/[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 1636 by ph10, Sat Feb 27 17:38:11 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 xx-xxxxxx-201x
8    ---------------------------
9    
10    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12        misbehaved. This bug was found by the LLVM fuzzer.
13    
14    2.  Similar to the above, if an isolated \E was present between an item and its
15        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16        bug was found by the LLVM fuzzer.
17    
18    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19        working correctly in UCP mode.
20    
21    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22        was set when the pmatch argument was NULL. It now returns REG_INVARG.
23    
24    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25    
26    6.  An empty \Q\E sequence between an item and its qualifier caused
27        pcre_compile() to misbehave when auto callouts were enabled. This bug was
28        found by the LLVM fuzzer.
29    
30    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
31        space or a #-type comment that was followed by (?-x), which turns off
32        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33        pcre_compile() assumed that (?-x) applied to the whole pattern and
34        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
35    
36    8.  An call of pcre_copy_named_substring() for a named substring whose number
37        was greater than the space in the ovector could cause a crash.
38    
39    9.  Yet another buffer overflow bug involved duplicate named groups with a
40        group that reset capture numbers (compare 8.38/7 below). Once again, I have
41        just allowed for more memory, even if not needed. (A proper fix is
42        implemented in PCRE2, but it involves a lot of refactoring.)
43    
44    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45        start of the match to be earlier than the end.
46    
47    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
48    
49    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50        assertion, caused pcretest to generate incorrect output, and also to read
51        uninitialized memory (detected by ASAN or valgrind).
52    
53    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
54        nested set of parentheses of sufficient size caused an overflow of the
55        compiling workspace (which was diagnosed, but of course is not desirable).
56    
57    14. And yet another buffer overflow bug involving duplicate named groups, this
58        time nested, with a nested back reference. Yet again, I have just allowed
59        for more memory, because anything more needs all the refactoring that has
60        been done for PCRE2. An example pattern that provoked this bug is:
61        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
62        registered as CVE-2016-1283.
63    
64    
65    Version 8.38 23-November-2015
66    -----------------------------
67    
68    1.  If a group that contained a recursive back reference also contained a
69        forward reference subroutine call followed by a non-forward-reference
70        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
71        compile correct code, leading to undefined behaviour or an internally
72        detected error. This bug was discovered by the LLVM fuzzer.
73    
74    2.  Quantification of certain items (e.g. atomic back references) could cause
75        incorrect code to be compiled when recursive forward references were
76        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
77        This bug was discovered by the LLVM fuzzer.
78    
79    3.  A repeated conditional group whose condition was a reference by name caused
80        a buffer overflow if there was more than one group with the given name.
81        This bug was discovered by the LLVM fuzzer.
82    
83    4.  A recursive back reference by name within a group that had the same name as
84        another group caused a buffer overflow. For example:
85        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
86    
87    5.  A forward reference by name to a group whose number is the same as the
88        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
89        a buffer overflow at compile time. This bug was discovered by the LLVM
90        fuzzer.
91    
92    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
93        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
94    
95    7.  Another buffer overflow bug involved duplicate named groups with a
96        reference between their definition, with a group that reset capture
97        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
98        fixed by always allowing for more memory, even if not needed. (A proper fix
99        is implemented in PCRE2, but it involves more refactoring.)
100    
101    8.  There was no check for integer overflow in subroutine calls such as (?123).
102    
103    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
104        being treated as a literal 'l' instead of causing an error.
105    
106    10. There was a buffer overflow if pcre_exec() was called with an ovector of
107        size 1. This bug was found by american fuzzy lop.
108    
109    11. If a non-capturing group containing a conditional group that could match
110        an empty string was repeated, it was not identified as matching an empty
111        string itself. For example: /^(?:(?(1)x|)+)+$()/.
112    
113    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
114        \a and \e in test subject lines.
115    
116    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
117        instead of the EBCDIC value.
118    
119    14. The handling of \c in an EBCDIC environment has been revised so that it is
120        now compatible with the specification in Perl's perlebcdic page.
121    
122    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
123        ASCII/Unicode. This has now been added to the list of characters that are
124        recognized as white space in EBCDIC.
125    
126    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
127        error (correctly) when used outside a class, but did not give an error
128        within a class.
129    
130    17. \h within a class was incorrectly compiled in EBCDIC environments.
131    
132    18. A pattern with an unmatched closing parenthesis that contained a backward
133        assertion which itself contained a forward reference caused buffer
134        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
135    
136    19. JIT should return with error when the compiled pattern requires more stack
137        space than the maximum.
138    
139    20. A possessively repeated conditional group that could match an empty string,
140        for example, /(?(R))*+/, was incorrectly compiled.
141    
142    21. Fix infinite recursion in the JIT compiler when certain patterns such as
143        /(?:|a|){100}x/ are analysed.
144    
145    22. Some patterns with character classes involving [: and \\ were incorrectly
146        compiled and could cause reading from uninitialized memory or an incorrect
147        error diagnosis.
148    
149    23. Pathological patterns containing many nested occurrences of [: caused
150        pcre_compile() to run for a very long time.
151    
152    24. A conditional group with only one branch has an implicit empty alternative
153        branch and must therefore be treated as potentially matching an empty
154        string.
155    
156    25. If (?R was followed by - or + incorrect behaviour happened instead of a
157        diagnostic.
158    
159    26. Arrange to give up on finding the minimum matching length for overly
160        complex patterns.
161    
162    27. Similar to (4) above: in a pattern with duplicated named groups and an
163        occurrence of (?| it is possible for an apparently non-recursive back
164        reference to become recursive if a later named group with the relevant
165        number is encountered. This could lead to a buffer overflow. Wen Guanxing
166        from Venustech ADLAB discovered this bug.
167    
168    28. If pcregrep was given the -q option with -c or -l, or when handling a
169        binary file, it incorrectly wrote output to stdout.
170    
171    29. The JIT compiler did not restore the control verb head in case of *THEN
172        control verbs. This issue was found by Karl Skomski with a custom LLVM
173        fuzzer.
174    
175    30. Error messages for syntax errors following \g and \k were giving inaccurate
176        offsets in the pattern.
177    
178    31. Added a check for integer overflow in conditions (?(<digits>) and
179        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
180        fuzzer.
181    
182    32. Handling recursive references such as (?2) when the reference is to a group
183        later in the pattern uses code that is very hacked about and error-prone.
184        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
185        give an internal error if it is obvious that compiling has gone wrong.
186    
187    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
188        This issue was found by Karl Skomski with a custom LLVM fuzzer.
189    
190    34. The JIT compiler should restore the control chain for empty possessive
191        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
192    
193    35. Match limit check added to JIT recursion. This issue was found by Karl
194        Skomski with a custom LLVM fuzzer.
195    
196    36. Yet another case similar to 27 above has been circumvented by an
197        unconditional allocation of extra memory. This issue is fixed "properly" in
198        PCRE2 by refactoring the way references are handled. Wen Guanxing
199        from Venustech ADLAB discovered this bug.
200    
201    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
202        with a custom LLVM fuzzer.
203    
204    38. Fixed a corner case of range optimization in JIT.
205    
206    39. An incorrect error "overran compiling workspace" was given if there were
207        exactly enough group forward references such that the last one extended
208        into the workspace safety margin. The next one would have expanded the
209        workspace. The test for overflow was not including the safety margin.
210    
211    40. A match limit issue is fixed in JIT which was found by Karl Skomski
212        with a custom LLVM fuzzer.
213    
214    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
215        work under Windows. (Why has it taken so long for anyone to notice?)
216    
217    42. In a character class such as [\W\p{Any}] where both a negative-type escape
218        ("not a word character") and a property escape were present, the property
219        escape was being ignored.
220    
221    43. Fix crash caused by very long (*MARK) or (*THEN) names.
222    
223    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
224        by a single ASCII character in a class item, was incorrectly compiled in
225        UCP mode. The POSIX class got lost, but only if the single character
226        followed it.
227    
228    45. [:punct:] in UCP mode was matching some characters in the range 128-255
229        that should not have been matched.
230    
231    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
232        class, all characters with code points greater than 255 are in the class.
233        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
234        such as \w are turned into Unicode properties), wide characters were not
235        correctly handled, and could fail to match.
236    
237    
238    Version 8.37 28-April-2015
239    --------------------------
240    
241    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
242        for those parentheses to be closed with whatever has been captured so far.
243        However, it was failing to mark any other groups between the hightest
244        capture so far and the currrent group as "unset". Thus, the ovector for
245        those groups contained whatever was previously there. An example is the
246        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
247    
248    2.  If an assertion condition was quantified with a minimum of zero (an odd
249        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
250    
251    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
252        unrecognized modifier, a crash could occur.
253    
254    4.  An attempt to do global matching in pcretest with a zero-length ovector
255        caused a crash.
256    
257    5.  Fixed a memory leak during matching that could occur for a subpattern
258        subroutine call (recursive or otherwise) if the number of captured groups
259        that had to be saved was greater than ten.
260    
261    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
262        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
263        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
264    
265    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
266        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
267    
268    8. When a pattern is compiled, it remembers the highest back reference so that
269       when matching, if the ovector is too small, extra memory can be obtained to
270       use instead. A conditional subpattern whose condition is a check on a
271       capture having happened, such as, for example in the pattern
272       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
273       setting the highest backreference number. This mattered only if pcre_exec()
274       was called with an ovector that was too small to hold the capture, and there
275       was no other kind of back reference (a situation which is probably quite
276       rare). The effect of the bug was that the condition was always treated as
277       FALSE when the capture could not be consulted, leading to a incorrect
278       behaviour by pcre_exec(). This bug has been fixed.
279    
280    9. A reference to a duplicated named group (either a back reference or a test
281       for being set in a conditional) that occurred in a part of the pattern where
282       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
283       to be incorrectly calculated, leading to overwriting.
284    
285    10. A mutually recursive set of back references such as (\2)(\1) caused a
286        segfault at study time (while trying to find the minimum matching length).
287        The infinite loop is now broken (with the minimum length unset, that is,
288        zero).
289    
290    11. If an assertion that was used as a condition was quantified with a minimum
291        of zero, matching went wrong. In particular, if the whole group had
292        unlimited repetition and could match an empty string, a segfault was
293        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
294        assertions to be quantified, but not if they are being used as conditions,
295        so the above pattern is faulted by Perl. PCRE has now been changed so that
296        it also rejects such patterns.
297    
298    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
299        failed to allow the zero-repeat case if pcre2_exec() was called with an
300        ovector too small to capture the group.
301    
302    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
303        Red Hat Product Security:
304    
305        (a) A crash if /K and /F were both set with the option to save the compiled
306        pattern.
307    
308        (b) Another crash if the option to print captured substrings in a callout
309        was combined with setting a null ovector, for example \O\C+ as a subject
310        string.
311    
312    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
313        forward reference repeated a large (but limited) number of times within a
314        repeated outer group that has a zero minimum quantifier, caused incorrect
315        code to be compiled, leading to the error "internal error:
316        previously-checked referenced subpattern not found" when an incorrect
317        memory address was read. This bug was reported as "heap overflow",
318        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
319        CVE-2015-2325.
320    
321    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
322        call within a group that also contained a recursive back reference caused
323        incorrect code to be compiled. This bug was reported as "heap overflow",
324        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
325        number CVE-2015-2326.
326    
327    24. Computing the size of the JIT read-only data in advance has been a source
328        of various issues, and new ones are still appear unfortunately. To fix
329        existing and future issues, size computation is eliminated from the code,
330        and replaced by on-demand memory allocation.
