whitespace characters used for \h and \v (which
745        are defined only in one place) are now also used for \s.
746    
747    27. Add JIT support for the 64 bit TileGX architecture.
748        Patch by Jiong Wang (Tilera Corporation).
749    
750    28. Possessive quantifiers for classes (both explicit and automatically
751        generated) now use special opcodes instead of wrapping in ONCE brackets.
752    
753    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
754        (because it's meaningless), this was not happening when PCRE_CASELESS was
755        set. Not wrong, but inefficient.
756    
757    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
758        properties for \w, \d, etc) is present in a test regex. Otherwise if the
759        test contains no characters greater than 255, Perl doesn't realise it
760        should be using Unicode semantics.
761    
762    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
763        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
764        does in Unicode mode.
765    
766    32. Added the "forbid" facility to pcretest so that putting tests into the
767        wrong test files can sometimes be quickly detected.
768    
769    33. There is now a limit (default 250) on the depth of nesting of parentheses.
770        This limit is imposed to control the amount of system stack used at compile
771        time. It can be changed at build time by --with-parens-nest-limit=xxx or
772        the equivalent in CMake.
773    
774    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
775        errors. Perl warns for these when in warning mode, but PCRE has no facility
776        for giving warnings.
777    
778    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
779        However, this was not working for (?!) because it is optimized to (*FAIL),
780        for which PCRE does not allow quantifiers. The optimization is now disabled
781        when a quantifier follows (?!). I can't see any use for this, but it makes
782        things uniform.
783    
784    36. Perl no longer allows group names to start with digits, so I have made this
785        change also in PCRE. It simplifies the code a bit.
786    
787    37. In extended mode, Perl ignores spaces before a + that indicates a
788        possessive quantifier. PCRE allowed a space before the quantifier, but not
789        before the possessive +. It now does.
790    
791    38. The use of \K (reset reported match start) within a repeated possessive
792        group such as (a\Kb)*+ was not working.
793    
794    40. Document that the same character tables must be used at compile time and
795        run time, and that the facility to pass tables to pcre_exec() and
796        pcre_dfa_exec() is for use only with saved/restored patterns.
797    
798    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
799        features for Builds with MSVC:
800    
801        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
802           builds). The .rc files can be used to set FileDescription and many other
803           attributes.
804    
805        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
806           This allows higher-level build scripts which want .pdb files to avoid
807           hard-coding the exact files needed."
808    
809    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
810        mean "start of word" and "end of word", respectively, as a transition aid.
811    
812    43. A minimizing repeat of a class containing codepoints greater than 255 in
813        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
814        compiled to use the heap for recursion.
815    
816    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
817        is configured.
818    
819    
820    Version 8.33 28-May-2013
821    ------------------------
822    
823    1.  Added 'U' to some constants that are compared to unsigned integers, to
824        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
825        variables that are added to signed variables, to ensure the result is
826        signed and can be negated.
827    
828    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
829        CMake config files.
830    
831    3.  Revise the creation of config.h.generic so that all boolean macros are
832        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
833        overriding via -D on the command line possible.
834    
835    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
836        to unsigned int is reported to make a quite noticeable speed difference in
837        a specific Windows environment. Testing on Linux did also appear to show
838        some benefit (and it is clearly not harmful). Also fixed the definition of
839        Xop which should be unsigned.
840    
841    5.  Related to (4), changing the definition of the intermediate variable cc
842        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
843        improvements.
844    
845    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
846        unnecessary spaces.
847    
848    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
849        and later.
850    
851    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
852    
853    9.  Optimizing fast_forward_start_bits in JIT.
854    
855    10. Adding support for callouts in JIT, and fixing some issues revealed
856        during this work. Namely:
857    
858        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
859    
860        (b) Minimum length was not checked before the matching is started.
861    
862    11. The value of capture_last that is passed to callouts was incorrect in some
863        cases when there was a capture on one path that was subsequently abandoned
864        after a backtrack. Also, the capture_last value is now reset after a
865        recursion, since all captures are also reset in this case.
866    
867    12. The interpreter no longer returns the "too many substrings" error in the
868        case when an overflowing capture is in a branch that is subsequently
869        abandoned after a backtrack.
870    
871    13. In the pathological case when an offset vector of size 2 is used, pcretest
872        now prints out the matched string after a yield of 0 or 1.
873    
874    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
875        Only supported by the JIT compiler at the moment.
876    
877    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
878    
879    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
880        offset of the starting point of the matching process, provided the offsets
881        vector is large enough.
882    
883    17. The \A escape now records a lookbehind value of 1, though its execution
884        does not actually inspect the previous character. This is to ensure that,
885        in partial multi-segment matching, at least one character from the old
886        segment is retained when a new segment is processed. Otherwise, if there
887        are no lookbehinds in the pattern, \A might match incorrectly at the start
888        of a new segment.
889    
890    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
891    
892    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
893        gives some modest performance improvement in 8-bit mode.
894    
895    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
896        be expressed in certain programming languages using Universal Character
897        Names.
898    
899    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
900        which points out that "non characters" are not "characters that may not
901        appear in Unicode strings" but rather "characters that are reserved for
902        internal use and have only local meaning".
903    
904    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
905        there was a conditional group that depended on an assertion, if the
906        assertion was false, the callout that immediately followed the alternation
907        in the condition was skipped when pcre_exec() was used for matching.
908    
909    23. Allow an explicit callout to be inserted before an assertion that is the
910        condition for a conditional group, for compatibility with automatic
911        callouts, which always insert a callout at this point.
912    
913    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
914        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
915    
916    25. (*PRUNE) is now supported by the JIT compiler.
917    
918    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
919    
920    27. Fix the case where there are two or more SKIPs with arguments that may be
921        ignored.
922    
923    28. (*SKIP) is now supported by the JIT compiler.
924    
925    29. (*THEN) is now supported by the JIT compiler.
926    
927    30. Update RunTest with additional test selector options.
928    
929    31. The way PCRE handles backtracking verbs has been changed in two ways.
930    
931        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
932        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
933        backtracking. In some cases this makes it more Perl-compatible, but Perl's
934        rather obscure rules do not always do the same thing.
935    
936        (2) Previously, backtracking verbs were confined within assertions. This is
937        no longer the case for positive assertions, except for (*ACCEPT). Again,
938        this sometimes improves Perl compatibility, and sometimes does not.
939    
940    32. A number of tests that were in test 2 because Perl did things differently
941        have been moved to test 1, because either Perl or PCRE has changed, and
942        these tests are now compatible.
943    
944    32. Backtracking control verbs are now handled in the same way in JIT and
945        interpreter.
946    
947    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
948        contained a forward subroutine reference caused a compile error.
949    
950    34. Auto-detect and optimize limited repetitions in JIT.
951    
952    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
953        blocking (*UTF) etc.
954    
955    36. In the interpreter, maximizing pattern repetitions for characters and
956        character types now use tail recursion, which reduces stack usage.
957    
958    37. The value of the max lookbehind was not correctly preserved if a compiled
959        and saved regex was reloaded on a host of different endianness.
960    
961    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
962        of the compiled pattern block, expand the flags field from 16 to 32 bits
963        because it was almost full.
964    
965    39. Try madvise first before posix_madvise.
966    
967    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
968        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
969        original change locked out all matches of empty strings. This has been
970        changed so that one match of an empty string per line is recognized.
971        Subsequent searches on the same line (for colouring or for --only-matching,
972        for example) do not recognize empty strings.
973    
974    41. Applied a user patch to fix a number of spelling mistakes in comments.
975    
976    42. Data lines longer than 65536 caused pcretest to crash.
977    
978    43. Clarified the data type for length and startoffset arguments for pcre_exec
979        and pcre_dfa_exec in the function-specific man pages, where they were
980        explicitly stated to be in bytes, never having been updated. I also added
981        some clarification to the pcreapi man page.
982    
983    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
984        a segmentation fault.
985    
986    
987    Version 8.32 30-November-2012
988    -----------------------------
989    
990    1.  Improved JIT compiler optimizations for first character search and single
991        character iterators.
992    
993    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
994        Patch by Daniel Richard G.
995    
996    3.  Single character iterator optimizations in the JIT compiler.
997    
998    4.  Improved JIT compiler optimizations for character ranges.
999    
1000    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1001        Reported by Giuseppe D'Angelo.
1002    
1003    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1004        of a line, was being set incorrectly in cases where .* appeared inside
1005        atomic brackets at the start of a pattern, or where there was a subsequent
1006        *PRUNE or *SKIP.
1007    
1008    7.  Improved instruction cache flush for POWER/PowerPC.
1009        Patch by Daniel Richard G.
1010    
1011    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1012        grep:
1013    
1014        (a) There is now no limit to the number of patterns to be matched.
1015    
1016        (b) An error is given if a pattern is too long.
1017    
1018        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1019            are now supported.
1020    
1021        (d) --exclude-from and --include-from (multiple use) have been added.
1022    
1023        (e) Exclusions and inclusions now apply to all files and directories, not
1024            just to those obtained from scanning a directory recursively.
1025    
1026        (f) Multiple uses of -f and --file-list are now supported.
1027    
1028        (g) In a Windows environment, the default for -d has been changed from
1029            "read" (the GNU grep default) to "skip", because otherwise the presence
1030            of a directory in the file list provokes an error.
1031    
1032        (h) The documentation has been revised and clarified in places.
1033    
1034    9.  Improve the matching speed of capturing brackets.
1035    
1036    10. Changed the meaning of \X so that it now matches a Unicode extended
1037        grapheme cluster.
1038    
1039    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1040        out POSIX threads when JIT support is configured.
1041    
1042    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1043    
1044    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1045        structure could go wrong in environments where size_t is not the same size
1046        as int.
1047    
1048    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1049        set.
1050    
1051    15. The EBCDIC support had decayed; later updates to the code had included
1052        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1053        general tidy up of EBCDIC-related issues, and the documentation was also
1054        not quite right. There is now a test that can be run on ASCII systems to
1055        check some of the EBCDIC-related things (but is it not a full test).
1056    
1057    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1058        in a small tidy to the code.
1059    
1060    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1061    
1062    18. If the --only-matching (-o) option in pcregrep is specified multiple
1063        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1064        substrings matched by the 1st and 2nd capturing parentheses. A separating
1065        string can be specified by --om-separator (default empty).
1066    
1067    19. Improving the first n character searches.
1068    
1069    20. Turn case lists for horizontal and vertical white space into macros so that
1070        they are defined only once.
1071    
1072    21. This set of changes together give more compatible Unicode case-folding
1073        behaviour for characters that have more than one other case when UCP
1074        support is available.