331    
332    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
333        adjacent to the end of the range, and the range contained characters with
334        more than one other case, caused incorrect behaviour when compiled in UTF
335        mode. In that example, the range a-j was left out of the class.
336    
337    26. Fix JIT compilation of conditional blocks, which assertion
338        is converted to (*FAIL). E.g: /(?(?!))/.
339    
340    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
341        discovered by the LLVM fuzzer.
342    
343    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
344        when this assertion was used as a condition, for example (?(?!)a|b). In
345        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
346        error about an unsupported item.
347    
348    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
349        possessification code could take exponential time to complete. A recursion
350        depth limit of 1000 has been imposed to limit the resources used by this
351        optimization.
352    
353    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
354        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
355        because \S ensures they are all in the class. The code for doing this was
356        interacting badly with the code for computing the amount of space needed to
357        compile the pattern, leading to a buffer overflow. This bug was discovered
358        by the LLVM fuzzer.
359    
360    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
361        other kinds of group caused stack overflow at compile time. This bug was
362        discovered by the LLVM fuzzer.
363    
364    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
365        between a subroutine call and its quantifier was incorrectly compiled,
366        leading to buffer overflow or other errors. This bug was discovered by the
367        LLVM fuzzer.
368    
369    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
370        assertion after (?(. The code was failing to check the character after
371        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
372        was discovered by the LLVM fuzzer.
373    
374    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
375        a fixed maximum following a group that contains a subroutine reference was
376        incorrectly compiled and could trigger buffer overflow. This bug was
377        discovered by the LLVM fuzzer.
378    
379    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
380        caused a stack overflow instead of the diagnosis of a non-fixed length
381        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
382    
383    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
384        (e.g. /(?<=\Ka)/) could make pcregrep loop.
385    
386    37. There was a similar problem to 36 in pcretest for global matches.
387    
388    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
389        and a subsequent item in the pattern caused a non-match, backtracking over
390        the repeated \X did not stop, but carried on past the start of the subject,
391        causing reference to random memory and/or a segfault. There were also some
392        other cases where backtracking after \C could crash. This set of bugs was
393        discovered by the LLVM fuzzer.
394    
395    39. The function for finding the minimum length of a matching string could take
396        a very long time if mutual recursion was present many times in a pattern,
397        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
398        method has been implemented. This infelicity was discovered by the LLVM
399        fuzzer.
400    
401    40. Static linking against the PCRE library using the pkg-config module was
402        failing on missing pthread symbols.
403    
404    
405    Version 8.36 26-September-2014
406    ------------------------------
407    
408    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
409        -Wmissing-field-initializers and -Wunused-parameter.
410    
411    2.  The tests for quantifiers being too big (greater than 65535) were being
412        applied after reading the number, and stupidly assuming that integer
413        overflow would give a negative number. The tests are now applied as the
414        numbers are read.
415    
416    3.  Tidy code in pcre_exec.c where two branches that used to be different are
417        now the same.
418    
419    4.  The JIT compiler did not generate match limit checks for certain
420        bracketed expressions with quantifiers. This may lead to exponential
421        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
422        issue should be resolved now.
423    
424    5.  Fixed an issue, which occures when nested alternatives are optimized
425        with table jumps.
426    
427    6.  Inserted two casts and changed some ints to size_t in the light of some
428        reported 64-bit compiler warnings (Bugzilla 1477).
429    
430    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
431        an empty string, which sometimes were behaving incorrectly in the
432        interpreter (though correctly in the JIT matcher). This pcretest input is
433        an example:
434    
435          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
436          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
437    
438        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
439        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
440        for an empty string was breaking the inner loop and carrying on at a lower
441        level, when possessive repeated groups should always return to a higher
442        level as they have no backtrack points in them. The empty string test now
443        occurs at the outer level.
444    
445    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
446        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
447    
448    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
449        doesn't).
450    
451    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
452        to hold a pattern to be incorrectly computed (too small) when there were
453        named back references to duplicated names. This could cause "internal
454        error: code overflow" or "double free or corruption" or other memory
455        handling errors.
456    
457    11. When named subpatterns had the same prefixes, back references could be
458        confused. For example, in this pattern:
459    
460          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
461    
462        the reference to 'Name' was incorrectly treated as a reference to a
463        duplicate name.
464    
465    12. A pattern such as /^s?c/mi8 where the optional character has more than
466        one "other case" was incorrectly compiled such that it would only try to
467        match starting at "c".
468    
469    13. When a pattern starting with \s was studied, VT was not included in the
470        list of possible starting characters; this should have been part of the
471        8.34/18 patch.
472    
473    14. If a character class started [\Qx]... where x is any character, the class
474        was incorrectly terminated at the ].
475    
476    15. If a pattern that started with a caseless match for a character with more
477        than one "other case" was studied, PCRE did not set up the starting code
478        unit bit map for the list of possible characters. Now it does. This is an
479        optimization improvement, not a bug fix.
480    
481    16. The Unicode data tables have been updated to Unicode 7.0.0.
482    
483    17. Fixed a number of memory leaks in pcregrep.
484    
485    18. Avoid a compiler warning (from some compilers) for a function call with
486        a cast that removes "const" from an lvalue by using an intermediate
487        variable (to which the compiler does not object).
488    
489    19. Incorrect code was compiled if a group that contained an internal recursive
490        back reference was optional (had quantifier with a minimum of zero). This
491        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
492        caused segmentation faults because of stack overflows at compile time.
493    
494    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
495        group that is quantified with an indefinite repeat, caused a compile-time
496        loop which used up all the system stack and provoked a segmentation fault.
497        This was not the same bug as 19 above.
498    
499    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
500        Patch by Mike Frysinger.
501    
502    
503    Version 8.35 04-April-2014
504    --------------------------
505    
506    1.  A new flag is set, when property checks are present in an XCLASS.
507        When this flag is not set, PCRE can perform certain optimizations
508        such as studying these XCLASS-es.
509    
510    2.  The auto-possessification of character sets were improved: a normal
511        and an extended character set can be compared now. Furthermore
512        the JIT compiler optimizes more character set checks.
513    
514    3.  Got rid of some compiler warnings for potentially uninitialized variables
515        that show up only when compiled with -O2.
516    
517    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
518        of a match later then the end of the match. The pcretest program was not
519        handling the case sensibly - it was outputting from the start to the next
520        binary zero. It now reports this situation in a message, and outputs the
521        text from the end to the start.
522    
523    5.  Fast forward search is improved in JIT. Instead of the first three
524        characters, any three characters with fixed position can be searched.
525        Search order: first, last, middle.
526    
527    6.  Improve character range checks in JIT. Characters are read by an inprecise
528        function now, which returns with an unknown value if the character code is
529        above a certain threshold (e.g: 256). The only limitation is that the value
530        must be bigger than the threshold as well. This function is useful when
531        the characters above the threshold are handled in the same way.
532    
533    7.  The macros whose names start with RAWUCHAR are placeholders for a future
534        mode in which only the bottom 21 bits of 32-bit data items are used. To
535        make this more memorable for those maintaining the code, the names have
536        been changed to start with UCHAR21, and an extensive comment has been added
537        to their definition.
538    
539    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
540        to the export list in Makefile.am (they were accidentally omitted from the
541        8.34 tarball).
542    
543    9.  The informational output from pcretest used the phrase "starting byte set"
544        which is inappropriate for the 16-bit and 32-bit libraries. As the output
545        for "first char" and "need char" really means "non-UTF-char", I've changed
546        "byte" to "char", and slightly reworded the output. The documentation about
547        these values has also been (I hope) clarified.
548    
549    10. Another JIT related optimization: use table jumps for selecting the correct
550        backtracking path, when more than four alternatives are present inside a
551        bracket.
552    
553    11. Empty match is not possible, when the minimum length is greater than zero,
554        and there is no \K in the pattern. JIT should avoid empty match checks in
555        such cases.
556    
557    12. In a caseless character class with UCP support, when a character with more
558        than one alternative case was not the first character of a range, not all
559        the alternative cases were added to the class. For example, s and \x{17f}
560        are both alternative cases for S: the class [RST] was handled correctly,
561        but [R-T] was not.
562    
563    13. The configure.ac file always checked for pthread support when JIT was
564        enabled. This is not used in Windows, so I have put this test inside a
565        check for the presence of windows.h (which was already tested for).
566    
567    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
568        The algorithm provides a way to skip certain starting offsets, and usually
569        faster than linear prefix searches.
570    
571    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
572        as for 'fr_FR' and 'french'. For some reason, however, it then used the
573        Windows-specific input and output files, which have 'french' screwed in.
574        So this could never have worked. One of the problems with locales is that
575        they aren't always the same. I have now updated RunTest so that it checks
576        the output of the locale test (test 3) against three different output
577        files, and it allows the test to pass if any one of them matches. With luck
578        this should make the test pass on some versions of Solaris where it was
579        failing. Because of the uncertainty, the script did not used to stop if
580        test 3 failed; it now does. If further versions of a French locale ever
581        come to light, they can now easily be added.
582    
583    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
584        there was a message during ./configure, but it did not stop. This now
585        provokes an error. The invalid example in README has been corrected.
586        If a value less than the minimum is given, the minimum value has always
587        been used, but now a warning is given.
588    
589    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
590        was a bug in the test system, which is now fixed. Also, the list of various
591        configurations that are tested for each release did not have one with both
592        16/32 bits and --enable-bar-anycrlf. It now does.
593    
594    18. pcretest was missing "-C bsr" for displaying the \R default setting.
595    
596    19. Little endian PowerPC systems are supported now by the JIT compiler.
597    
598    20. The fast forward newline mechanism could enter to an infinite loop on
599        certain invalid UTF-8 input. Although we don't support these cases
600        this issue can be fixed by a performance optimization.
601    
602    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
603        not take account if existing stack usage. There is now a new global
604        variable called pcre_stack_guard that can be set to point to an external
605        function to check stack availability. It is called at the start of
606        processing every parenthesized group.
607    
608    22. A typo in the code meant that in ungreedy mode the max/min qualifier
609        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
610        not match "ab".
611    
612    23. When UTF was disabled, the JIT program reported some incorrect compile
613        errors. These messages are silenced now.
614    
615    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
616        compiler.
617    
618    25. Change all the temporary files used in RunGrepTest to be different to those
619        used by RunTest so that the tests can be run simultaneously, for example by
620        "make -j check".
621    
622    
623    Version 8.34 15-December-2013
624    -----------------------------
625    
626    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
627        executable memory. Patch inspired by Carsten Klein.
628    
629    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
630        this macro is never tested and has no effect, because the work to support
631        coverage involves only compiling and linking options and special targets in
632        the Makefile. The comment in config.h implied that defining the macro would
633        enable coverage support, which is totally false. There was also support for
634        setting this macro in the CMake files (my fault, I just copied it from
635        configure). SUPPORT_GCOV has now been removed.
636    
637    3.  Make a small performance improvement in strlen16() and strlen32() in
638        pcretest.
639    
640    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
641        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
642        these cases). There was also one in pcretest.c.
643    
644    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
645    
646    6.  In UTF mode, the code for checking whether a group could match an empty
647        string (which is used for indefinitely repeated groups to allow for
648        breaking an infinite loop) was broken when the group contained a repeated
649        negated single-character class with a character that occupied more than one
650        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
651        UTF-8 mode). The effect was undefined: the group might or might not be
652        deemed as matching an empty string, or the program might have crashed.
653    
654    7.  The code for checking whether a group could match an empty string was not
655        recognizing that \h, \H, \v, \V, and \R must match a character.