1075    
1076        (a) The Unicode property table now has offsets into a new table of sets of
1077            three or more characters that are case-equivalent. The MultiStage2.py
1078            script that generates these tables (the pcre_ucd.c file) now scans
1079            CaseFolding.txt instead of UnicodeData.txt for character case
1080            information.
1081    
1082        (b) The code for adding characters or ranges of characters to a character
1083            class has been abstracted into a generalized function that also handles
1084            case-independence. In UTF-mode with UCP support, this uses the new data
1085            to handle characters with more than one other case.
1086    
1087        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1088            whose other case is greater than 256 are now correctly matched
1089            caselessly. Previously, the high codepoint matched the low one, but not
1090            vice versa.
1091    
1092        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1093            of the new class addition function, using character lists defined as
1094            macros alongside the case definitions of 20 above.
1095    
1096        (e) Caseless back references now work with characters that have more than
1097            one other case.
1098    
1099        (f) General caseless matching of characters with more than one other case
1100            is supported.
1101    
1102    22. Unicode character properties were updated from Unicode 6.2.0
1103    
1104    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1105    
1106    24. Add support for 32-bit character strings, and UTF-32
1107    
1108    25. Major JIT compiler update (code refactoring and bugfixing).
1109        Experimental Sparc 32 support is added.
1110    
1111    26. Applied a modified version of Daniel Richard G's patch to create
1112        pcre.h.generic and config.h.generic by "make" instead of in the
1113        PrepareRelease script.
1114    
1115    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1116        pcre_compile.c when checking for a zero character.
1117    
1118    28. Introducing a native interface for JIT. Through this interface, the compiled
1119        machine code can be directly executed. The purpose of this interface is to
1120        provide fast pattern matching, so several sanity checks are not performed.
1121        However, feature tests are still performed. The new interface provides
1122        1.4x speedup compared to the old one.
1123    
1124    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1125        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1126        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1127    
1128    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1129        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1130        the "old" RFC 2279). Instead, it ended up passing a negative length to
1131        pcre_exec().
1132    
1133    31. Add support for GCC's visibility feature to hide internal functions.
1134    
1135    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1136        "unknown -C option" after outputting 0 or 1.
1137    
1138    33. There is now support for generating a code coverage report for the test
1139        suite in environments where gcc is the compiler and lcov is installed. This
1140        is mainly for the benefit of the developers.
1141    
1142    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1143        unaddressable using valgrind annotations, allowing valgrind to detect
1144        invalid memory accesses. This is mainly for the benefit of the developers.
1145    
1146    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1147    
1148    26. Give configure error if --enable-cpp but no C++ compiler found.
1149    
1150    
1151    Version 8.31 06-July-2012
1152    -------------------------
1153    
1154    1.  Fixing a wrong JIT test case and some compiler warnings.
1155    
1156    2.  Removed a bashism from the RunTest script.
1157    
1158    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1159        to unsigned type, result still unsigned" that was given by an MS compiler
1160        on encountering the code "-sizeof(xxx)".
1161    
1162    4.  Partial matching support is added to the JIT compiler.
1163    
1164    5.  Fixed several bugs concerned with partial matching of items that consist
1165        of more than one character:
1166    
1167        (a) /^(..)\1/ did not partially match "aba" because checking references was
1168            done on an "all or nothing" basis. This also applied to repeated
1169            references.
1170    
1171        (b) \R did not give a hard partial match if \r was found at the end of the
1172            subject.
1173    
1174        (c) \X did not give a hard partial match after matching one or more
1175            characters at the end of the subject.
1176    
1177        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1178            a partial match for the string "\r".
1179    
1180        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1181            a partial match for a CR character at the end of the subject string.
1182    
1183    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1184        running pcretest, the text "(JIT)" added to the output whenever JIT is
1185        actually used to run the match.
1186    
1187    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1188        or /S+[+] with a digit between 1 and 7.
1189    
1190    8.  OP_NOT now supports any UTF character not just single-byte ones.
1191    
1192    9.  (*MARK) control verb is now supported by the JIT compiler.
1193    
1194    10. The command "./RunTest list" lists the available tests without actually
1195        running any of them. (Because I keep forgetting what they all are.)
1196    
1197    11. Add PCRE_INFO_MAXLOOKBEHIND.
1198    
1199    12. Applied a (slightly modified) user-supplied patch that improves performance
1200        when the heap is used for recursion (compiled with --disable-stack-for-
1201        recursion). Instead of malloc and free for each heap frame each time a
1202        logical recursion happens, frames are retained on a chain and re-used where
1203        possible. This sometimes gives as much as 30% improvement.
1204    
1205    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1206        call.
1207    
1208    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1209    
1210    15. It is now possible to link pcretest with libedit as an alternative to
1211        libreadline.
1212    
1213    16. (*COMMIT) control verb is now supported by the JIT compiler.
1214    
1215    17. The Unicode data tables have been updated to Unicode 6.1.0.
1216    
1217    18. Added --file-list option to pcregrep.
1218    
1219    19. Added binary file support to pcregrep, including the -a, --binary-files,
1220        -I, and --text options.
1221    
1222    20. The madvise function is renamed for posix_madvise for QNX compatibility
1223        reasons. Fixed by Giuseppe D'Angelo.
1224    
1225    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1226    
1227    22. Changed the option for creating symbolic links for 16-bit man pages from
1228        -s to -sf so that re-installing does not cause issues.
1229    
1230    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1231    
1232    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1233        matches in certain environments (the workspace was not being correctly
1234        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1235        some of the workspace data at the beginning of a restart.
1236    
1237    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1238        was not doing so when it should - probably a typo introduced by SVN 528
1239        (change 8.10/14).
1240    
1241    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1242        \w+ when the character tables indicated that \x{c4} was a word character.
1243        There were several related cases, all because the tests for doing a table
1244        lookup were testing for characters less than 127 instead of 255.
1245    
1246    27. If a pattern contains capturing parentheses that are not used in a match,
1247        their slots in the ovector are set to -1. For those that are higher than
1248        any matched groups, this happens at the end of processing. In the case when
1249        there were back references that the ovector was too small to contain
1250        (causing temporary malloc'd memory to be used during matching), and the
1251        highest capturing number was not used, memory off the end of the ovector
1252        was incorrectly being set to -1. (It was using the size of the temporary
1253        memory instead of the true size.)
1254    
1255    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1256        ovector size, it uses memory at the end of the block that it has got.
1257    
1258    29. Check for an overlong MARK name and give an error at compile time. The
1259        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1260    
1261    30. JIT compiler update.
1262    
1263    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1264        Rill for the patch.
1265    
1266    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1267    
1268    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1269    
1270    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1271        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1272        was enabled.
1273    
1274    35. Improve JIT code generation for greedy plus quantifier.
1275    
1276    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1277        1 to "aa" instead of to an empty string. The bug affected repeated groups
1278        that could potentially match an empty string.
1279    
1280    37. Optimizing single character iterators in JIT.
1281    
1282    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1283        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1284        codepoints that are too big for the mode are faulted, and in a UTF mode,
1285        disallowed codepoints are also faulted.
1286    
1287    39. If PCRE was compiled with UTF support, in three places in the DFA
1288        matcher there was code that should only have been obeyed in UTF mode, but
1289        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1290        processing when bytes with values greater than 127 were present. In 16-bit
1291        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1292        both cases the values are those that cannot be the first data item in a UTF
1293        character. The three items that might have provoked this were recursions,
1294        possessively repeated groups, and atomic groups.
1295    
1296    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1297        and pcregrep, because some OS require shared objects to be explicitly
1298        passed to ld, causing the link step to fail if they are not.
1299    
1300    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1301        mode, patterns that started with \h* or \R* might be incorrectly matched.
1302    
1303    
1304    Version 8.30 04-February-2012
1305    -----------------------------
1306    
1307    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1308        name is defined in ctype.h.
1309    
1310    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1311        only in quite long subpatterns.
1312    
1313    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1314        since it was replaced by pcre_fullinfo() in February 2000.
1315    
1316    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1317        match a (*MARK), and the match failed at the start of the subject, a
1318        reference to memory before the start of the subject could occur. This bug
1319        was introduced by fix 17 of release 8.21.
1320    
1321    5.  A reference to an unset group with zero minimum repetition was giving
1322        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1323        /(another)?(\1?)test/ matched against "hello world test". This bug was
1324        introduced in release 8.13.
1325    
1326    6.  Add support for 16-bit character strings (a large amount of work involving
1327        many changes and refactorings).
1328    
1329    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1330        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1331        from a file.
1332    
1333    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1334        rounding is not applied in this particular case).
1335    
1336    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1337        if they appear, or are escaped, in patterns.
1338    
1339    10. Get rid of a number of -Wunused-but-set-variable warnings.
1340    
1341    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1342        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1343        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1344        also returns the mark "x". This bug applied to capturing parentheses,
1345        non-capturing parentheses, and atomic parentheses. It also applied to some
1346        assertions.
1347    
1348    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1349        information out of configure.ac instead of relying on pcre.h.generic, which
1350        is not stored in the repository.
1351    
1352    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1353        -lreadline.
1354    
1355    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1356    
1357    15. Applied Graycode's patch to put the top-level frame on the stack rather
1358        than the heap when not using the stack for recursion. This gives a
1359        performance improvement in many cases when recursion is not deep.
1360    
1361    16. Experimental code added to "pcretest -C" to output the stack frame size.
1362    
1363    
1364    Version 8.21 12-Dec-2011
1365    ------------------------
1366    
1367    1.  Updating the JIT compiler.
1368    
1369    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1370        are added as well.
1371    
1372    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1373        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1374        calling _pcre_jit_exec. Some extra comments are added.
1375    
1376    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1377        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1378        was introduced by change 18 for 8.20.
1379    
1380    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1381        ECMA-262 standard.
1382    
1383    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1384        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1385        This bug was probably introduced by change 9 of 8.13.
1386    
1387    7.  While fixing 6 above, I noticed that a number of other items were being
1388        incorrectly rejected as "not fixed length". This arose partly because newer
1389        opcodes had not been added to the fixed-length checking code. I have (a)
1390        corrected the bug and added tests for these items, and (b) arranged for an
1391        error to occur if an unknown opcode is encountered while checking for fixed
1392        length instead of just assuming "not fixed length". The items that were
1393        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1394        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1395        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1396    
1397    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1398        being incorrectly compiled and would have given unpredicatble results.
1399    
1400    9.  A possessively repeated subpattern with minimum repeat count greater than
1401        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1402        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1403        the first (A) could occur when it should not.
1404    
1405    10. Add a cast and remove a redundant test from the code.
1406    
1407    11. JIT should use pcre_malloc/pcre_free for allocation.
1408    
1409    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1410        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1411        is anything other than /usr, -L is still shown).
1412    
1413    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1414    
1415    14. Perl does not support \N without a following name in a [] class; PCRE now
1416        also gives an error.
1417    
1418    15. If a forward reference was repeated with an upper limit of around 2000,
1419        it caused the error "internal error: overran compiling workspace". The
1420        maximum number of forward references (including repeats) was limited by the
1421        internal workspace, and dependent on the LINK_SIZE. The code has been
1422        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1423        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1424        of around 200,000 forward references. While doing this, I also speeded up
1425        the filling in of repeated forward references.
1426    
1427    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1428        incorrectly expecting the subject to contain another "a" after the start.
1429    
1430    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1431        in the match, the SKIP should be ignored. This was not happening; instead
1432        the SKIP was being treated as NOMATCH. For patterns such as
1433        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1434        tested.
1435    
1436    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1437        now much more compatible with Perl, in particular in cases where the result
1438        is a non-match for a non-anchored pattern. For example, if
1439        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1440        "m", where previously it did not return a name. A side effect of this
1441        change is that for partial matches, the last encountered mark name is
1442        returned, as for non matches. A number of tests that were previously not
1443        Perl-compatible have been moved into the Perl-compatible test files. The
1444        refactoring has had the pleasing side effect of removing one argument from
1445        the match() function, thus reducing its stack requirements.
1446    
1447    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1448        subsequent uses of /S (without +) incorrectly behaved like /S+.
1449    
1450    21. Retrieve executable code size support for the JIT compiler and fixing
1451        some warnings.
1452    
1453    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1454        not work when the shorter character appeared right at the end of the
1455        subject string.
1456    
1457    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1458        systems.
1459    
1460    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1461        output it when the /M option is used in pcretest.
1462    
1463    25. The CheckMan script was not being included in the distribution. Also, added
1464        an explicit "perl" to run Perl scripts from the PrepareRelease script
1465        because this is reportedly needed in Windows.
1466    
1467    26. If study data was being save in a file and studying had not found a set of
1468        "starts with" bytes for the pattern, the data written to the file (though
1469        never used) was taken from uninitialized memory and so caused valgrind to
1470        complain.
1471    
1472    27. Updated RunTest.bat as provided by Sheri Pierce.
1473    
1474    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1475    
1476    29. Computation of memory usage for the table of capturing group names was
1477        giving an unnecessarily large value.
1478    
1479    
1480    Version 8.20 21-Oct-2011
1481    ------------------------
1482    
1483    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1484        a POSIX class. After further experiments with Perl, which convinced me that
1485        Perl has bugs and confusions, a closing square bracket is no longer allowed
1486        in a POSIX name. This bug also affected patterns with classes that started
1487        with full stops.
1488    
1489    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1490        captured substring, but while checking the failing first alternative,
1491        substring 1 is temporarily captured. If the output vector supplied to
1492        pcre_exec() was not big enough for this capture, the yield of the function
1493        was still zero ("insufficient space for captured substrings"). This cannot
1494        be totally fixed without adding another stack variable, which seems a lot
1495        of expense for a edge case. However, I have improved the situation in cases
1496        such as /(a)(b)x|abc/ matched against "abc", where the return code
1497        indicates that fewer than the maximum number of slots in the ovector have
1498        been set.
1499    
1500    3.  Related to (2) above: when there are more back references in a pattern than
1501        slots in the output vector, pcre_exec() uses temporary memory during
1502        matching, and copies in the captures as far as possible afterwards. It was
1503        using the entire output vector, but this conflicts with the specification
1504        that only 2/3 is used for passing back captured substrings. Now it uses
1505        only the first 2/3, for compatibility. This is, of course, another edge
1506        case.
1507    
1508    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1509        main code base, and can be used by building with --enable-jit. When this is
1510        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1511        runtime --no-jit option is given.
1512    
1513    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1514        ovector, the return from the function was zero, implying that there were
1515        other matches that did not fit. The correct "exactly full" value is now
1516        returned.
1517    
1518    6.  If a subpattern that was called recursively or as a subroutine contained
1519        (*PRUNE) or any other control that caused it to give a non-standard return,
1520        invalid errors such as "Error -26 (nested recursion at the same subject
1521        position)" or even infinite loops could occur.
1522    
1523    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1524        computing the minimum length on reaching *ACCEPT, and so ended up with the
1525        wrong value of 1 rather than 0. Further investigation indicates that
1526        computing a minimum subject length in the presence of *ACCEPT is difficult
1527        (think back references, subroutine calls), and so I have changed the code
1528        so that no minimum is registered for a pattern that contains *ACCEPT.
1529    
1530    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1531        it was not handled as intended. [But see 16 below.]
1532    
1533    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1534        Sheri Pierce.
1535    
1536    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1537        the first byte in a match must be "a".
1538    
1539    11. Change 17 for 8.13 increased the recursion depth for patterns like
1540        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1541        pattern contains any instances of (*THEN). If it does not, the old
1542        optimizations are restored. It would be nice to do this on a per-group
1543        basis, but at the moment that is not feasible.
1544    
1545    12. In some environments, the output of pcretest -C is CRLF terminated. This
1546        broke RunTest's code that checks for the link size. A single white space
1547        character after the value is now allowed for.
1548    
1549    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1550        For "fr", it uses the Windows-specific input and output files.
1551    
1552    14. If (*THEN) appeared in a group that was called recursively or as a
1553        subroutine, it did not work as intended. [But see next item.]
1554    
1555    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1556        pattern fragments (but not containing any | characters). If A and B are
1557        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1558        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1559        D. In other words, Perl considers parentheses that do not contain any |
1560        characters to be part of a surrounding alternative, whereas PCRE was
1561        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1562        differently. PCRE now behaves in the same way as Perl, except in the case
1563        of subroutine/recursion calls such as (?1) which have in any case always
1564        been different (but PCRE had them first :-).
1565    
1566    16. Related to 15 above: Perl does not treat the | in a conditional group as
1567        creating alternatives. Such a group is treated in the same way as an
1568        ordinary group without any | characters when processing (*THEN). PCRE has
1569        been changed to match Perl's behaviour.
1570    
1571    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1572        RunGrepTest script failed.
1573    
1574    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1575        inevitable for groups that contain captures, but it can lead to a lot of
1576        stack use in large patterns. The old behaviour has been restored for atomic
1577        groups that do not contain any capturing parentheses.
1578    
1579    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1580        suppress the check for a minimum subject length at run time. (If it was
1581        given to pcre_exec() or pcre_dfa_exec() it did work.)
1582    
1583    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1584        fail to work when decoding hex characters in data strings in EBCDIC
1585        environments.
1586    
1587    21. It appears that in at least one Mac OS environment, the isxdigit() function
1588        is implemented as a macro that evaluates to its argument more than once,
1589        contravening the C 90 Standard (I haven't checked a later standard). There
1590        was an instance in pcretest which caused it to go wrong when processing
1591        \x{...} escapes in subject strings. The has been rewritten to avoid using
1592        things like p++ in the argument of isxdigit().
1593    
1594    
1595    Version 8.13 16-Aug-2011
1596    ------------------------
1597    
1598    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1599    
1600    2.  Two minor typos in pcre_internal.h have been fixed.
1601    
1602    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1603        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1604        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1605    
1606    4.  There were a number of related bugs in the code for matching backrefences
1607        caselessly in UTF-8 mode when codes for the characters concerned were
1608        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1609        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1610        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1611        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1612        2-byte code at the end of the subject (it thought there wasn't enough data
1613        left).
1614    
1615    5.  Comprehensive information about what went wrong is now returned by
1616        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1617        as the output vector has at least 2 elements. The offset of the start of
1618        the failing character and a reason code are placed in the vector.
1619    
1620    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1621        now returned is for the first byte of the failing character, instead of the
1622        last byte inspected. This is an incompatible change, but I hope it is small
1623        enough not to be a problem. It makes the returned offset consistent with
1624        pcre_exec() and pcre_dfa_exec().
1625    
1626    7.  pcretest now gives a text phrase as well as the error number when
1627        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1628        failure, the offset and reason code are output.
1629    
1630    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1631        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1632        back over a single character (\n). This seems wrong (because it treated the
1633        two characters as a single entity when going forwards), conflicts with the
1634        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1635        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1636        has been changed.
1637    
1638    9.  Some internal refactoring has changed the processing so that the handling
1639        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1640        time (the PCRE_DOTALL option was changed this way some time ago: version
1641        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1642        which was always a bit of a fudge. It also means that there is one less
1643        argument for the match() function, which reduces its stack requirements
1644        slightly. This change also fixes an incompatibility with Perl: the pattern
1645        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1646    
1647    10. More internal refactoring has drastically reduced the number of recursive
1648        calls to match() for possessively repeated groups such as (abc)++ when
1649        using pcre_exec().
1650    
1651    11. While implementing 10, a number of bugs in the handling of groups were
1652        discovered and fixed:
1653    
1654        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1655        (a|)*(?1) gave a compile-time internal error.
1656        ((a|)+)+  did not notice that the outer group could match an empty string.
1657        (^a|^)+   was not marked as anchored.
1658        (.*a|.*)+ was not marked as matching at start or after a newline.
1659    
1660    12. Yet more internal refactoring has removed another argument from the match()
1661        function. Special calls to this function are now indicated by setting a
1662        value in a variable in the "match data" data block.
1663    
1664    13. Be more explicit in pcre_study() instead of relying on "default" for
1665        opcodes that mean there is no starting character; this means that when new
1666        ones are added and accidentally left out of pcre_study(), testing should
1667        pick them up.
1668    
1669    14. The -s option of pcretest has been documented for ages as being an old
1670        synonym of -m (show memory usage). I have changed it to mean "force study
1671        for every regex", that is, assume /S for every regex. This is similar to -i
1672        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1673        using it. It makes it easier to run collections of tests with and without
1674        study enabled, and thereby test pcre_study() more easily. All the standard
1675        tests are now run with and without -s (but some patterns can be marked as
1676        "never study" - see 20 below).
1677    
1678    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1679        restoration of the capturing data to the outer values was not happening
1680        correctly.
1681    
1682    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1683        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1684        pattern had matched an empty string, and so incorrectly returned a no
1685        match.
1686    
1687    17. There was optimizing code for the last branch of non-capturing parentheses,
1688        and also for the obeyed branch of a conditional subexpression, which used
1689        tail recursion to cut down on stack usage. Unfortunately, now that there is
1690        the possibility of (*THEN) occurring in these branches, tail recursion is
1691        no longer possible because the return has to be checked for (*THEN). These
1692        two optimizations have therefore been removed. [But see 8.20/11 above.]