656    
657    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
658        an empty string. If it can, pcretest shows this in its information output.
659    
660    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
661        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
662        matched by pcre_exec() without using JIT:
663    
664        (a) If the rest of the pattern did not match after a maximal run of
665            grapheme clusters, the code for backing up to try with fewer of them
666            did not always back up over a full grapheme when characters that do not
667            have the modifier quality were involved, e.g. Hangul syllables.
668    
669        (b) If the match point in a subject started with modifier character, and
670            there was no match, the code could incorrectly back up beyond the match
671            point, and potentially beyond the first character in the subject,
672            leading to a segfault or an incorrect match result.
673    
674    10. A conditional group with an assertion condition could lead to PCRE
675        recording an incorrect first data item for a match if no other first data
676        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
677        first data item, and therefore matched "ca" after "c" instead of at the
678        start.
679    
680    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
681        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
682    
683    12. The source of pcregrep now includes z/OS-specific code so that it can be
684        compiled for z/OS as part of the special z/OS distribution.
685    
686    13. Added the -T and -TM options to pcretest.
687    
688    14. The code in pcre_compile.c for creating the table of named capturing groups
689        has been refactored. Instead of creating the table dynamically during the
690        actual compiling pass, the information is remembered during the pre-compile
691        pass (on the stack unless there are more than 20 named groups, in which
692        case malloc() is used) and the whole table is created before the actual
693        compile happens. This has simplified the code (it is now nearly 150 lines
694        shorter) and prepared the way for better handling of references to groups
695        with duplicate names.
696    
697    15. A back reference to a named subpattern when there is more than one of the
698        same name now checks them in the order in which they appear in the pattern.
699        The first one that is set is used for the reference. Previously only the
700        first one was inspected. This change makes PCRE more compatible with Perl.
701    
702    16. Unicode character properties were updated from Unicode 6.3.0.
703    
704    17. The compile-time code for auto-possessification has been refactored, based
705        on a patch by Zoltan Herczeg. It now happens after instead of during
706        compilation. The code is cleaner, and more cases are handled. The option
707        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
708        options in pcretest are provided to set it. It can also be set by
709        (*NO_AUTO_POSSESS) at the start of a pattern.
710    
711    18. The character VT has been added to the default ("C" locale) set of
712        characters that match \s and are generally treated as white space,
713        following this same change in Perl 5.18. There is now no difference between
714        "Perl space" and "POSIX space". Whether VT is treated as white space in
715        other locales depends on the locale.
716    
717    19. The code for checking named groups as conditions, either for being set or
718        for being recursed, has been refactored (this is related to 14 and 15
719        above). Processing unduplicated named groups should now be as fast at
720        numerical groups, and processing duplicated groups should be faster than
721        before.
722    
723    20. Two patches to the CMake build system, by Alexander Barkov:
724    
725          (1) Replace the "source" command by "." in CMakeLists.txt because
726              "source" is a bash-ism.
727    
728          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
729              without these the CMake build does not work on Solaris.
730    
731    21. Perl has changed its handling of \8 and \9. If there is no previously
732        encountered capturing group of those numbers, they are treated as the
733        literal characters 8 and 9 instead of a binary zero followed by the
734        literals. PCRE now does the same.
735    
736    22. Following Perl, added \o{} to specify codepoints in octal, making it
737        possible to specify values greater than 0777 and also making them
738        unambiguous.
739    
740    23. Perl now gives an error for missing closing braces after \x{... instead of
741        treating the string as literal. PCRE now does the same.
742    
743    24. RunTest used to grumble if an inappropriate test was selected explicitly,
744        but just skip it when running all tests. This make it awkward to run ranges
745        of tests when one of them was inappropriate. Now it just skips any
746        inappropriate tests, as it always did when running all tests.
747    
748    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
749        character types such as \d or \w, too many callouts were inserted, and the
750        data that they returned was rubbish.
751    
752    26. In UCP mode, \s was not matching two of the characters that Perl matches,
753        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
754        were matched by \h. The code has now been refactored so that the lists of
755        the horizontal and vertical whitespace characters used for \h and \v (which
756        are defined only in one place) are now also used for \s.
757    
758    27. Add JIT support for the 64 bit TileGX architecture.
759        Patch by Jiong Wang (Tilera Corporation).
760    
761    28. Possessive quantifiers for classes (both explicit and automatically
762        generated) now use special opcodes instead of wrapping in ONCE brackets.
763    
764    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
765        (because it's meaningless), this was not happening when PCRE_CASELESS was
766        set. Not wrong, but inefficient.
767    
768    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
769        properties for \w, \d, etc) is present in a test regex. Otherwise if the
770        test contains no characters greater than 255, Perl doesn't realise it
771        should be using Unicode semantics.
772    
773    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
774        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
775        does in Unicode mode.
776    
777    32. Added the "forbid" facility to pcretest so that putting tests into the
778        wrong test files can sometimes be quickly detected.
779    
780    33. There is now a limit (default 250) on the depth of nesting of parentheses.
781        This limit is imposed to control the amount of system stack used at compile
782        time. It can be changed at build time by --with-parens-nest-limit=xxx or
783        the equivalent in CMake.
784    
785    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
786        errors. Perl warns for these when in warning mode, but PCRE has no facility
787        for giving warnings.
788    
789    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
790        However, this was not working for (?!) because it is optimized to (*FAIL),
791        for which PCRE does not allow quantifiers. The optimization is now disabled
792        when a quantifier follows (?!). I can't see any use for this, but it makes
793        things uniform.
794    
795    36. Perl no longer allows group names to start with digits, so I have made this
796        change also in PCRE. It simplifies the code a bit.
797    
798    37. In extended mode, Perl ignores spaces before a + that indicates a
799        possessive quantifier. PCRE allowed a space before the quantifier, but not
800        before the possessive +. It now does.
801    
802    38. The use of \K (reset reported match start) within a repeated possessive
803        group such as (a\Kb)*+ was not working.
804    
805    40. Document that the same character tables must be used at compile time and
806        run time, and that the facility to pass tables to pcre_exec() and
807        pcre_dfa_exec() is for use only with saved/restored patterns.
808    
809    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
810        features for Builds with MSVC:
811    
812        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
813           builds). The .rc files can be used to set FileDescription and many other
814           attributes.
815    
816        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
817           This allows higher-level build scripts which want .pdb files to avoid
818           hard-coding the exact files needed."
819    
820    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
821        mean "start of word" and "end of word", respectively, as a transition aid.
822    
823    43. A minimizing repeat of a class containing codepoints greater than 255 in
824        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
825        compiled to use the heap for recursion.
826    
827    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
828        is configured.
829    
830    
831    Version 8.33 28-May-2013
832    ------------------------
833    
834    1.  Added 'U' to some constants that are compared to unsigned integers, to
835        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
836        variables that are added to signed variables, to ensure the result is
837        signed and can be negated.
838    
839    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
840        CMake config files.
841    
842    3.  Revise the creation of config.h.generic so that all boolean macros are
843        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
844        overriding via -D on the command line possible.
845    
846    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
847        to unsigned int is reported to make a quite noticeable speed difference in
848        a specific Windows environment. Testing on Linux did also appear to show
849        some benefit (and it is clearly not harmful). Also fixed the definition of
850        Xop which should be unsigned.
851    
852    5.  Related to (4), changing the definition of the intermediate variable cc
853        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
854        improvements.
855    
856    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
857        unnecessary spaces.
858    
859    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
860        and later.
861    
862    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
863    
864    9.  Optimizing fast_forward_start_bits in JIT.
865    
866    10. Adding support for callouts in JIT, and fixing some issues revealed
867        during this work. Namely:
868    
869        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
870    
871        (b) Minimum length was not checked before the matching is started.
872    
873    11. The value of capture_last that is passed to callouts was incorrect in some
874        cases when there was a capture on one path that was subsequently abandoned
875        after a backtrack. Also, the capture_last value is now reset after a
876        recursion, since all captures are also reset in this case.
877    
878    12. The interpreter no longer returns the "too many substrings" error in the
879        case when an overflowing capture is in a branch that is subsequently
880        abandoned after a backtrack.
881    
882    13. In the pathological case when an offset vector of size 2 is used, pcretest
883        now prints out the matched string after a yield of 0 or 1.
884    
885    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
886        Only supported by the JIT compiler at the moment.
887    
888    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
889    
890    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
891        offset of the starting point of the matching process, provided the offsets
892        vector is large enough.
893    
894    17. The \A escape now records a lookbehind value of 1, though its execution
895        does not actually inspect the previous character. This is to ensure that,
896        in partial multi-segment matching, at least one character from the old
897        segment is retained when a new segment is processed. Otherwise, if there
898        are no lookbehinds in the pattern, \A might match incorrectly at the start
899        of a new segment.
900    
901    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
902    
903    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
904        gives some modest performance improvement in 8-bit mode.
905    
906    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
907        be expressed in certain programming languages using Universal Character
908        Names.
909    
910    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
911        which points out that "non characters" are not "characters that may not
912        appear in Unicode strings" but rather "characters that are reserved for
913        internal use and have only local meaning".
914    
915    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
916        there was a conditional group that depended on an assertion, if the
917        assertion was false, the callout that immediately followed the alternation
918        in the condition was skipped when pcre_exec() was used for matching.
919    
920    23. Allow an explicit callout to be inserted before an assertion that is the
921        condition for a conditional group, for compatibility with automatic
922        callouts, which always insert a callout at this point.
923    
924    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
925        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
926    
927    25. (*PRUNE) is now supported by the JIT compiler.
928    
929    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
930    
931    27. Fix the case where there are two or more SKIPs with arguments that may be
932        ignored.
933    
934    28. (*SKIP) is now supported by the JIT compiler.
935    
936    29. (*THEN) is now supported by the JIT compiler.
937    
938    30. Update RunTest with additional test selector options.
939    
940    31. The way PCRE handles backtracking verbs has been changed in two ways.
941    
942        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
943        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
944        backtracking. In some cases this makes it more Perl-compatible, but Perl's
945        rather obscure rules do not always do the same thing.
946    
947        (2) Previously, backtracking verbs were confined within assertions. This is
948        no longer the case for positive assertions, except for (*ACCEPT). Again,
949        this sometimes improves Perl compatibility, and sometimes does not.
950    
951    32. A number of tests that were in test 2 because Perl did things differently
952        have been moved to test 1, because either Perl or PCRE has changed, and
953        these tests are now compatible.
954    
955    32. Backtracking control verbs are now handled in the same way in JIT and
956        interpreter.
957    
958    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
959        contained a forward subroutine reference caused a compile error.
960    
961    34. Auto-detect and optimize limited repetitions in JIT.
962    
963    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
964        blocking (*UTF) etc.
965    
966    36. In the interpreter, maximizing pattern repetitions for characters and
967        character types now use tail recursion, which reduces stack usage.
968    
969    37. The value of the max lookbehind was not correctly preserved if a compiled
970        and saved regex was reloaded on a host of different endianness.
971    
972    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
973        of the compiled pattern block, expand the flags field from 16 to 32 bits
974        because it was almost full.
975    
976    39. Try madvise first before posix_madvise.
977    
978    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
979        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
980        original change locked out all matches of empty strings. This has been
981        changed so that one match of an empty string per line is recognized.
982        Subsequent searches on the same line (for colouring or for --only-matching,
983        for example) do not recognize empty strings.
984    
985    41. Applied a user patch to fix a number of spelling mistakes in comments.
986    
987    42. Data lines longer than 65536 caused pcretest to crash.