1693    
1694    18. If a pattern containing \R was studied, it was assumed that \R always
1695        matched two bytes, thus causing the minimum subject length to be
1696        incorrectly computed because \R can also match just one byte.
1697    
1698    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1699        was incorrectly computed.
1700    
1701    20. If /S is present twice on a test pattern in pcretest input, it now
1702        *disables* studying, thereby overriding the use of -s on the command line
1703        (see 14 above). This is necessary for one or two tests to keep the output
1704        identical in both cases.
1705    
1706    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1707        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1708    
1709    22. When an atomic group that contained a capturing parenthesis was
1710        successfully matched, but the branch in which it appeared failed, the
1711        capturing was not being forgotten if a higher numbered group was later
1712        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1713        group 1 to "a", when in fact it should be unset. This applied to multi-
1714        branched capturing and non-capturing groups, repeated or not, and also to
1715        positive assertions (capturing in negative assertions does not happen
1716        in PCRE) and also to nested atomic groups.
1717    
1718    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1719        subject after a captured substring, to make it easier to tell which of a
1720        number of identical substrings has been captured.
1721    
1722    24. The way atomic groups are processed by pcre_exec() has been changed so that
1723        if they are repeated, backtracking one repetition now resets captured
1724        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1725        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1726        "aaa". Previously, it would have been "a". As part of this code
1727        refactoring, the way recursive calls are handled has also been changed.
1728    
1729    25. If an assertion condition captured any substrings, they were not passed
1730        back unless some other capturing happened later. For example, if
1731        (?(?=(a))a) was matched against "a", no capturing was returned.
1732    
1733    26. When studying a pattern that contained subroutine calls or assertions,
1734        the code for finding the minimum length of a possible match was handling
1735        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1736        group 1 called group 2 while simultaneously a separate group 2 called group
1737        1). A stack overflow occurred in this case. I have fixed this by limiting
1738        the recursion depth to 10.
1739    
1740    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1741        Fortmann. This supports explicit test numbers on the command line, and has
1742        argument validation and error reporting.
1743    
1744    28. An instance of \X with an unlimited repeat could fail if at any point the
1745        first character it looked at was a mark character.
1746    
1747    29. Some minor code refactoring concerning Unicode properties and scripts
1748        should reduce the stack requirement of match() slightly.
1749    
1750    30. Added the '=' option to pcretest to check the setting of unused capturing
1751        slots at the end of the pattern, which are documented as being -1, but are
1752        not included in the return count.
1753    
1754    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1755        compiled something random. Now it gives a compile-time error (as does
1756        Perl).
1757    
1758    32. A *MARK encountered during the processing of a positive assertion is now
1759        recorded and passed back (compatible with Perl).
1760    
1761    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1762        had alternative anchored branches, the search for a second match in a line
1763        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1764        matched the line "0102" twice. The same bug affected patterns that started
1765        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1766        twice.
1767    
1768    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1769        does, and because of capturing effects, quantifying parenthesized
1770        assertions may at times be useful. Quantifiers are now allowed for
1771        parenthesized assertions.
1772    
1773    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1774    
1775    36. \g was being checked for fancy things in a character class, when it should
1776        just be a literal "g".
1777    
1778    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1779        appearance of a nested POSIX class supersedes an apparent external class.
1780        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1781        unescaped square brackets may also appear as part of class names. For
1782        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1783        more like Perl. (But see 8.20/1 above.)
1784    
1785    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1786        was because it thought it was \N{name}, which is not supported).
1787    
1788    39. Add minix to OS list not supporting the -S option in pcretest.
1789    
1790    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1791        cannot analyze patterns in sufficient detail to catch mutual recursions
1792        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1793        subgroup is called recursively as a subpattern for a second time at the
1794        same position in the subject string. In previous releases this might have
1795        been caught by the recursion limit, or it might have run out of stack.
1796    
1797    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1798        happen only once. PCRE was, however incorrectly giving a compile time error
1799        "recursive call could loop indefinitely" because it cannot analyze the
1800        pattern in sufficient detail. The compile time test no longer happens when
1801        PCRE is compiling a conditional subpattern, but actual runaway loops are
1802        now caught at runtime (see 40 above).
1803    
1804    42. It seems that Perl allows any characters other than a closing parenthesis
1805        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1806        has been changed to be the same.
1807    
1808    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1809        as not to get warnings when autogen.sh is called. Also changed
1810        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1811    
1812    44. To help people who use pcregrep to scan files containing exceedingly long
1813        lines, the following changes have been made:
1814    
1815        (a) The default value of the buffer size parameter has been increased from
1816            8K to 20K. (The actual buffer used is three times this size.)
1817    
1818        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1819            PCRE is built.
1820    
1821        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1822            to be set at run time.
1823    
1824        (d) Numerical values in pcregrep options can be followed by K or M, for
1825            example --buffer-size=50K.
1826    
1827        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1828            given and the return code is set to 2.
1829    
1830    45. Add a pointer to the latest mark to the callout data block.
1831    
1832    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1833        partial match of an empty string instead of no match. This was specific to
1834        the use of ".".
1835    
1836    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1837        complete match instead of a partial match. This bug was dependent on both
1838        the PCRE_UTF8 and PCRE_DOTALL options being set.
1839    
1840    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1841        starting byte set, because \b was not being ignored.
1842    
1843    
1844    Version 8.12 15-Jan-2011
1845    ------------------------
1846    
1847    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1848        checks for such things as part of the documentation building process.
1849    
1850    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1851        --match-limit and --recursion-limit options (added for 8.11). In
1852        particular, this made one of the standard tests fail. (The integer value
1853        went into the wrong half of a long int.)
1854    
1855    3.  If the --colour option was given to pcregrep with -v (invert match), it
1856        did strange things, either producing crazy output, or crashing. It should,
1857        of course, ignore a request for colour when reporting lines that do not
1858        match.
1859    
1860    4.  Another pcregrep bug caused similar problems if --colour was specified with
1861        -M (multiline) and the pattern match finished with a line ending.
1862    
1863    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1864        matched in multiline mode, the following line was shown as part of the
1865        match. This seems wrong, so I have changed it.
1866    
1867    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1868        the check for further matches in the same line (so they could be coloured)
1869        to overrun the end of the current line. If another match was found, it was
1870        incorrectly shown (and then shown again when found in the next line).
1871    
1872    7.  If pcregrep was compiled under Windows, there was a reference to the
1873        function pcregrep_exit() before it was defined. I am assuming this was
1874        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1875        reported by a user. I've moved the definition above the reference.
1876    
1877    
1878    Version 8.11 10-Dec-2010
1879    ------------------------
1880    
1881    1.  (*THEN) was not working properly if there were untried alternatives prior
1882        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1883        backtracked to try for "b" instead of moving to the next alternative branch
1884        at the same level (in this case, to look for "c"). The Perl documentation
1885        is clear that when (*THEN) is backtracked onto, it goes to the "next
1886        alternative in the innermost enclosing group".
1887    
1888    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1889        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1890        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1891        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1892        (*THEN).
1893    
1894    3.  If \s appeared in a character class, it removed the VT character from
1895        the class, even if it had been included by some previous item, for example
1896        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1897        of \s, but is part of the POSIX "space" class.)
1898    
1899    4.  A partial match never returns an empty string (because you can always
1900        match an empty string at the end of the subject); however the checking for
1901        an empty string was starting at the "start of match" point. This has been
1902        changed to the "earliest inspected character" point, because the returned
1903        data for a partial match starts at this character. This means that, for
1904        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1905        (previously it gave "no match").
1906    
1907    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1908        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1909        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1910        has an implication that the given string is incomplete (because a partial
1911        match is preferred over a full match). For this reason, these items now
1912        give a partial match in this situation. [Aside: previously, the one case
1913        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1914        match rather than a full match, which was wrong by the old rules, but is
1915        now correct.]
1916    
1917    6.  There was a bug in the handling of #-introduced comments, recognized when
1918        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1919        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1920        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1921        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1922        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1923        places in pcre_compile().
1924    
1925    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1926        comments when looking ahead for named forward references to subpatterns,
1927        the only newline sequence it recognized was NL. It now handles newlines
1928        according to the set newline convention.
1929    
1930    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1931        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1932        cater for a lack of strerror(). These oversights have been fixed.
1933    
1934    9.  Added --match-limit and --recursion-limit to pcregrep.
1935    
1936    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1937    
1938    11. When the -o option was used, pcregrep was setting a return code of 1, even
1939        when matches were found, and --line-buffered was not being honoured.
1940    
1941    12. Added an optional parentheses number to the -o and --only-matching options
1942        of pcregrep.
1943    
1944    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1945        can match an empty string. The code to do it in pcretest and pcredemo
1946        needed fixing:
1947    
1948        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1949            only one byte after an empty string match just before CRLF (this case
1950            just got forgotten; "any" and "anycrlf" were OK).
1951    
1952        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1953            mode when an empty string match preceded an ASCII character followed by
1954            a non-ASCII character. (The code for advancing by one character rather
1955            than one byte was nonsense.)
1956    
1957        (c) The pcredemo.c sample program did not have any code at all to handle
1958            the cases when CRLF is a valid newline sequence.
1959    
1960    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1961        as a starting offset was within the subject string. There is now a new
1962        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1963        negative or greater than the length of the string. In order to test this,
1964        pcretest is extended to allow the setting of negative starting offsets.
1965    
1966    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1967        starting offset points to the beginning of a UTF-8 character was
1968        unnecessarily clumsy. I tidied it up.
1969    
1970    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1971        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1972    
1973    17. Nobody had reported that the --include_dir option, which was added in
1974        release 7.7 should have been called --include-dir (hyphen, not underscore)
1975        for compatibility with GNU grep. I have changed it to --include-dir, but
1976        left --include_dir as an undocumented synonym, and the same for
1977        --exclude-dir, though that is not available in GNU grep, at least as of
1978        release 2.5.4.
1979    
1980    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1981        characters from a string of bytes) have been redefined so as not to use
1982        loops, in order to improve performance in some environments. At the same
1983        time, I abstracted some of the common code into auxiliary macros to save
1984        repetition (this should not affect the compiled code).
1985    
1986    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1987        compile-time error is now given if \c is not followed by an ASCII
1988        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1989        different, and any byte value is allowed.)
1990    
1991    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1992        START_OPTIMIZE option, which is now allowed at compile time - but just
1993        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1994        to pcregrep and other applications that have no direct access to PCRE
1995        options. The new /Y option in pcretest sets this option when calling
1996        pcre_compile().
1997    
1998    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1999        back references. Groups containing recursive back references were forced to
2000        be atomic by that change, but in the case of named groups, the amount of
2001        memory required was incorrectly computed, leading to "Failed: internal
2002        error: code overflow". This has been fixed.
2003    
2004    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2005        pcretest.c, to avoid build problems in some Borland environments.
2006    
2007    
2008    Version 8.10 25-Jun-2010
2009    ------------------------
2010    
2011    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2012        THEN.
2013    
2014    2.  (*ACCEPT) was not working when inside an atomic group.
2015    
2016    3.  Inside a character class, \B is treated as a literal by default, but
2017        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2018        causes the error). The code is unchanged, but I tidied the documentation.
2019    
2020    4.  Inside a character class, PCRE always treated \R and \X as literals,
2021        whereas Perl faults them if its -w option is set. I have changed PCRE so
2022        that it faults them when PCRE_EXTRA is set.
2023    
2024    5.  Added support for \N, which always matches any character other than
2025        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2026    
2027    6.  When compiling pcregrep with newer versions of gcc which may have
2028        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2029        declared with attribute warn_unused_result" were given. Just casting the
2030        result to (void) does not stop the warnings; a more elaborate fudge is
2031        needed. I've used a macro to implement this.
2032    
2033    7.  Minor change to pcretest.c to avoid a compiler warning.
2034    
2035    8.  Added four artifical Unicode properties to help with an option to make
2036        \s etc use properties (see next item). The new properties are: Xan
2037        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2038    
2039    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2040        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2041        this option. Modified pcretest to add /W to test this facility. Added
2042        REG_UCP to make it available via the POSIX interface.
2043    
2044    10. Added --line-buffered to pcregrep.
2045    
2046    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2047        studied, and the match started with a letter with a code point greater than
2048        127 whose first byte was different to the first byte of the other case of
2049        the letter, the other case of this starting letter was not recognized
2050        (#976).
2051    
2052    12. If a pattern that was studied started with a repeated Unicode property
2053        test, for example, \p{Nd}+, there was the theoretical possibility of
2054        setting up an incorrect bitmap of starting bytes, but fortunately it could
2055        not have actually happened in practice until change 8 above was made (it
2056        added property types that matched character-matching opcodes).
2057    
2058    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2059        possible starting bytes for non-anchored patterns.
2060    
2061    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2062        \R, and also a number of cases that involve Unicode properties, both
2063        explicit and implicit when PCRE_UCP is set.
2064    
2065    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2066        input, it could crash or give wrong results if characters with values
2067        greater than 0xc0 were present in the subject string. (Detail: it assumed
2068        UTF-8 input when processing these items.)
2069    
2070    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2071        size_t is 64-bit (#991).
2072    
2073    17. Added a check for running out of memory when PCRE is compiled with
2074        --disable-stack-for-recursion (#990).
2075    
2076    18. If the last data line in a file for pcretest does not have a newline on
2077        the end, a newline was missing in the output.
2078    
2079    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2080        less than 128) in its various bitmaps. However, there is a facility for
2081        generating tables according to the current locale when PCRE is compiled. It
2082        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2083        characters, are recognized by isspace() and therefore were getting set in
2084        these tables, and indeed these tables seem to approximate to ISO 8859. This
2085        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2086        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2087        which of course cannot start UTF-8 characters. I have changed the code so
2088        that only real ASCII characters (less than 128) and the correct starting
2089        bytes for UTF-8 encodings are set for characters greater than 127 when in
2090        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2091        altogether.)
2092    
2093    20. Added the /T option to pcretest so as to be able to run tests with non-
2094        standard character tables, thus making it possible to include the tests
2095        used for 19 above in the standard set of tests.
2096    
2097    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2098        reference to a subpattern the other side of a comment that contains an
2099        opening parenthesis caused either an internal compiling error, or a
2100        reference to the wrong subpattern.
2101    
2102    
2103    Version 8.02 19-Mar-2010
2104    ------------------------
2105    
2106    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2107    
2108    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2109        configured.
2110    
2111    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2112        original author of that file, following a query about its status.
2113    
2114    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2115        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2116    
2117    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2118        quantifier applied to a forward-referencing subroutine call, could compile
2119        incorrect code or give the error "internal error: previously-checked
2120        referenced subpattern not found".
2121    
2122    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2123        variables to point to external functions. For these systems, therefore,
2124        pcre_malloc etc. are now initialized to local functions that call the
2125        relevant global functions.
2126    
2127    7.  There were two entries missing in the vectors called coptable and poptable
2128        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2129        I've fixed the data, and added a kludgy way of testing at compile time that
2130        the lengths are correct (equal to the number of opcodes).
2131    
2132    8.  Following on from 7, I added a similar kludge to check the length of the
2133        eint vector in pcreposix.c.
2134    
2135    9.  Error texts for pcre_compile() are held as one long string to avoid too
2136        much relocation at load time. To find a text, the string is searched,
2137        counting zeros. There was no check for running off the end of the string,
2138        which could happen if a new error number was added without updating the
2139        string.
2140    
2141    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2142    
2143    11. \K was not working if it appeared in an atomic group or in a group that
2144        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2145        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2146        the assertion is positive, but not if it is negative.
2147    
2148    12. Change 11 fortuitously reduced the size of the stack frame used in the
2149        "match()" function of pcre_exec.c by one pointer. Forthcoming
2150        implementation of support for (*MARK) will need an extra pointer on the
2151        stack; I have reserved it now, so that the stack frame size does not
2152        decrease.
2153    
2154    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2155        item in branch that calls a recursion is a subroutine call - as in the
2156        second branch in the above example - was incorrectly given the compile-
2157        time error "recursive call could loop indefinitely" because pcre_compile()
2158        was not correctly checking the subroutine for matching a non-empty string.
2159    
2160    14. The checks for overrunning compiling workspace could trigger after an
2161        overrun had occurred. This is a "should never occur" error, but it can be
2162        triggered by pathological patterns such as hundreds of nested parentheses.
2163        The checks now trigger 100 bytes before the end of the workspace.
2164    
2165    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2166    
2167    
2168    Version 8.01 19-Jan-2010
2169    ------------------------
2170    
2171    1.  If a pattern contained a conditional subpattern with only one branch (in
2172        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2173        computed the wrong minimum data length (which is of course zero for such
2174        subpatterns). This could cause incorrect "no match" results.
2175    
2176    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2177        the pattern is reset in the first branch, pcre_compile() failed with
2178        "internal error: code overflow at offset...". This happened only when
2179        the reset was to the original external option setting. (An optimization
2180        abstracts leading options settings into an external setting, which was the
2181        cause of this.)
2182    
2183    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2184        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2185        assertion pattern did not match (meaning that the assertion was true), it
2186        was incorrectly treated as false if the SKIP had been reached during the
2187        matching. This also applied to assertions used as conditions.
2188    
2189    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2190        assertion subpattern, including such a pattern used as a condition,
2191        unpredictable results occurred, instead of the error return
2192        PCRE_ERROR_DFA_UITEM.
2193    
2194    5.  The C++ GlobalReplace function was not working like Perl for the special
2195        situation when an empty string is matched. It now does the fancy magic
2196        stuff that is necessary.
2197    
2198    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2199        removed. (These were left over from very, very early versions of PCRE.)
2200    
2201    7.  Some cosmetic changes to the code to make life easier when compiling it
2202        as part of something else:
2203    
2204        (a) Change DEBUG to PCRE_DEBUG.
2205    
2206        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2207            called "current" as "current_branch", to prevent a collision with the
2208            Linux macro when compiled as a kernel module.
2209    
2210        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2211            prevent a collision with the Linux macro when compiled as a kernel
2212            module.
2213    
2214    8.  In pcre_compile() there are some checks for integer overflows that used to
2215        cast potentially large values to (double). This has been changed to that
2216        when building, a check for int64_t is made, and if it is found, it is used
2217        instead, thus avoiding the use of floating point arithmetic. (There is no
2218        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2219        double.
2220    
2221    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2222        2005 (difference between two addresses compared to an unsigned value).
2223    
2224    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2225        custom one, because of the following reported problem in Windows:
2226    
2227          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2228              under Win32.
2229          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2230              therefore missing the function definition.
2231          - The compiler thus generates a "C" signature for the test function.
2232          - The linker fails to find the "C" function.
2233          - PCRE fails to configure if asked to do so against libbz2.
2234    
2235    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2236        messages were output:
2237    
2238          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2239          rerunning libtoolize, to keep the correct libtool macros in-tree.
2240          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2241    
2242        I have done both of these things.
2243    
2244    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2245        most of the time, it *can* run out if it is given a pattern that contains a
2246        runaway infinite recursion. I updated the discussion in the pcrestack man
2247        page.
2248    
2249    13. Now that we have gone to the x.xx style of version numbers, the minor
2250        version may start with zero. Using 08 or 09 is a bad idea because users
2251        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2252        interpreted as invalid octal numbers. I've updated the previous comment in
2253        configure.ac, and also added a check that gives an error if 08 or 09 are
2254        used.
2255    
2256    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2257        causing partial matching to fail when the end of the subject matched \W
2258        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2259    
2260    15. There were some discrepancies between the declarations in pcre_internal.h
2261        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2262        their definitions. The declarations used "const uschar *" and the
2263        definitions used USPTR. Even though USPTR is normally defined as "const
2264        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2265        reported that: "This difference in casting confuses some C++ compilers, for
2266        example, SunCC recognizes above declarations as different functions and
2267        generates broken code for hbpcre." I have changed the declarations to use
2268        USPTR.
2269    
2270    16. GNU libtool is named differently on some systems. The autogen.sh script now
2271        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2272        (FreeBSD).
2273    
2274    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2275        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2276        comment: "Figure out how to create a longlong from a string: strtoll and
2277        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2278        instance, but it only takes 2 args instead of 3!"
2279    
2280    18. A subtle bug concerned with back references has been fixed by a change of
2281        specification, with a corresponding code fix. A pattern such as
2282        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2283        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2284        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2285        same bug. Such groups have to be quantified to be useful, or contained
2286        inside another quantified group. (If there's no repetition, the reference
2287        can never match.) The problem arises because, having left the group and
2288        moved on to the rest of the pattern, a later failure that backtracks into
2289        the group uses the captured value from the final iteration of the group
2290        rather than the correct earlier one. I have fixed this in PCRE by forcing
2291        any group that contains a reference to itself to be an atomic group; that
2292        is, there cannot be any backtracking into it once it has completed. This is
2293        similar to recursive and subroutine calls.