988    
989    43. Clarified the data type for length and startoffset arguments for pcre_exec
990        and pcre_dfa_exec in the function-specific man pages, where they were
991        explicitly stated to be in bytes, never having been updated. I also added
992        some clarification to the pcreapi man page.
993    
994    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
995        a segmentation fault.
996    
997    
998    Version 8.32 30-November-2012
999    -----------------------------
1000    
1001    1.  Improved JIT compiler optimizations for first character search and single
1002        character iterators.
1003    
1004    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1005        Patch by Daniel Richard G.
1006    
1007    3.  Single character iterator optimizations in the JIT compiler.
1008    
1009    4.  Improved JIT compiler optimizations for character ranges.
1010    
1011    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1012        Reported by Giuseppe D'Angelo.
1013    
1014    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1015        of a line, was being set incorrectly in cases where .* appeared inside
1016        atomic brackets at the start of a pattern, or where there was a subsequent
1017        *PRUNE or *SKIP.
1018    
1019    7.  Improved instruction cache flush for POWER/PowerPC.
1020        Patch by Daniel Richard G.
1021    
1022    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1023        grep:
1024    
1025        (a) There is now no limit to the number of patterns to be matched.
1026    
1027        (b) An error is given if a pattern is too long.
1028    
1029        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1030            are now supported.
1031    
1032        (d) --exclude-from and --include-from (multiple use) have been added.
1033    
1034        (e) Exclusions and inclusions now apply to all files and directories, not
1035            just to those obtained from scanning a directory recursively.
1036    
1037        (f) Multiple uses of -f and --file-list are now supported.
1038    
1039        (g) In a Windows environment, the default for -d has been changed from
1040            "read" (the GNU grep default) to "skip", because otherwise the presence
1041            of a directory in the file list provokes an error.
1042    
1043        (h) The documentation has been revised and clarified in places.
1044    
1045    9.  Improve the matching speed of capturing brackets.
1046    
1047    10. Changed the meaning of \X so that it now matches a Unicode extended
1048        grapheme cluster.
1049    
1050    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1051        out POSIX threads when JIT support is configured.
1052    
1053    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1054    
1055    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1056        structure could go wrong in environments where size_t is not the same size
1057        as int.
1058    
1059    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1060        set.
1061    
1062    15. The EBCDIC support had decayed; later updates to the code had included
1063        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1064        general tidy up of EBCDIC-related issues, and the documentation was also
1065        not quite right. There is now a test that can be run on ASCII systems to
1066        check some of the EBCDIC-related things (but is it not a full test).
1067    
1068    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1069        in a small tidy to the code.
1070    
1071    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1072    
1073    18. If the --only-matching (-o) option in pcregrep is specified multiple
1074        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1075        substrings matched by the 1st and 2nd capturing parentheses. A separating
1076        string can be specified by --om-separator (default empty).
1077    
1078    19. Improving the first n character searches.
1079    
1080    20. Turn case lists for horizontal and vertical white space into macros so that
1081        they are defined only once.
1082    
1083    21. This set of changes together give more compatible Unicode case-folding
1084        behaviour for characters that have more than one other case when UCP
1085        support is available.
1086    
1087        (a) The Unicode property table now has offsets into a new table of sets of
1088            three or more characters that are case-equivalent. The MultiStage2.py
1089            script that generates these tables (the pcre_ucd.c file) now scans
1090            CaseFolding.txt instead of UnicodeData.txt for character case
1091            information.
1092    
1093        (b) The code for adding characters or ranges of characters to a character
1094            class has been abstracted into a generalized function that also handles
1095            case-independence. In UTF-mode with UCP support, this uses the new data
1096            to handle characters with more than one other case.
1097    
1098        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1099            whose other case is greater than 256 are now correctly matched
1100            caselessly. Previously, the high codepoint matched the low one, but not
1101            vice versa.
1102    
1103        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1104            of the new class addition function, using character lists defined as
1105            macros alongside the case definitions of 20 above.
1106    
1107        (e) Caseless back references now work with characters that have more than
1108            one other case.
1109    
1110        (f) General caseless matching of characters with more than one other case
1111            is supported.
1112    
1113    22. Unicode character properties were updated from Unicode 6.2.0
1114    
1115    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1116    
1117    24. Add support for 32-bit character strings, and UTF-32
1118    
1119    25. Major JIT compiler update (code refactoring and bugfixing).
1120        Experimental Sparc 32 support is added.
1121    
1122    26. Applied a modified version of Daniel Richard G's patch to create
1123        pcre.h.generic and config.h.generic by "make" instead of in the
1124        PrepareRelease script.
1125    
1126    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1127        pcre_compile.c when checking for a zero character.
1128    
1129    28. Introducing a native interface for JIT. Through this interface, the compiled
1130        machine code can be directly executed. The purpose of this interface is to
1131        provide fast pattern matching, so several sanity checks are not performed.
1132        However, feature tests are still performed. The new interface provides
1133        1.4x speedup compared to the old one.
1134    
1135    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1136        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1137        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1138    
1139    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1140        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1141        the "old" RFC 2279). Instead, it ended up passing a negative length to
1142        pcre_exec().
1143    
1144    31. Add support for GCC's visibility feature to hide internal functions.
1145    
1146    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1147        "unknown -C option" after outputting 0 or 1.
1148    
1149    33. There is now support for generating a code coverage report for the test
1150        suite in environments where gcc is the compiler and lcov is installed. This
1151        is mainly for the benefit of the developers.
1152    
1153    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1154        unaddressable using valgrind annotations, allowing valgrind to detect
1155        invalid memory accesses. This is mainly for the benefit of the developers.
1156    
1157    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1158    
1159    26. Give configure error if --enable-cpp but no C++ compiler found.
1160    
1161    
1162    Version 8.31 06-July-2012
1163    -------------------------
1164    
1165    1.  Fixing a wrong JIT test case and some compiler warnings.
1166    
1167    2.  Removed a bashism from the RunTest script.
1168    
1169    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1170        to unsigned type, result still unsigned" that was given by an MS compiler
1171        on encountering the code "-sizeof(xxx)".
1172    
1173    4.  Partial matching support is added to the JIT compiler.
1174    
1175    5.  Fixed several bugs concerned with partial matching of items that consist
1176        of more than one character:
1177    
1178        (a) /^(..)\1/ did not partially match "aba" because checking references was
1179            done on an "all or nothing" basis. This also applied to repeated
1180            references.
1181    
1182        (b) \R did not give a hard partial match if \r was found at the end of the
1183            subject.
1184    
1185        (c) \X did not give a hard partial match after matching one or more
1186            characters at the end of the subject.
1187    
1188        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1189            a partial match for the string "\r".
1190    
1191        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1192            a partial match for a CR character at the end of the subject string.
1193    
1194    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1195        running pcretest, the text "(JIT)" added to the output whenever JIT is
1196        actually used to run the match.
1197    
1198    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1199        or /S+[+] with a digit between 1 and 7.
1200    
1201    8.  OP_NOT now supports any UTF character not just single-byte ones.
1202    
1203    9.  (*MARK) control verb is now supported by the JIT compiler.
1204    
1205    10. The command "./RunTest list" lists the available tests without actually
1206        running any of them. (Because I keep forgetting what they all are.)
1207    
1208    11. Add PCRE_INFO_MAXLOOKBEHIND.
1209    
1210    12. Applied a (slightly modified) user-supplied patch that improves performance
1211        when the heap is used for recursion (compiled with --disable-stack-for-
1212        recursion). Instead of malloc and free for each heap frame each time a
1213        logical recursion happens, frames are retained on a chain and re-used where
1214        possible. This sometimes gives as much as 30% improvement.
1215    
1216    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1217        call.
1218    
1219    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1220    
1221    15. It is now possible to link pcretest with libedit as an alternative to
1222        libreadline.
1223    
1224    16. (*COMMIT) control verb is now supported by the JIT compiler.
1225    
1226    17. The Unicode data tables have been updated to Unicode 6.1.0.
1227    
1228    18. Added --file-list option to pcregrep.
1229    
1230    19. Added binary file support to pcregrep, including the -a, --binary-files,
1231        -I, and --text options.
1232    
1233    20. The madvise function is renamed for posix_madvise for QNX compatibility
1234        reasons. Fixed by Giuseppe D'Angelo.
1235    
1236    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1237    
1238    22. Changed the option for creating symbolic links for 16-bit man pages from
1239        -s to -sf so that re-installing does not cause issues.
1240    
1241    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1242    
1243    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1244        matches in certain environments (the workspace was not being correctly
1245        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1246        some of the workspace data at the beginning of a restart.
1247    
1248    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1249        was not doing so when it should - probably a typo introduced by SVN 528
1250        (change 8.10/14).
1251    
1252    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1253        \w+ when the character tables indicated that \x{c4} was a word character.
1254        There were several related cases, all because the tests for doing a table
1255        lookup were testing for characters less than 127 instead of 255.
1256    
1257    27. If a pattern contains capturing parentheses that are not used in a match,
1258        their slots in the ovector are set to -1. For those that are higher than
1259        any matched groups, this happens at the end of processing. In the case when
1260        there were back references that the ovector was too small to contain
1261        (causing temporary malloc'd memory to be used during matching), and the
1262        highest capturing number was not used, memory off the end of the ovector
1263        was incorrectly being set to -1. (It was using the size of the temporary
1264        memory instead of the true size.)
1265    
1266    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1267        ovector size, it uses memory at the end of the block that it has got.
1268    
1269    29. Check for an overlong MARK name and give an error at compile time. The
1270        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1271    
1272    30. JIT compiler update.
1273    
1274    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1275        Rill for the patch.
1276    
1277    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1278    
1279    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1280    
1281    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1282        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1283        was enabled.
1284    
1285    35. Improve JIT code generation for greedy plus quantifier.
1286    
1287    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1288        1 to "aa" instead of to an empty string. The bug affected repeated groups
1289        that could potentially match an empty string.
1290    
1291    37. Optimizing single character iterators in JIT.
1292    
1293    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1294        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1295        codepoints that are too big for the mode are faulted, and in a UTF mode,
1296        disallowed codepoints are also faulted.
1297    
1298    39. If PCRE was compiled with UTF support, in three places in the DFA
1299        matcher there was code that should only have been obeyed in UTF mode, but
1300        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1301        processing when bytes with values greater than 127 were present. In 16-bit
1302        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1303        both cases the values are those that cannot be the first data item in a UTF
1304        character. The three items that might have provoked this were recursions,
1305        possessively repeated groups, and atomic groups.
1306    
1307    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1308        and pcregrep, because some OS require shared objects to be explicitly
1309        passed to ld, causing the link step to fail if they are not.
1310    
1311    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1312        mode, patterns that started with \h* or \R* might be incorrectly matched.
1313    
1314    
1315    Version 8.30 04-February-2012
1316    -----------------------------
1317    
1318    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1319        name is defined in ctype.h.
1320    
1321    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1322        only in quite long subpatterns.
1323    
1324    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1325        since it was replaced by pcre_fullinfo() in February 2000.
1326    
1327    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1328        match a (*MARK), and the match failed at the start of the subject, a
1329        reference to memory before the start of the subject could occur. This bug
1330        was introduced by fix 17 of release 8.21.
1331    
1332    5.  A reference to an unset group with zero minimum repetition was giving
1333        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1334        /(another)?(\1?)test/ matched against "hello world test". This bug was
1335        introduced in release 8.13.
1336    
1337    6.  Add support for 16-bit character strings (a large amount of work involving
1338        many changes and refactorings).
1339    
1340    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1341        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1342        from a file.
1343    
1344    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1345        rounding is not applied in this particular case).
1346    
1347    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1348        if they appear, or are escaped, in patterns.
1349    
1350    10. Get rid of a number of -Wunused-but-set-variable warnings.
1351    
1352    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1353        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1354        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1355        also returns the mark "x". This bug applied to capturing parentheses,
1356        non-capturing parentheses, and atomic parentheses. It also applied to some
1357        assertions.
1358    
1359    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1360        information out of configure.ac instead of relying on pcre.h.generic, which
1361        is not stored in the repository.
1362    
1363    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1364        -lreadline.
1365    
1366    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1367    
1368    15. Applied Graycode's patch to put the top-level frame on the stack rather
1369        than the heap when not using the stack for recursion. This gives a
1370        performance improvement in many cases when recursion is not deep.
1371    
1372    16. Experimental code added to "pcretest -C" to output the stack frame size.
1373    
1374    
1375    Version 8.21 12-Dec-2011
1376    ------------------------
1377    
1378    1.  Updating the JIT compiler.
1379    
1380    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1381        are added as well.
1382    
1383    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1384        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1385        calling _pcre_jit_exec. Some extra comments are added.
1386    
1387    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1388        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1389        was introduced by change 18 for 8.20.
1390    
1391    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1392        ECMA-262 standard.
1393    
1394    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1395        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1396        This bug was probably introduced by change 9 of 8.13.
1397    
1398    7.  While fixing 6 above, I noticed that a number of other items were being
1399        incorrectly rejected as "not fixed length". This arose partly because newer
1400        opcodes had not been added to the fixed-length checking code. I have (a)
1401        corrected the bug and added tests for these items, and (b) arranged for an
1402        error to occur if an unknown opcode is encountered while checking for fixed
1403        length instead of just assuming "not fixed length". The items that were
1404        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1405        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1406        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1407    
1408    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1409        being incorrectly compiled and would have given unpredicatble results.
1410    
1411    9.  A possessively repeated subpattern with minimum repeat count greater than
1412        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1413        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1414        the first (A) could occur when it should not.
1415    
1416    10. Add a cast and remove a redundant test from the code.
1417    
1418    11. JIT should use pcre_malloc/pcre_free for allocation.
1419    
1420    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1421        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1422        is anything other than /usr, -L is still shown).
1423    
1424    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1425    
1426    14. Perl does not support \N without a following name in a [] class; PCRE now
1427        also gives an error.
1428    
1429    15. If a forward reference was repeated with an upper limit of around 2000,
1430        it caused the error "internal error: overran compiling workspace". The
1431        maximum number of forward references (including repeats) was limited by the
1432        internal workspace, and dependent on the LINK_SIZE. The code has been
1433        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1434        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1435        of around 200,000 forward references. While doing this, I also speeded up
1436        the filling in of repeated forward references.
1437    
1438    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1439        incorrectly expecting the subject to contain another "a" after the start.
1440    
1441    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1442        in the match, the SKIP should be ignored. This was not happening; instead
1443        the SKIP was being treated as NOMATCH. For patterns such as
1444        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1445        tested.
1446    
1447    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1448        now much more compatible with Perl, in particular in cases where the result
1449        is a non-match for a non-anchored pattern. For example, if
1450        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1451        "m", where previously it did not return a name. A side effect of this
1452        change is that for partial matches, the last encountered mark name is
1453        returned, as for non matches. A number of tests that were previously not
1454        Perl-compatible have been moved into the Perl-compatible test files. The
1455        refactoring has had the pleasing side effect of removing one argument from
1456        the match() function, thus reducing its stack requirements.
1457    
1458    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1459        subsequent uses of /S (without +) incorrectly behaved like /S+.
1460    
1461    21. Retrieve executable code size support for the JIT compiler and fixing
1462        some warnings.
1463    
1464    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1465        not work when the shorter character appeared right at the end of the
1466        subject string.
1467    
1468    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1469        systems.
1470    
1471    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1472        output it when the /M option is used in pcretest.
1473    
1474    25. The CheckMan script was not being included in the distribution. Also, added
1475        an explicit "perl" to run Perl scripts from the PrepareRelease script
1476        because this is reportedly needed in Windows.
1477    
1478    26. If study data was being save in a file and studying had not found a set of
1479        "starts with" bytes for the pattern, the data written to the file (though
1480        never used) was taken from uninitialized memory and so caused valgrind to
1481        complain.
1482    
1483    27. Updated RunTest.bat as provided by Sheri Pierce.
1484    
1485    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1486    
1487    29. Computation of memory usage for the table of capturing group names was
1488        giving an unnecessarily large value.
1489    
1490    
1491    Version 8.20 21-Oct-2011
1492    ------------------------
1493    
1494    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1495        a POSIX class. After further experiments with Perl, which convinced me that
1496        Perl has bugs and confusions, a closing square bracket is no longer allowed
1497        in a POSIX name. This bug also affected patterns with classes that started
1498        with full stops.
1499    
1500    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1501        captured substring, but while checking the failing first alternative,
1502        substring 1 is temporarily captured. If the output vector supplied to
1503        pcre_exec() was not big enough for this capture, the yield of the function
1504        was still zero ("insufficient space for captured substrings"). This cannot
1505        be totally fixed without adding another stack variable, which seems a lot
1506        of expense for a edge case. However, I have improved the situation in cases
1507        such as /(a)(b)x|abc/ matched against "abc", where the return code
1508        indicates that fewer than the maximum number of slots in the ovector have
1509        been set.
1510    
1511    3.  Related to (2) above: when there are more back references in a pattern than
1512        slots in the output vector, pcre_exec() uses temporary memory during
1513        matching, and copies in the captures as far as possible afterwards. It was
1514        using the entire output vector, but this conflicts with the specification
1515        that only 2/3 is used for passing back captured substrings. Now it uses
1516        only the first 2/3, for compatibility. This is, of course, another edge
1517        case.
1518    
1519    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1520        main code base, and can be used by building with --enable-jit. When this is
1521        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1522        runtime --no-jit option is given.
1523    
1524    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1525        ovector, the return from the function was zero, implying that there were
1526        other matches that did not fit. The correct "exactly full" value is now
1527        returned.
1528    
1529    6.  If a subpattern that was called recursively or as a subroutine contained
1530        (*PRUNE) or any other control that caused it to give a non-standard return,
1531        invalid errors such as "Error -26 (nested recursion at the same subject
1532        position)" or even infinite loops could occur.
1533    
1534    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1535        computing the minimum length on reaching *ACCEPT, and so ended up with the
1536        wrong value of 1 rather than 0. Further investigation indicates that
1537        computing a minimum subject length in the presence of *ACCEPT is difficult
1538        (think back references, subroutine calls), and so I have changed the code
1539        so that no minimum is registered for a pattern that contains *ACCEPT.
1540    
1541    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1542        it was not handled as intended. [But see 16 below.]
1543    
1544    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1545        Sheri Pierce.
1546    
1547    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1548        the first byte in a match must be "a".
1549    
1550    11. Change 17 for 8.13 increased the recursion depth for patterns like
1551        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1552        pattern contains any instances of (*THEN). If it does not, the old
1553        optimizations are restored. It would be nice to do this on a per-group
1554        basis, but at the moment that is not feasible.
1555    
1556    12. In some environments, the output of pcretest -C is CRLF terminated. This
1557        broke RunTest's code that checks for the link size. A single white space
1558        character after the value is now allowed for.
1559    
1560    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1561        For "fr", it uses the Windows-specific input and output files.
1562    
1563    14. If (*THEN) appeared in a group that was called recursively or as a
1564        subroutine, it did not work as intended. [But see next item.]
1565    
1566    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1567        pattern fragments (but not containing any | characters). If A and B are
1568        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1569        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1570        D. In other words, Perl considers parentheses that do not contain any |
1571        characters to be part of a surrounding alternative, whereas PCRE was
1572        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1573        differently. PCRE now behaves in the same way as Perl, except in the case
1574        of subroutine/recursion calls such as (?1) which have in any case always
1575        been different (but PCRE had them first :-).
1576    
1577    16. Related to 15 above: Perl does not treat the | in a conditional group as
1578        creating alternatives. Such a group is treated in the same way as an
1579        ordinary group without any | characters when processing (*THEN). PCRE has
1580        been changed to match Perl's behaviour.
1581    
1582    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1583        RunGrepTest script failed.
1584    
1585    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1586        inevitable for groups that contain captures, but it can lead to a lot of
1587        stack use in large patterns. The old behaviour has been restored for atomic
1588        groups that do not contain any capturing parentheses.
1589    
1590    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1591        suppress the check for a minimum subject length at run time. (If it was
1592        given to pcre_exec() or pcre_dfa_exec() it did work.)
1593    
1594    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1595        fail to work when decoding hex characters in data strings in EBCDIC
1596        environments.
1597    
1598    21. It appears that in at least one Mac OS environment, the isxdigit() function
1599        is implemented as a macro that evaluates to its argument more than once,
1600        contravening the C 90 Standard (I haven't checked a later standard). There
1601        was an instance in pcretest which caused it to go wrong when processing
1602        \x{...} escapes in subject strings. The has been rewritten to avoid using
1603        things like p++ in the argument of isxdigit().
1604    
1605    
1606    Version 8.13 16-Aug-2011
1607    ------------------------
1608    
1609    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1610    
1611    2.  Two minor typos in pcre_internal.h have been fixed.
1612    
1613    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1614        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1615        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1616    
1617    4.  There were a number of related bugs in the code for matching backrefences
1618        caselessly in UTF-8 mode when codes for the characters concerned were
1619        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1620        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1621        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1622        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1623        2-byte code at the end of the subject (it thought there wasn't enough data
1624        left).
1625    
1626    5.  Comprehensive information about what went wrong is now returned by
1627        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1628        as the output vector has at least 2 elements. The offset of the start of
1629        the failing character and a reason code are placed in the vector.
1630    
1631    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1632        now returned is for the first byte of the failing character, instead of the
1633        last byte inspected. This is an incompatible change, but I hope it is small
1634        enough not to be a problem. It makes the returned offset consistent with
1635        pcre_exec() and pcre_dfa_exec().
1636    
1637    7.  pcretest now gives a text phrase as well as the error number when
1638        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1639        failure, the offset and reason code are output.
1640    
1641    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1642        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1643        back over a single character (\n). This seems wrong (because it treated the
1644        two characters as a single entity when going forwards), conflicts with the
1645        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1646        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1647        has been changed.
1648    
1649    9.  Some internal refactoring has changed the processing so that the handling
1650        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1651        time (the PCRE_DOTALL option was changed this way some time ago: version
1652        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1653        which was always a bit of a fudge. It also means that there is one less
1654        argument for the match() function, which reduces its stack requirements
1655        slightly. This change also fixes an incompatibility with Perl: the pattern
1656        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1657    
1658    10. More internal refactoring has drastically reduced the number of recursive
1659        calls to match() for possessively repeated groups such as (abc)++ when
1660        using pcre_exec().
1661    
1662    11. While implementing 10, a number of bugs in the handling of groups were
1663        discovered and fixed:
1664    
1665        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1666        (a|)*(?1) gave a compile-time internal error.
1667        ((a|)+)+  did not notice that the outer group could match an empty string.
1668        (^a|^)+   was not marked as anchored.
1669        (.*a|.*)+ was not marked as matching at start or after a newline.
1670    
1671    12. Yet more internal refactoring has removed another argument from the match()
1672        function. Special calls to this function are now indicated by setting a
1673        value in a variable in the "match data" data block.