2294    
2295    
2296    Version 8.00 19-Oct-09
2297    ----------------------
2298    
2299    1.  The table for translating pcre_compile() error codes into POSIX error codes
2300        was out-of-date, and there was no check on the pcre_compile() error code
2301        being within the table. This could lead to an OK return being given in
2302        error.
2303    
2304    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2305        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2306        in a Windows environment.
2307    
2308    3.  The pcregrep --count option prints the count for each file even when it is
2309        zero, as does GNU grep. However, pcregrep was also printing all files when
2310        --files-with-matches was added. Now, when both options are given, it prints
2311        counts only for those files that have at least one match. (GNU grep just
2312        prints the file name in this circumstance, but including the count seems
2313        more useful - otherwise, why use --count?) Also ensured that the
2314        combination -clh just lists non-zero counts, with no names.
2315    
2316    4.  The long form of the pcregrep -F option was incorrectly implemented as
2317        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2318        but it seems right to fix it, and I didn't think it was worth preserving
2319        the old behaviour.
2320    
2321    5.  The command line items --regex=pattern and --regexp=pattern were not
2322        recognized by pcregrep, which required --regex pattern or --regexp pattern
2323        (with a space rather than an '='). The man page documented the '=' forms,
2324        which are compatible with GNU grep; these now work.
2325    
2326    6.  No libpcreposix.pc file was created for pkg-config; there was just
2327        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2328    
2329    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2330        when UCP support is not needed, by modifying the Python script that
2331        generates it from Unicode data files. This should not matter if the module
2332        is correctly used as a library, but I received one complaint about 50K of
2333        unwanted data. My guess is that the person linked everything into his
2334        program rather than using a library. Anyway, it does no harm.
2335    
2336    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2337        was a minimum greater than 1 for a wide character in a possessive
2338        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2339        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2340        character. Chaos in the form of incorrect output or a compiling loop could
2341        result.
2342    
2343    9.  The restrictions on what a pattern can contain when partial matching is
2344        requested for pcre_exec() have been removed. All patterns can now be
2345        partially matched by this function. In addition, if there are at least two
2346        slots in the offset vector, the offset of the earliest inspected character
2347        for the match and the offset of the end of the subject are set in them when
2348        PCRE_ERROR_PARTIAL is returned.
2349    
2350    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2351        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2352        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2353        and may be more useful for multi-segment matching.
2354    
2355    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2356        used to be given if ever the end of the subject was reached; now it is
2357        given only if matching could not proceed because another character was
2358        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2359        string "Z", which now yields "no match" instead of "partial match". In the
2360        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2361        final character ended with (*FAIL).
2362    
2363    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2364        if the pattern had a "must contain" character that was already found in the
2365        earlier partial match, unless partial matching was again requested. For
2366        example, with the pattern /dog.(body)?/, the "must contain" character is
2367        "g". If the first part-match was for the string "dog", restarting with
2368        "sbody" failed. This bug has been fixed.
2369    
2370    13. The string returned by pcre_dfa_exec() after a partial match has been
2371        changed so that it starts at the first inspected character rather than the
2372        first character of the match. This makes a difference only if the pattern
2373        starts with a lookbehind assertion or \b or \B (\K is not supported by
2374        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2375        matching functions compatible, and I think it's the right thing to do.
2376    
2377    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2378        so that the demonstration program is easily available in environments where
2379        PCRE has not been installed from source.
2380    
2381    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2382        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2383        library.
2384    
2385    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2386        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2387        is not the first non-POSIX option to be added. Clearly some people find
2388        these options useful.
2389    
2390    17. If a caller to the POSIX matching function regexec() passes a non-zero
2391        value for nmatch with a NULL value for pmatch, the value of
2392        nmatch is forced to zero.
2393    
2394    18. RunGrepTest did not have a test for the availability of the -u option of
2395        the diff command, as RunTest does. It now checks in the same way as
2396        RunTest, and also checks for the -b option.
2397    
2398    19. If an odd number of negated classes containing just a single character
2399        interposed, within parentheses, between a forward reference to a named
2400        subpattern and the definition of the subpattern, compilation crashed with
2401        an internal error, complaining that it could not find the referenced
2402        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2403        [The bug was that it was starting one character too far in when skipping
2404        over the character class, thus treating the ] as data rather than
2405        terminating the class. This meant it could skip too much.]
2406    
2407    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2408        /g option in pcretest when the pattern contains \K, which makes it possible
2409        to have an empty string match not at the start, even when the pattern is
2410        anchored. Updated pcretest and pcredemo to use this option.
2411    
2412    21. If the maximum number of capturing subpatterns in a recursion was greater
2413        than the maximum at the outer level, the higher number was returned, but
2414        with unset values at the outer level. The correct (outer level) value is
2415        now given.
2416    
2417    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2418        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2419        make it do so. The string so far is captured, making this feature
2420        compatible with Perl.
2421    
2422    23. The tests have been re-organized, adding tests 11 and 12, to make it
2423        possible to check the Perl 5.10 features against Perl 5.10.
2424    
2425    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2426        pattern matches a fixed length string. PCRE did not allow this; now it
2427        does. Neither allows recursion.
2428    
2429    25. I finally figured out how to implement a request to provide the minimum
2430        length of subject string that was needed in order to match a given pattern.
2431        (It was back references and recursion that I had previously got hung up
2432        on.) This code has now been added to pcre_study(); it finds a lower bound
2433        to the length of subject needed. It is not necessarily the greatest lower
2434        bound, but using it to avoid searching strings that are too short does give
2435        some useful speed-ups. The value is available to calling programs via
2436        pcre_fullinfo().
2437    
2438    26. While implementing 25, I discovered to my embarrassment that pcretest had
2439        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2440        study optimizations had never been tested with that matching function.
2441        Oops. What is worse, even when it was passed study data, there was a bug in
2442        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2443        were also very few tests of studied patterns with pcre_dfa_exec().
2444    
2445    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2446        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2447        on the other side of the coin, they are no longer allowed to have different
2448        names, because these cannot be distinguished in PCRE, and this has caused
2449        confusion. (This is a difference from Perl.)
2450    
2451    28. When duplicate subpattern names are present (necessarily with different
2452        numbers, as required by 27 above), and a test is made by name in a
2453        conditional pattern, either for a subpattern having been matched, or for
2454        recursion in such a pattern, all the associated numbered subpatterns are
2455        tested, and the overall condition is true if the condition is true for any
2456        one of them. This is the way Perl works, and is also more like the way
2457        testing by number works.
2458    
2459    
2460    Version 7.9 11-Apr-09
2461    ---------------------
2462    
2463    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2464        (pcretest), all targets were linked against these libraries. This included
2465        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2466        libraries. This caused unwanted dependencies to be created. This problem
2467        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2468        pcretest is linked with readline.
2469    
2470    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2471        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2472        moved outside it again, because FALSE and TRUE are already defined in AIX,
2473        but BOOL is not.
2474    
2475    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2476        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2477    
2478    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2479        hyphens) following file names and line numbers when outputting matching
2480        lines. This is not true; no spaces are inserted. I have also clarified the
2481        wording for the --colour (or --color) option.
2482    
2483    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2484        was not coloured; this is different to GNU grep, so I have changed it to be
2485        the same.
2486    
2487    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2488        each matching line was coloured. Now it goes on to look for further matches
2489        of any of the test patterns, which is the same behaviour as GNU grep.
2490    
2491    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2492        doesn't make sense to accept an empty string match in pcregrep, so I have
2493        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2494        seems to be how GNU grep behaves. [But see later change 40 for release
2495        8.33.]
2496    
2497    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2498        start or after a newline", because the conditional assertion was not being
2499        correctly handled. The rule now is that both the assertion and what follows
2500        in the first alternative must satisfy the test.
2501    
2502    9.  If auto-callout was enabled in a pattern with a conditional group whose
2503        condition was an assertion, PCRE could crash during matching, both with
2504        pcre_exec() and pcre_dfa_exec().
2505    
2506    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2507        used for matching.
2508    
2509    11. Unicode property support in character classes was not working for
2510        characters (bytes) greater than 127 when not in UTF-8 mode.
2511    
2512    12. Added the -M command line option to pcretest.
2513    
2514    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2515    
2516    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2517    
2518    16. Added comments and documentation about mis-use of no_arg in the C++
2519        wrapper.
2520    
2521    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2522        from Martin Jerabek that uses macro names for all relevant character and
2523        string constants.
2524    
2525    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2526        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2527        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2528        these, but not everybody uses configure.
2529    
2530    19. A conditional group that had only one branch was not being correctly
2531        recognized as an item that could match an empty string. This meant that an
2532        enclosing group might also not be so recognized, causing infinite looping
2533        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2534        with the subject "ab", where knowledge that the repeated group can match
2535        nothing is needed in order to break the loop.
2536    
2537    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2538        exec(), but without supplying a callout function, matching went wrong.
2539    
2540    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2541        leak if the size of the offset vector was greater than 30. When the vector
2542        is smaller, the saved offsets during recursion go onto a local stack
2543        vector, but for larger vectors malloc() is used. It was failing to free
2544        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2545        error, in fact).
2546    
2547    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2548        heapframe that is used only when UTF-8 support is enabled. This caused no
2549        problem, but was untidy.
2550    
2551    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2552        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2553        included within another project.
2554    
2555    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2556        slightly modified by me:
2557    
2558          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2559              not building pcregrep.
2560    
2561          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2562              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2563    
2564    25. Forward references, both numeric and by name, in patterns that made use of
2565        duplicate group numbers, could behave incorrectly or give incorrect errors,
2566        because when scanning forward to find the reference group, PCRE was not
2567        taking into account the duplicate group numbers. A pattern such as
2568        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2569    
2570    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2571        the feature of a custom pointer more persuasive (as requested by a user).
2572    
2573    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2574        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2575        pre-defined.
2576    
2577    28. Added support for (*UTF8) at the start of a pattern.
2578    
2579    29. Arrange for flags added by the "release type" setting in CMake to be shown
2580        in the configuration summary.
2581    
2582    
2583    Version 7.8 05-Sep-08
2584    ---------------------
2585    
2586    1.  Replaced UCP searching code with optimized version as implemented for Ad
2587        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2588        stage table and inline lookup instead of a function, giving speed ups of 2
2589        to 5 times on some simple patterns that I tested. Permission was given to
2590        distribute the MultiStage2.py script that generates the tables (it's not in
2591        the tarball, but is in the Subversion repository).
2592    
2593    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2594        scripts.
2595    
2596    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2597        a group with a zero qualifier. The result of the study could be incorrect,
2598        or the function might crash, depending on the pattern.
2599    
2600    4.  Caseless matching was not working for non-ASCII characters in back
2601        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2602        It now works when Unicode Property Support is available.