1674    
1675    13. Be more explicit in pcre_study() instead of relying on "default" for
1676        opcodes that mean there is no starting character; this means that when new
1677        ones are added and accidentally left out of pcre_study(), testing should
1678        pick them up.
1679    
1680    14. The -s option of pcretest has been documented for ages as being an old
1681        synonym of -m (show memory usage). I have changed it to mean "force study
1682        for every regex", that is, assume /S for every regex. This is similar to -i
1683        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1684        using it. It makes it easier to run collections of tests with and without
1685        study enabled, and thereby test pcre_study() more easily. All the standard
1686        tests are now run with and without -s (but some patterns can be marked as
1687        "never study" - see 20 below).
1688    
1689    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1690        restoration of the capturing data to the outer values was not happening
1691        correctly.
1692    
1693    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1694        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1695        pattern had matched an empty string, and so incorrectly returned a no
1696        match.
1697    
1698    17. There was optimizing code for the last branch of non-capturing parentheses,
1699        and also for the obeyed branch of a conditional subexpression, which used
1700        tail recursion to cut down on stack usage. Unfortunately, now that there is
1701        the possibility of (*THEN) occurring in these branches, tail recursion is
1702        no longer possible because the return has to be checked for (*THEN). These
1703        two optimizations have therefore been removed. [But see 8.20/11 above.]
1704    
1705    18. If a pattern containing \R was studied, it was assumed that \R always
1706        matched two bytes, thus causing the minimum subject length to be
1707        incorrectly computed because \R can also match just one byte.
1708    
1709    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1710        was incorrectly computed.
1711    
1712    20. If /S is present twice on a test pattern in pcretest input, it now
1713        *disables* studying, thereby overriding the use of -s on the command line
1714        (see 14 above). This is necessary for one or two tests to keep the output
1715        identical in both cases.
1716    
1717    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1718        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1719    
1720    22. When an atomic group that contained a capturing parenthesis was
1721        successfully matched, but the branch in which it appeared failed, the
1722        capturing was not being forgotten if a higher numbered group was later
1723        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1724        group 1 to "a", when in fact it should be unset. This applied to multi-
1725        branched capturing and non-capturing groups, repeated or not, and also to
1726        positive assertions (capturing in negative assertions does not happen
1727        in PCRE) and also to nested atomic groups.
1728    
1729    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1730        subject after a captured substring, to make it easier to tell which of a
1731        number of identical substrings has been captured.
1732    
1733    24. The way atomic groups are processed by pcre_exec() has been changed so that
1734        if they are repeated, backtracking one repetition now resets captured
1735        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1736        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1737        "aaa". Previously, it would have been "a". As part of this code
1738        refactoring, the way recursive calls are handled has also been changed.
1739    
1740    25. If an assertion condition captured any substrings, they were not passed
1741        back unless some other capturing happened later. For example, if
1742        (?(?=(a))a) was matched against "a", no capturing was returned.
1743    
1744    26. When studying a pattern that contained subroutine calls or assertions,
1745        the code for finding the minimum length of a possible match was handling
1746        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1747        group 1 called group 2 while simultaneously a separate group 2 called group
1748        1). A stack overflow occurred in this case. I have fixed this by limiting
1749        the recursion depth to 10.
1750    
1751    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1752        Fortmann. This supports explicit test numbers on the command line, and has
1753        argument validation and error reporting.
1754    
1755    28. An instance of \X with an unlimited repeat could fail if at any point the
1756        first character it looked at was a mark character.
1757    
1758    29. Some minor code refactoring concerning Unicode properties and scripts
1759        should reduce the stack requirement of match() slightly.
1760    
1761    30. Added the '=' option to pcretest to check the setting of unused capturing
1762        slots at the end of the pattern, which are documented as being -1, but are
1763        not included in the return count.
1764    
1765    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1766        compiled something random. Now it gives a compile-time error (as does
1767        Perl).
1768    
1769    32. A *MARK encountered during the processing of a positive assertion is now
1770        recorded and passed back (compatible with Perl).
1771    
1772    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1773        had alternative anchored branches, the search for a second match in a line
1774        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1775        matched the line "0102" twice. The same bug affected patterns that started
1776        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1777        twice.
1778    
1779    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1780        does, and because of capturing effects, quantifying parenthesized
1781        assertions may at times be useful. Quantifiers are now allowed for
1782        parenthesized assertions.
1783    
1784    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1785    
1786    36. \g was being checked for fancy things in a character class, when it should
1787        just be a literal "g".
1788    
1789    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1790        appearance of a nested POSIX class supersedes an apparent external class.
1791        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1792        unescaped square brackets may also appear as part of class names. For
1793        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1794        more like Perl. (But see 8.20/1 above.)
1795    
1796    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1797        was because it thought it was \N{name}, which is not supported).
1798    
1799    39. Add minix to OS list not supporting the -S option in pcretest.
1800    
1801    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1802        cannot analyze patterns in sufficient detail to catch mutual recursions
1803        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1804        subgroup is called recursively as a subpattern for a second time at the
1805        same position in the subject string. In previous releases this might have
1806        been caught by the recursion limit, or it might have run out of stack.
1807    
1808    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1809        happen only once. PCRE was, however incorrectly giving a compile time error
1810        "recursive call could loop indefinitely" because it cannot analyze the
1811        pattern in sufficient detail. The compile time test no longer happens when
1812        PCRE is compiling a conditional subpattern, but actual runaway loops are
1813        now caught at runtime (see 40 above).
1814    
1815    42. It seems that Perl allows any characters other than a closing parenthesis
1816        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1817        has been changed to be the same.
1818    
1819    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1820        as not to get warnings when autogen.sh is called. Also changed
1821        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1822    
1823    44. To help people who use pcregrep to scan files containing exceedingly long
1824        lines, the following changes have been made:
1825    
1826        (a) The default value of the buffer size parameter has been increased from
1827            8K to 20K. (The actual buffer used is three times this size.)
1828    
1829        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1830            PCRE is built.
1831    
1832        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1833            to be set at run time.
1834    
1835        (d) Numerical values in pcregrep options can be followed by K or M, for
1836            example --buffer-size=50K.
1837    
1838        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1839            given and the return code is set to 2.
1840    
1841    45. Add a pointer to the latest mark to the callout data block.
1842    
1843    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1844        partial match of an empty string instead of no match. This was specific to
1845        the use of ".".
1846    
1847    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1848        complete match instead of a partial match. This bug was dependent on both
1849        the PCRE_UTF8 and PCRE_DOTALL options being set.
1850    
1851    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1852        starting byte set, because \b was not being ignored.
1853    
1854    
1855    Version 8.12 15-Jan-2011
1856    ------------------------
1857    
1858    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1859        checks for such things as part of the documentation building process.
1860    
1861    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1862        --match-limit and --recursion-limit options (added for 8.11). In
1863        particular, this made one of the standard tests fail. (The integer value
1864        went into the wrong half of a long int.)
1865    
1866    3.  If the --colour option was given to pcregrep with -v (invert match), it
1867        did strange things, either producing crazy output, or crashing. It should,
1868        of course, ignore a request for colour when reporting lines that do not
1869        match.
1870    
1871    4.  Another pcregrep bug caused similar problems if --colour was specified with
1872        -M (multiline) and the pattern match finished with a line ending.
1873    
1874    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1875        matched in multiline mode, the following line was shown as part of the
1876        match. This seems wrong, so I have changed it.
1877    
1878    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1879        the check for further matches in the same line (so they could be coloured)
1880        to overrun the end of the current line. If another match was found, it was
1881        incorrectly shown (and then shown again when found in the next line).
1882    
1883    7.  If pcregrep was compiled under Windows, there was a reference to the
1884        function pcregrep_exit() before it was defined. I am assuming this was
1885        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1886        reported by a user. I've moved the definition above the reference.
1887    
1888    
1889    Version 8.11 10-Dec-2010
1890    ------------------------
1891    
1892    1.  (*THEN) was not working properly if there were untried alternatives prior
1893        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1894        backtracked to try for "b" instead of moving to the next alternative branch
1895        at the same level (in this case, to look for "c"). The Perl documentation
1896        is clear that when (*THEN) is backtracked onto, it goes to the "next
1897        alternative in the innermost enclosing group".
1898    
1899    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1900        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1901        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1902        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1903        (*THEN).
1904    
1905    3.  If \s appeared in a character class, it removed the VT character from
1906        the class, even if it had been included by some previous item, for example
1907        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1908        of \s, but is part of the POSIX "space" class.)
1909    
1910    4.  A partial match never returns an empty string (because you can always
1911        match an empty string at the end of the subject); however the checking for
1912        an empty string was starting at the "start of match" point. This has been
1913        changed to the "earliest inspected character" point, because the returned
1914        data for a partial match starts at this character. This means that, for
1915        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1916        (previously it gave "no match").
1917    
1918    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1919        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1920        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1921        has an implication that the given string is incomplete (because a partial
1922        match is preferred over a full match). For this reason, these items now
1923        give a partial match in this situation. [Aside: previously, the one case
1924        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1925        match rather than a full match, which was wrong by the old rules, but is
1926        now correct.]
1927    
1928    6.  There was a bug in the handling of #-introduced comments, recognized when
1929        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1930        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1931        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1932        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1933        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1934        places in pcre_compile().
1935    
1936    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1937        comments when looking ahead for named forward references to subpatterns,
1938        the only newline sequence it recognized was NL. It now handles newlines
1939        according to the set newline convention.
1940    
1941    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1942        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1943        cater for a lack of strerror(). These oversights have been fixed.
1944    
1945    9.  Added --match-limit and --recursion-limit to pcregrep.
1946    
1947    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1948    
1949    11. When the -o option was used, pcregrep was setting a return code of 1, even
1950        when matches were found, and --line-buffered was not being honoured.
1951    
1952    12. Added an optional parentheses number to the -o and --only-matching options
1953        of pcregrep.
1954    
1955    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1956        can match an empty string. The code to do it in pcretest and pcredemo
1957        needed fixing:
1958    
1959        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1960            only one byte after an empty string match just before CRLF (this case
1961            just got forgotten; "any" and "anycrlf" were OK).
1962    
1963        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1964            mode when an empty string match preceded an ASCII character followed by
1965            a non-ASCII character. (The code for advancing by one character rather
1966            than one byte was nonsense.)
1967    
1968        (c) The pcredemo.c sample program did not have any code at all to handle
1969            the cases when CRLF is a valid newline sequence.
1970    
1971    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1972        as a starting offset was within the subject string. There is now a new
1973        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1974        negative or greater than the length of the string. In order to test this,
1975        pcretest is extended to allow the setting of negative starting offsets.
1976    
1977    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1978        starting offset points to the beginning of a UTF-8 character was
1979        unnecessarily clumsy. I tidied it up.
1980    
1981    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1982        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1983    
1984    17. Nobody had reported that the --include_dir option, which was added in
1985        release 7.7 should have been called --include-dir (hyphen, not underscore)
1986        for compatibility with GNU grep. I have changed it to --include-dir, but
1987        left --include_dir as an undocumented synonym, and the same for
1988        --exclude-dir, though that is not available in GNU grep, at least as of
1989        release 2.5.4.
1990    
1991    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1992        characters from a string of bytes) have been redefined so as not to use
1993        loops, in order to improve performance in some environments. At the same
1994        time, I abstracted some of the common code into auxiliary macros to save
1995        repetition (this should not affect the compiled code).
1996    
1997    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1998        compile-time error is now given if \c is not followed by an ASCII
1999        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2000        different, and any byte value is allowed.)