2603    
2604    5.  In pcretest, an escape such as \x{de} in the data was always generating
2605        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2606        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2607        truncation.
2608    
2609    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2610    
2611    7.  Added two (int) casts to pcregrep when printing the difference of two
2612        pointers, in case they are 64-bit values.
2613    
2614    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2615        test 2 if it fails.
2616    
2617    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2618        and a #define of that name to empty if it is not externally set. This is to
2619        allow users of MSVC to set it if necessary.
2620    
2621    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2622        the convenience functions in the pcre_get.c source file.
2623    
2624    11. An option change at the start of a pattern that had top-level alternatives
2625        could cause overwriting and/or a crash. This command provoked a crash in
2626        some environments:
2627    
2628          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2629    
2630        This potential security problem was recorded as CVE-2008-2371.
2631    
2632    12. For a pattern where the match had to start at the beginning or immediately
2633        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2634        pcre_dfa_exec() could read past the end of the passed subject if there was
2635        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2636        pcretest so that it places the subject at the end of its malloc-ed buffer.
2637    
2638    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2639        exec() might read past the end of the data buffer in UTF-8 mode.
2640    
2641    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2642        the data contained the byte 0x85 as part of a UTF-8 character within its
2643        first line. This applied both to normal and DFA matching.
2644    
2645    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2646        /^[^d]*?$/8 failed to match "abc".
2647    
2648    16. Added a missing copyright notice to pcrecpp_internal.h.
2649    
2650    17. Make it more clear in the documentation that values returned from
2651        pcre_exec() in ovector are byte offsets, not character counts.
2652    
2653    18. Tidied a few places to stop certain compilers from issuing warnings.
2654    
2655    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2656        supplied by Stefan Weber. I made a further small update for 7.8 because
2657        there is a change of source arrangements: the pcre_searchfuncs.c module is
2658        replaced by pcre_ucd.c.
2659    
2660    
2661    Version 7.7 07-May-08
2662    ---------------------
2663    
2664    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2665        a string to a long long, pretend we don't even have a long long." This is
2666        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2667    
2668    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2669        pre-7.6 versions, which defined a global no_arg variable instead of putting
2670        it in the RE class. (See also #8 below.)
2671    
2672    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2673        Lopes.
2674    
2675    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2676    
2677        (1) The include/exclude patterns were being applied to the whole pathnames
2678            of files, instead of just to the final components.
2679    
2680        (2) If there was more than one level of directory, the subdirectories were
2681            skipped unless they satisfied the include/exclude conditions. This is
2682            inconsistent with GNU grep (and could even be seen as contrary to the
2683            pcregrep specification - which I improved to make it absolutely clear).
2684            The action now is always to scan all levels of directory, and just
2685            apply the include/exclude patterns to regular files.
2686    
2687    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2688        --exclude_dir in the tests to avoid scanning .svn directories.
2689    
2690    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2691        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2692        doesn't support NULs in patterns.
2693    
2694    7.  Added some missing "const"s to declarations of static tables in
2695        pcre_compile.c and pcre_dfa_exec.c.
2696    
2697    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2698        caused by fix #2  above. (Subsequently also a second patch to fix the
2699        first patch. And a third patch - this was a messy problem.)
2700    
2701    9.  Applied Craig's patch to remove the use of push_back().
2702    
2703    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2704        matching function regexec().
2705    
2706    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2707        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2708        references. PCRE supports relative numbers with this syntax (I don't think
2709        Oniguruma does).
2710    
2711    12. Previously, a group with a zero repeat such as (...){0} was completely
2712        omitted from the compiled regex. However, this means that if the group
2713        was called as a subroutine from elsewhere in the pattern, things went wrong
2714        (an internal error was given). Such groups are now left in the compiled
2715        pattern, with a new opcode that causes them to be skipped at execution
2716        time.
2717    
2718    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2719        to the way PCRE behaves:
2720    
2721        (a) A lone ] character is dis-allowed (Perl treats it as data).
2722    
2723        (b) A back reference to an unmatched subpattern matches an empty string
2724            (Perl fails the current match path).
2725    
2726        (c) A data ] in a character class must be notated as \] because if the
2727            first data character in a class is ], it defines an empty class. (In
2728            Perl it is not possible to have an empty class.) The empty class []
2729            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2730            The negative empty class [^] matches any one character, independently
2731            of the DOTALL setting.
2732    
2733    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2734        non-existent subpattern following a character class starting with ']' and
2735        containing () gave an internal compiling error instead of "reference to
2736        non-existent subpattern". Fortunately, when the pattern did exist, the
2737        compiled code was correct. (When scanning forwards to check for the
2738        existence of the subpattern, it was treating the data ']' as terminating
2739        the class, so got the count wrong. When actually compiling, the reference
2740        was subsequently set up correctly.)
2741    
2742    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2743        it was being rejected as not supported by pcre_dfa_exec(), even though
2744        other assertions are supported. I have made pcre_dfa_exec() support
2745        (*FAIL).
2746    
2747    16. The implementation of 13c above involved the invention of a new opcode,
2748        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2749        cannot be changed at match time, I realized I could make a small
2750        improvement to matching performance by compiling OP_ALLANY instead of
2751        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2752        on the OP_ANY path.
2753    
2754    17. Compiling pcretest on Windows with readline support failed without the
2755        following two fixes: (1) Make the unistd.h include conditional on
2756        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2757    
2758    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2759        ncurses library to be included for pcretest when ReadLine support is
2760        requested, but also to allow for it to be overridden. This patch came from
2761        Daniel Bergström.
2762    
2763    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2764        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2765        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2766        spotting this.
2767    
2768    
2769    Version 7.6 28-Jan-08
2770    ---------------------
2771    
2772    1.  A character class containing a very large number of characters with
2773        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2774        overflow.
2775    
2776    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2777        HAVE_LONG_LONG is not defined.
2778    
2779    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2780        bring them up to date and include new features. This patch includes:
2781    
2782        - Fixed PH's badly added libz and libbz2 support.
2783        - Fixed a problem with static linking.
2784        - Added pcredemo. [But later removed - see 7 below.]
2785        - Fixed dftables problem and added an option.
2786        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2787            HAVE_LONG_LONG.
2788        - Added readline support for pcretest.
2789        - Added an listing of the option settings after cmake has run.
2790    
2791    4.  A user submitted a patch to Makefile that makes it easy to create
2792        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2793        Makefile.am that cause it to include this special target, without
2794        affecting anything else. Note that the same mingw target plus all
2795        the other distribution libraries and programs are now supported
2796        when configuring with CMake (see 6 below) instead of with
2797        Configure/Make.
2798    
2799    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2800        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2801        exported in the Windows port". It has not yet been confirmed that the patch
2802        solves the problem, but it does no harm.
2803    
2804    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2805        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2806        with CMake, and also correct the comment about stack recursion.
2807    
2808    7.  Remove the automatic building of pcredemo from the ./configure system and
2809        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2810        of a program that users should build themselves after PCRE is installed, so
2811        building it automatically is not really right. What is more, it gave
2812        trouble in some build environments.
2813    
2814    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2815    
2816    
2817    Version 7.5 10-Jan-08
2818    ---------------------
2819    
2820    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2821        values in parens when parsing an RE using the C++ wrapper."
2822    
2823    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2824        Characters greater than 255 were excluded from the class instead of being
2825        included.
2826    
2827    3.  The same bug as (2) above applied to negated POSIX classes such as
2828        [:^space:].
2829    
2830    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2831        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2832        I have changed it.
2833    
2834    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2835        first named subpattern) and a construct such as (?&a) would reference the
2836        first named subpattern whose name started with "a" (in other words, the
2837        length check was missing). Both these problems are fixed. "Subpattern name
2838        expected" is now given for (?&) (a zero-length name), and this patch also
2839        makes it give the same error for \k'' (previously it complained that that
2840        was a reference to a non-existent subpattern).
2841    
2842    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2843        this is right because (?- can be followed by option settings as well as by
2844        digits. I have, however, made the messages clearer.
2845    
2846    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2847        than the number used in the conditional) now cause a compile-time error.
2848        This is actually not compatible with Perl, which accepts such patterns, but
2849        treats the conditional as always being FALSE (as PCRE used to), but it
2850        seems to me that giving a diagnostic is better.
2851    
2852    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2853        and messages.
2854    
2855    9.  Fix two occurrences of "backslash" in comments that should have been
2856        "backspace".
2857    
2858    10. Remove two redundant lines of code that can never be obeyed (their function
2859        was moved elsewhere).
2860    
2861    11. The program that makes PCRE's Unicode character property table had a bug
2862        which caused it to generate incorrect table entries for sequences of
2863        characters that have the same character type, but are in different scripts.
2864        It amalgamated them into a single range, with the script of the first of
2865        them. In other words, some characters were in the wrong script. There were
2866        thirteen such cases, affecting characters in the following ranges:
2867    
2868          U+002b0 - U+002c1
2869          U+0060c - U+0060d
2870          U+0061e - U+00612
2871          U+0064b - U+0065e
2872          U+0074d - U+0076d
2873          U+01800 - U+01805
2874          U+01d00 - U+01d77
2875          U+01d9b - U+01dbf
2876          U+0200b - U+0200f
2877          U+030fc - U+030fe
2878          U+03260 - U+0327f
2879          U+0fb46 - U+0fbb1
2880          U+10450 - U+1049d
2881    
2882    12. The -o option (show only the matching part of a line) for pcregrep was not
2883        compatible with GNU grep in that, if there was more than one match in a
2884        line, it showed only the first of them. It now behaves in the same way as
2885        GNU grep.
2886    
2887    13. If the -o and -v options were combined for pcregrep, it printed a blank
2888        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2889        does the same. The return code can be used to tell if there were any
2890        non-matching lines.
2891    
2892    14. Added --file-offsets and --line-offsets to pcregrep.
2893    
2894    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2895        infinitely looping recursion. The bug was that positive lookaheads were not
2896        being skipped when checking for a possible empty match (negative lookaheads
2897        and both kinds of lookbehind were skipped).
2898    
2899    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2900        inclusion of <windows.h> to before rather than after the definition of
2901        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2902    
2903    17. Specifying a possessive quantifier with a specific limit for a Unicode
2904        character property caused pcre_compile() to compile bad code, which led at
2905        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2906        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2907        caused the error; without that there was no problem.
2908    
2909    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2910    
2911    19. Added --enable-pcretest-libreadline.
2912    
2913    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2914        RE::GlobalReplace(). As a result, the number of replacements returned was
2915        double what it should be. I removed one of the increments, but Craig sent a
2916        later patch that removed the other one (the right fix) and added unit tests
2917        that check the return values (which was not done before).