2001    
2002    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2003        START_OPTIMIZE option, which is now allowed at compile time - but just
2004        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2005        to pcregrep and other applications that have no direct access to PCRE
2006        options. The new /Y option in pcretest sets this option when calling
2007        pcre_compile().
2008    
2009    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2010        back references. Groups containing recursive back references were forced to
2011        be atomic by that change, but in the case of named groups, the amount of
2012        memory required was incorrectly computed, leading to "Failed: internal
2013        error: code overflow". This has been fixed.
2014    
2015    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2016        pcretest.c, to avoid build problems in some Borland environments.
2017    
2018    
2019    Version 8.10 25-Jun-2010
2020    ------------------------
2021    
2022    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2023        THEN.
2024    
2025    2.  (*ACCEPT) was not working when inside an atomic group.
2026    
2027    3.  Inside a character class, \B is treated as a literal by default, but
2028        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2029        causes the error). The code is unchanged, but I tidied the documentation.
2030    
2031    4.  Inside a character class, PCRE always treated \R and \X as literals,
2032        whereas Perl faults them if its -w option is set. I have changed PCRE so
2033        that it faults them when PCRE_EXTRA is set.
2034    
2035    5.  Added support for \N, which always matches any character other than
2036        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2037    
2038    6.  When compiling pcregrep with newer versions of gcc which may have
2039        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2040        declared with attribute warn_unused_result" were given. Just casting the
2041        result to (void) does not stop the warnings; a more elaborate fudge is
2042        needed. I've used a macro to implement this.
2043    
2044    7.  Minor change to pcretest.c to avoid a compiler warning.
2045    
2046    8.  Added four artifical Unicode properties to help with an option to make
2047        \s etc use properties (see next item). The new properties are: Xan
2048        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2049    
2050    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2051        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2052        this option. Modified pcretest to add /W to test this facility. Added
2053        REG_UCP to make it available via the POSIX interface.
2054    
2055    10. Added --line-buffered to pcregrep.
2056    
2057    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2058        studied, and the match started with a letter with a code point greater than
2059        127 whose first byte was different to the first byte of the other case of
2060        the letter, the other case of this starting letter was not recognized
2061        (#976).
2062    
2063    12. If a pattern that was studied started with a repeated Unicode property
2064        test, for example, \p{Nd}+, there was the theoretical possibility of
2065        setting up an incorrect bitmap of starting bytes, but fortunately it could
2066        not have actually happened in practice until change 8 above was made (it
2067        added property types that matched character-matching opcodes).
2068    
2069    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2070        possible starting bytes for non-anchored patterns.
2071    
2072    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2073        \R, and also a number of cases that involve Unicode properties, both
2074        explicit and implicit when PCRE_UCP is set.
2075    
2076    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2077        input, it could crash or give wrong results if characters with values
2078        greater than 0xc0 were present in the subject string. (Detail: it assumed
2079        UTF-8 input when processing these items.)
2080    
2081    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2082        size_t is 64-bit (#991).
2083    
2084    17. Added a check for running out of memory when PCRE is compiled with
2085        --disable-stack-for-recursion (#990).
2086    
2087    18. If the last data line in a file for pcretest does not have a newline on
2088        the end, a newline was missing in the output.
2089    
2090    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2091        less than 128) in its various bitmaps. However, there is a facility for
2092        generating tables according to the current locale when PCRE is compiled. It
2093        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2094        characters, are recognized by isspace() and therefore were getting set in
2095        these tables, and indeed these tables seem to approximate to ISO 8859. This
2096        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2097        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2098        which of course cannot start UTF-8 characters. I have changed the code so
2099        that only real ASCII characters (less than 128) and the correct starting
2100        bytes for UTF-8 encodings are set for characters greater than 127 when in
2101        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2102        altogether.)
2103    
2104    20. Added the /T option to pcretest so as to be able to run tests with non-
2105        standard character tables, thus making it possible to include the tests
2106        used for 19 above in the standard set of tests.
2107    
2108    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2109        reference to a subpattern the other side of a comment that contains an
2110        opening parenthesis caused either an internal compiling error, or a
2111        reference to the wrong subpattern.
2112    
2113    
2114    Version 8.02 19-Mar-2010
2115    ------------------------
2116    
2117    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2118    
2119    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2120        configured.
2121    
2122    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2123        original author of that file, following a query about its status.
2124    
2125    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2126        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2127    
2128    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2129        quantifier applied to a forward-referencing subroutine call, could compile
2130        incorrect code or give the error "internal error: previously-checked
2131        referenced subpattern not found".
2132    
2133    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2134        variables to point to external functions. For these systems, therefore,
2135        pcre_malloc etc. are now initialized to local functions that call the
2136        relevant global functions.
2137    
2138    7.  There were two entries missing in the vectors called coptable and poptable
2139        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2140        I've fixed the data, and added a kludgy way of testing at compile time that
2141        the lengths are correct (equal to the number of opcodes).
2142    
2143    8.  Following on from 7, I added a similar kludge to check the length of the
2144        eint vector in pcreposix.c.
2145    
2146    9.  Error texts for pcre_compile() are held as one long string to avoid too
2147        much relocation at load time. To find a text, the string is searched,
2148        counting zeros. There was no check for running off the end of the string,
2149        which could happen if a new error number was added without updating the
2150        string.
2151    
2152    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2153    
2154    11. \K was not working if it appeared in an atomic group or in a group that
2155        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2156        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2157        the assertion is positive, but not if it is negative.
2158    
2159    12. Change 11 fortuitously reduced the size of the stack frame used in the
2160        "match()" function of pcre_exec.c by one pointer. Forthcoming
2161        implementation of support for (*MARK) will need an extra pointer on the
2162        stack; I have reserved it now, so that the stack frame size does not
2163        decrease.
2164    
2165    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2166        item in branch that calls a recursion is a subroutine call - as in the
2167        second branch in the above example - was incorrectly given the compile-
2168        time error "recursive call could loop indefinitely" because pcre_compile()
2169        was not correctly checking the subroutine for matching a non-empty string.
2170    
2171    14. The checks for overrunning compiling workspace could trigger after an
2172        overrun had occurred. This is a "should never occur" error, but it can be
2173        triggered by pathological patterns such as hundreds of nested parentheses.
2174        The checks now trigger 100 bytes before the end of the workspace.
2175    
2176    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2177    
2178    
2179    Version 8.01 19-Jan-2010
2180    ------------------------
2181    
2182    1.  If a pattern contained a conditional subpattern with only one branch (in
2183        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2184        computed the wrong minimum data length (which is of course zero for such
2185        subpatterns). This could cause incorrect "no match" results.
2186    
2187    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2188        the pattern is reset in the first branch, pcre_compile() failed with
2189        "internal error: code overflow at offset...". This happened only when
2190        the reset was to the original external option setting. (An optimization
2191        abstracts leading options settings into an external setting, which was the
2192        cause of this.)
2193    
2194    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2195        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2196        assertion pattern did not match (meaning that the assertion was true), it
2197        was incorrectly treated as false if the SKIP had been reached during the
2198        matching. This also applied to assertions used as conditions.
2199    
2200    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2201        assertion subpattern, including such a pattern used as a condition,
2202        unpredictable results occurred, instead of the error return
2203        PCRE_ERROR_DFA_UITEM.
2204    
2205    5.  The C++ GlobalReplace function was not working like Perl for the special
2206        situation when an empty string is matched. It now does the fancy magic
2207        stuff that is necessary.
2208    
2209    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2210        removed. (These were left over from very, very early versions of PCRE.)
2211    
2212    7.  Some cosmetic changes to the code to make life easier when compiling it
2213        as part of something else:
2214    
2215        (a) Change DEBUG to PCRE_DEBUG.
2216    
2217        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2218            called "current" as "current_branch", to prevent a collision with the
2219            Linux macro when compiled as a kernel module.
2220    
2221        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2222            prevent a collision with the Linux macro when compiled as a kernel
2223            module.
2224    
2225    8.  In pcre_compile() there are some checks for integer overflows that used to
2226        cast potentially large values to (double). This has been changed to that
2227        when building, a check for int64_t is made, and if it is found, it is used
2228        instead, thus avoiding the use of floating point arithmetic. (There is no
2229        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2230        double.
2231    
2232    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2233        2005 (difference between two addresses compared to an unsigned value).
2234    
2235    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2236        custom one, because of the following reported problem in Windows:
2237    
2238          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2239              under Win32.
2240          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2241              therefore missing the function definition.
2242          - The compiler thus generates a "C" signature for the test function.
2243          - The linker fails to find the "C" function.
2244          - PCRE fails to configure if asked to do so against libbz2.
2245    
2246    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2247        messages were output:
2248    
2249          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2250          rerunning libtoolize, to keep the correct libtool macros in-tree.
2251          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2252    
2253        I have done both of these things.
2254    
2255    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2256        most of the time, it *can* run out if it is given a pattern that contains a
2257        runaway infinite recursion. I updated the discussion in the pcrestack man
2258        page.
2259    
2260    13. Now that we have gone to the x.xx style of version numbers, the minor
2261        version may start with zero. Using 08 or 09 is a bad idea because users
2262        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2263        interpreted as invalid octal numbers. I've updated the previous comment in
2264        configure.ac, and also added a check that gives an error if 08 or 09 are
2265        used.
2266    
2267    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2268        causing partial matching to fail when the end of the subject matched \W
2269        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2270    
2271    15. There were some discrepancies between the declarations in pcre_internal.h
2272        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2273        their definitions. The declarations used "const uschar *" and the
2274        definitions used USPTR. Even though USPTR is normally defined as "const
2275        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2276        reported that: "This difference in casting confuses some C++ compilers, for
2277        example, SunCC recognizes above declarations as different functions and
2278        generates broken code for hbpcre." I have changed the declarations to use
2279        USPTR.
2280    
2281    16. GNU libtool is named differently on some systems. The autogen.sh script now
2282        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2283        (FreeBSD).
2284    
2285    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2286        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2287        comment: "Figure out how to create a longlong from a string: strtoll and
2288        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2289        instance, but it only takes 2 args instead of 3!"
2290    
2291    18. A subtle bug concerned with back references has been fixed by a change of
2292        specification, with a corresponding code fix. A pattern such as
2293        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2294        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2295        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2296        same bug. Such groups have to be quantified to be useful, or contained
2297        inside another quantified group. (If there's no repetition, the reference
2298        can never match.) The problem arises because, having left the group and
2299        moved on to the rest of the pattern, a later failure that backtracks into
2300        the group uses the captured value from the final iteration of the group
2301        rather than the correct earlier one. I have fixed this in PCRE by forcing
2302        any group that contains a reference to itself to be an atomic group; that
2303        is, there cannot be any backtracking into it once it has completed. This is
2304        similar to recursive and subroutine calls.
2305    
2306    
2307    Version 8.00 19-Oct-09
2308    ----------------------
2309    
2310    1.  The table for translating pcre_compile() error codes into POSIX error codes
2311        was out-of-date, and there was no check on the pcre_compile() error code
2312        being within the table. This could lead to an OK return being given in
2313        error.
2314    
2315    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2316        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2317        in a Windows environment.
2318    
2319    3.  The pcregrep --count option prints the count for each file even when it is
2320        zero, as does GNU grep. However, pcregrep was also printing all files when
2321        --files-with-matches was added. Now, when both options are given, it prints
2322        counts only for those files that have at least one match. (GNU grep just
2323        prints the file name in this circumstance, but including the count seems
2324        more useful - otherwise, why use --count?) Also ensured that the
2325        combination -clh just lists non-zero counts, with no names.