2918    
2919    21. Several CMake things:
2920    
2921        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2922            the names libpcre and libpcreposix, not just pcre and pcreposix.
2923    
2924        (2) The above change means that pcretest and pcregrep are now correctly
2925            linked with the newly-built libraries, not previously installed ones.
2926    
2927        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2928    
2929    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2930        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2931        UTF-8 newline character). The key issue is that the pattern starts .*;
2932        this means that the match must be either at the beginning, or after a
2933        newline. The bug was in the code for advancing after a failed match and
2934        checking that the new position followed a newline. It was not taking
2935        account of UTF-8 characters correctly.
2936    
2937    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2938        character classes. PCRE was not treating the sequence [:...:] as a
2939        character class unless the ... were all letters. Perl, however, seems to
2940        allow any characters between [: and :], though of course it rejects as
2941        unknown any "names" that contain non-letters, because all the known class
2942        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2943        for example, whereas PCRE did not - it did not recognize a POSIX character
2944        class. This seemed a bit dangerous, so the code has been changed to be
2945        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2946        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2947        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2948        Perl does, and where it didn't before.
2949    
2950    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2951        Windows environments %n is disabled by default.
2952    
2953    
2954    Version 7.4 21-Sep-07
2955    ---------------------
2956    
2957    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2958        means that a class such as [\s] counted as "explicit reference to CR or
2959        LF". That isn't really right - the whole point of the change was to try to
2960        help when there was an actual mention of one of the two characters. So now
2961        the change happens only if \r or \n (or a literal CR or LF) character is
2962        encountered.
2963    
2964    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2965        of both had grown to the point where there were only 3 bits left.
2966        Fortunately, there was spare space in the data structure, and so I have
2967        moved the internal flags into a new 16-bit field to free up more option
2968        bits.
2969    
2970    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2971        but did not set the internal JCHANGED flag - either of these is enough to
2972        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2973        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2974        start sets both bits.
2975    
2976    4.  Added options (at build time, compile time, exec time) to change \R from
2977        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2978    
2979    5.  doc/pcresyntax.html was missing from the distribution.
2980    
2981    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2982        compatibility, even though it is no longer used.
2983    
2984    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2985        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2986        windows.h file is present (where different names are used). [This was
2987        reversed later after testing - see 16 below.]
2988    
2989    8.  Changed all #include <config.h> to #include "config.h". There were also
2990        some further <pcre.h> cases that I changed to "pcre.h".
2991    
2992    9.  When pcregrep was used with the --colour option, it missed the line ending
2993        sequence off the lines that it output.
2994    
2995    10. It was pointed out to me that arrays of string pointers cause lots of
2996        relocations when a shared library is dynamically loaded. A technique of
2997        using a single long string with a table of offsets can drastically reduce
2998        these. I have refactored PCRE in four places to do this. The result is
2999        dramatic:
3000    
3001          Originally:                          290
3002          After changing UCP table:            187
3003          After changing error message table:   43
3004          After changing table of "verbs"       36
3005          After changing table of Posix names   22
3006    
3007        Thanks to the folks working on Gregex for glib for this insight.
3008    
3009    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3010        unicode-properties was also set.
3011    
3012    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3013    
3014    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3015        checked only for CRLF.
3016    
3017    14. Added casts to pcretest.c to avoid compiler warnings.
3018    
3019    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3020    
3021    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3022        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3023        entirely. This removes changes made in 7 above.
3024    
3025    17. The CMake files have been updated, and there is now more information about
3026        building with CMake in the NON-UNIX-USE document.
3027    
3028    
3029    Version 7.3 28-Aug-07
3030    ---------------------
3031    
3032     1. In the rejigging of the build system that eventually resulted in 7.1, the
3033        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3034        brackets there is not right, since it causes compilers to look for an
3035        installed pcre.h, not the version that is in the source that is being
3036        compiled (which of course may be different). I have changed it back to:
3037    
3038          #include "pcre.h"
3039    
3040        I have a vague recollection that the change was concerned with compiling in
3041        different directories, but in the new build system, that is taken care of
3042        by the VPATH setting the Makefile.
3043    
3044     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3045        when the subject happened to end in the byte 0x85 (e.g. if the last
3046        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3047        characters but of course it shouldn't be taken as a newline when it is part
3048        of another character. The bug was that, for an unlimited repeat of . in
3049        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3050        characters when looking for a newline.
3051    
3052     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3053    
3054     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3055        in debug output.
3056    
3057     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3058        long printing in the pcrecpp unittest when running under MinGW.
3059    
3060     6. ESC_K was left out of the EBCDIC table.
3061    
3062     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3063        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3064        limit also applies to "virtual nesting" when a pattern is recursive, and in
3065        this case 1000 isn't so big. I have been able to remove this limit at the
3066        expense of backing off one optimization in certain circumstances. Normally,
3067        when pcre_exec() would call its internal match() function recursively and
3068        immediately return the result unconditionally, it uses a "tail recursion"
3069        feature to save stack. However, when a subpattern that can match an empty
3070        string has an unlimited repetition quantifier, it no longer makes this
3071        optimization. That gives it a stack frame in which to save the data for
3072        checking that an empty string has been matched. Previously this was taken
3073        from the 1000-entry workspace that had been reserved. So now there is no
3074        explicit limit, but more stack is used.
3075    
3076     8. Applied Daniel's patches to solve problems with the import/export magic
3077        syntax that is required for Windows, and which was going wrong for the
3078        pcreposix and pcrecpp parts of the library. These were overlooked when this
3079        problem was solved for the main library.
3080    
3081     9. There were some crude static tests to avoid integer overflow when computing
3082        the size of patterns that contain repeated groups with explicit upper
3083        limits. As the maximum quantifier is 65535, the maximum group length was
3084        set at 30,000 so that the product of these two numbers did not overflow a
3085        32-bit integer. However, it turns out that people want to use groups that
3086        are longer than 30,000 bytes (though not repeat them that many times).
3087        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3088        made it possible to implement the integer overflow checks in a much more
3089        dynamic way, which I have now done. The artificial limitation on group
3090        length has been removed - we now have only the limit on the total length of
3091        the compiled pattern, which depends on the LINK_SIZE setting.
3092    
3093    10. Fixed a bug in the documentation for get/copy named substring when
3094        duplicate names are permitted. If none of the named substrings are set, the
3095        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3096        empty string.
3097    
3098    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3099        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3100        because the ] is interpreted as the first data character and the
3101        terminating ] is not found. PCRE has been made compatible with Perl in this
3102        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3103        cause memory overwriting.
3104    
3105    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3106        string has been matched (to stop an infinite loop). It was not recognizing
3107        a conditional subpattern that could match an empty string if that
3108        subpattern was within another subpattern. For example, it looped when
3109        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3110        condition was not nested. This bug has been fixed.
3111    
3112    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3113        past the start of the subject in the presence of bytes with the top bit
3114        set, for example "\x8aBCD".
3115    
3116    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3117        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3118    
3119    14. Optimized (?!) to (*FAIL).
3120    
3121    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3122        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3123        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3124        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3125        does: it's just the validity check that is more restrictive.
3126    
3127    16. Inserted checks for integer overflows during escape sequence (backslash)
3128        processing, and also fixed erroneous offset values for syntax errors during
3129        backslash processing.
3130    
3131    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3132        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3133    
3134    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3135        caused an overrun.
3136    
3137    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3138        something other than just ASCII characters) inside a group that had an
3139        unlimited repeat caused a loop at compile time (while checking to see
3140        whether the group could match an empty string).
3141    
3142    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3143        [\P{Any}] did so. (Error in the code for printing property names.)
3144    
3145    21. An orphan \E inside a character class could cause a crash.
3146    
3147    22. A repeated capturing bracket such as (A)? could cause a wild memory
3148        reference during compilation.
3149    
3150    23. There are several functions in pcre_compile() that scan along a compiled
3151        expression for various reasons (e.g. to see if it's fixed length for look
3152        behind). There were bugs in these functions when a repeated \p or \P was
3153        present in the pattern. These operators have additional parameters compared
3154        with \d, etc, and these were not being taken into account when moving along
3155        the compiled data. Specifically:
3156    
3157        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3158            length.
3159    
3160        (b) An item such as \pL+ within a repeated group could cause crashes or
3161            loops.
3162    
3163        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3164            "reference to non-existent subpattern" error.
3165    
3166        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3167    
3168    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3169        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3170    
3171    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3172    
3173    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3174        character were causing crashes (broken optimization).
3175    
3176    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3177        \p or \P) caused a compile-time loop.
3178    
3179    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3180        break. For example, the unstudied pattern [\r\n]A does not match the string
3181        "\r\nA" because change 7.0/46 below moves the current point on by two
3182        characters after failing to match at the start. However, the pattern \nA
3183        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3184        the same is true. There doesn't seem any very clean way out of this, but
3185        what I have chosen to do makes the common cases work: PCRE now takes note
3186        of whether there can be an explicit match for \r or \n anywhere in the
3187        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3188        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3189        pattern has explicit CR or LF references.
3190    
3191    29. Added (*CR) etc for changing newline setting at start of pattern.
3192    
3193    
3194    Version 7.2 19-Jun-07
3195  ---------------------  ---------------------
3196    
3197   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 3211  Version 7.2 01-May-07
3211      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3212      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
3213      in all cases.      in all cases.
3214    
3215   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3216    
3217      (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
3218          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3219    
3220      (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
3221          to be opened parentheses.          to be opened parentheses.
3222    
3223      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3224          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3225    
3226        (d) \K resets the start of the current match so that everything before
3227            is not part of it.
3228    
3229        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3230    
3231        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3232            reference syntax.
3233    
3234        (g) (?| introduces a group in which the numbering of parentheses in each
3235            alternative starts with the same number.
3236    
3237        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3238    
3239     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3240        PCRE_INFO_JCHANGED.
3241    
3242     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3243        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3244        for detecting groups that can match an empty string.
3245    
3246     9. A pattern with a very large number of alternatives (more than several
3247        hundred) was running out of internal workspace during the pre-compile
3248        phase, where pcre_compile() figures out how much memory will be needed. A
3249        bit of new cunning has reduced the workspace needed for groups with
3250        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3251        workspace instead of running out of the 4096 that are available.
3252    
3253    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3254    
3255    11. Applied patch from Google to remove an optimization that didn't quite work.
3256        The report of the bug said:
3257    
3258          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3259          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3260          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3261    
3262    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3263        it matched the wrong number of bytes.
3264    
3265    
3266  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3267  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1628

  ViewVC Help
Powered by ViewVC 1.1.5