2326    
2327    4.  The long form of the pcregrep -F option was incorrectly implemented as
2328        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2329        but it seems right to fix it, and I didn't think it was worth preserving
2330        the old behaviour.
2331    
2332    5.  The command line items --regex=pattern and --regexp=pattern were not
2333        recognized by pcregrep, which required --regex pattern or --regexp pattern
2334        (with a space rather than an '='). The man page documented the '=' forms,
2335        which are compatible with GNU grep; these now work.
2336    
2337    6.  No libpcreposix.pc file was created for pkg-config; there was just
2338        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2339    
2340    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2341        when UCP support is not needed, by modifying the Python script that
2342        generates it from Unicode data files. This should not matter if the module
2343        is correctly used as a library, but I received one complaint about 50K of
2344        unwanted data. My guess is that the person linked everything into his
2345        program rather than using a library. Anyway, it does no harm.
2346    
2347    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2348        was a minimum greater than 1 for a wide character in a possessive
2349        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2350        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2351        character. Chaos in the form of incorrect output or a compiling loop could
2352        result.
2353    
2354    9.  The restrictions on what a pattern can contain when partial matching is
2355        requested for pcre_exec() have been removed. All patterns can now be
2356        partially matched by this function. In addition, if there are at least two
2357        slots in the offset vector, the offset of the earliest inspected character
2358        for the match and the offset of the end of the subject are set in them when
2359        PCRE_ERROR_PARTIAL is returned.
2360    
2361    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2362        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2363        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2364        and may be more useful for multi-segment matching.
2365    
2366    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2367        used to be given if ever the end of the subject was reached; now it is
2368        given only if matching could not proceed because another character was
2369        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2370        string "Z", which now yields "no match" instead of "partial match". In the
2371        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2372        final character ended with (*FAIL).
2373    
2374    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2375        if the pattern had a "must contain" character that was already found in the
2376        earlier partial match, unless partial matching was again requested. For
2377        example, with the pattern /dog.(body)?/, the "must contain" character is
2378        "g". If the first part-match was for the string "dog", restarting with
2379        "sbody" failed. This bug has been fixed.
2380    
2381    13. The string returned by pcre_dfa_exec() after a partial match has been
2382        changed so that it starts at the first inspected character rather than the
2383        first character of the match. This makes a difference only if the pattern
2384        starts with a lookbehind assertion or \b or \B (\K is not supported by
2385        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2386        matching functions compatible, and I think it's the right thing to do.
2387    
2388    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2389        so that the demonstration program is easily available in environments where
2390        PCRE has not been installed from source.
2391    
2392    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2393        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2394        library.
2395    
2396    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2397        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2398        is not the first non-POSIX option to be added. Clearly some people find
2399        these options useful.
2400    
2401    17. If a caller to the POSIX matching function regexec() passes a non-zero
2402        value for nmatch with a NULL value for pmatch, the value of
2403        nmatch is forced to zero.
2404    
2405    18. RunGrepTest did not have a test for the availability of the -u option of
2406        the diff command, as RunTest does. It now checks in the same way as
2407        RunTest, and also checks for the -b option.
2408    
2409    19. If an odd number of negated classes containing just a single character
2410        interposed, within parentheses, between a forward reference to a named
2411        subpattern and the definition of the subpattern, compilation crashed with
2412        an internal error, complaining that it could not find the referenced
2413        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2414        [The bug was that it was starting one character too far in when skipping
2415        over the character class, thus treating the ] as data rather than
2416        terminating the class. This meant it could skip too much.]
2417    
2418    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2419        /g option in pcretest when the pattern contains \K, which makes it possible
2420        to have an empty string match not at the start, even when the pattern is
2421        anchored. Updated pcretest and pcredemo to use this option.
2422    
2423    21. If the maximum number of capturing subpatterns in a recursion was greater
2424        than the maximum at the outer level, the higher number was returned, but
2425        with unset values at the outer level. The correct (outer level) value is
2426        now given.
2427    
2428    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2429        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2430        make it do so. The string so far is captured, making this feature
2431        compatible with Perl.
2432    
2433    23. The tests have been re-organized, adding tests 11 and 12, to make it
2434        possible to check the Perl 5.10 features against Perl 5.10.
2435    
2436    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2437        pattern matches a fixed length string. PCRE did not allow this; now it
2438        does. Neither allows recursion.
2439    
2440    25. I finally figured out how to implement a request to provide the minimum
2441        length of subject string that was needed in order to match a given pattern.
2442        (It was back references and recursion that I had previously got hung up
2443        on.) This code has now been added to pcre_study(); it finds a lower bound
2444        to the length of subject needed. It is not necessarily the greatest lower
2445        bound, but using it to avoid searching strings that are too short does give
2446        some useful speed-ups. The value is available to calling programs via
2447        pcre_fullinfo().
2448    
2449    26. While implementing 25, I discovered to my embarrassment that pcretest had
2450        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2451        study optimizations had never been tested with that matching function.
2452        Oops. What is worse, even when it was passed study data, there was a bug in
2453        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2454        were also very few tests of studied patterns with pcre_dfa_exec().
2455    
2456    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2457        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2458        on the other side of the coin, they are no longer allowed to have different
2459        names, because these cannot be distinguished in PCRE, and this has caused
2460        confusion. (This is a difference from Perl.)
2461    
2462    28. When duplicate subpattern names are present (necessarily with different
2463        numbers, as required by 27 above), and a test is made by name in a
2464        conditional pattern, either for a subpattern having been matched, or for
2465        recursion in such a pattern, all the associated numbered subpatterns are
2466        tested, and the overall condition is true if the condition is true for any
2467        one of them. This is the way Perl works, and is also more like the way
2468        testing by number works.
2469    
2470    
2471    Version 7.9 11-Apr-09
2472    ---------------------
2473    
2474    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2475        (pcretest), all targets were linked against these libraries. This included
2476        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2477        libraries. This caused unwanted dependencies to be created. This problem
2478        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2479        pcretest is linked with readline.
2480    
2481    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2482        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2483        moved outside it again, because FALSE and TRUE are already defined in AIX,
2484        but BOOL is not.
2485    
2486    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2487        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2488    
2489    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2490        hyphens) following file names and line numbers when outputting matching
2491        lines. This is not true; no spaces are inserted. I have also clarified the
2492        wording for the --colour (or --color) option.
2493    
2494    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2495        was not coloured; this is different to GNU grep, so I have changed it to be
2496        the same.
2497    
2498    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2499        each matching line was coloured. Now it goes on to look for further matches
2500        of any of the test patterns, which is the same behaviour as GNU grep.
2501    
2502    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2503        doesn't make sense to accept an empty string match in pcregrep, so I have
2504        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2505        seems to be how GNU grep behaves. [But see later change 40 for release
2506        8.33.]
2507    
2508    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2509        start or after a newline", because the conditional assertion was not being
2510        correctly handled. The rule now is that both the assertion and what follows
2511        in the first alternative must satisfy the test.
2512    
2513    9.  If auto-callout was enabled in a pattern with a conditional group whose
2514        condition was an assertion, PCRE could crash during matching, both with
2515        pcre_exec() and pcre_dfa_exec().
2516    
2517    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2518        used for matching.
2519    
2520    11. Unicode property support in character classes was not working for
2521        characters (bytes) greater than 127 when not in UTF-8 mode.
2522    
2523    12. Added the -M command line option to pcretest.
2524    
2525    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2526    
2527    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2528    
2529    16. Added comments and documentation about mis-use of no_arg in the C++
2530        wrapper.
2531    
2532    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2533        from Martin Jerabek that uses macro names for all relevant character and
2534        string constants.
2535    
2536    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2537        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2538        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2539        these, but not everybody uses configure.
2540    
2541    19. A conditional group that had only one branch was not being correctly
2542        recognized as an item that could match an empty string. This meant that an
2543        enclosing group might also not be so recognized, causing infinite looping
2544        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2545        with the subject "ab", where knowledge that the repeated group can match
2546        nothing is needed in order to break the loop.
2547    
2548    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2549        exec(), but without supplying a callout function, matching went wrong.
2550    
2551    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2552        leak if the size of the offset vector was greater than 30. When the vector
2553        is smaller, the saved offsets during recursion go onto a local stack
2554        vector, but for larger vectors malloc() is used. It was failing to free
2555        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2556        error, in fact).
2557    
2558    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2559        heapframe that is used only when UTF-8 support is enabled. This caused no
2560        problem, but was untidy.
2561    
2562    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2563        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2564        included within another project.
2565    
2566    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2567        slightly modified by me:
2568    
2569          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2570              not building pcregrep.
2571    
2572          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2573              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2574    
2575    25. Forward references, both numeric and by name, in patterns that made use of
2576        duplicate group numbers, could behave incorrectly or give incorrect errors,
2577        because when scanning forward to find the reference group, PCRE was not
2578        taking into account the duplicate group numbers. A pattern such as
2579        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2580    
2581    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2582        the feature of a custom pointer more persuasive (as requested by a user).
2583    
2584    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2585        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2586        pre-defined.
2587    
2588    28. Added support for (*UTF8) at the start of a pattern.
2589    
2590    29. Arrange for flags added by the "release type" setting in CMake to be shown
2591        in the configuration summary.
2592    
2593    
2594    Version 7.8 05-Sep-08
2595    ---------------------
2596    
2597    1.  Replaced UCP searching code with optimized version as implemented for Ad
2598        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2599        stage table and inline lookup instead of a function, giving speed ups of 2
2600        to 5 times on some simple patterns that I tested. Permission was given to
2601        distribute the MultiStage2.py script that generates the tables (it's not in
2602        the tarball, but is in the Subversion repository).
2603    
2604    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2605        scripts.
2606    
2607    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2608        a group with a zero qualifier. The result of the study could be incorrect,
2609        or the function might crash, depending on the pattern.
2610    
2611    4.  Caseless matching was not working for non-ASCII characters in back
2612        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2613        It now works when Unicode Property Support is available.
2614    
2615    5.  In pcretest, an escape such as \x{de} in the data was always generating
2616        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2617        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2618        truncation.
2619    
2620    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2621    
2622    7.  Added two (int) casts to pcregrep when printing the difference of two
2623        pointers, in case they are 64-bit values.
2624    
2625    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2626        test 2 if it fails.
2627    
2628    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2629        and a #define of that name to empty if it is not externally set. This is to
2630        allow users of MSVC to set it if necessary.
2631    
2632    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2633        the convenience functions in the pcre_get.c source file.
2634    
2635    11. An option change at the start of a pattern that had top-level alternatives
2636        could cause overwriting and/or a crash. This command provoked a crash in
2637        some environments:
2638    
2639          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2640    
2641        This potential security problem was recorded as CVE-2008-2371.
2642    
2643    12. For a pattern where the match had to start at the beginning or immediately
2644        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2645        pcre_dfa_exec() could read past the end of the passed subject if there was
2646        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2647        pcretest so that it places the subject at the end of its malloc-ed buffer.
2648    
2649    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2650        exec() might read past the end of the data buffer in UTF-8 mode.
2651    
2652    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2653        the data contained the byte 0x85 as part of a UTF-8 character within its
2654        first line. This applied both to normal and DFA matching.
2655    
2656    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2657        /^[^d]*?$/8 failed to match "abc".
2658    
2659    16. Added a missing copyright notice to pcrecpp_internal.h.
2660    
2661    17. Make it more clear in the documentation that values returned from
2662        pcre_exec() in ovector are byte offsets, not character counts.
2663    
2664    18. Tidied a few places to stop certain compilers from issuing warnings.
2665    
2666    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2667        supplied by Stefan Weber. I made a further small update for 7.8 because
2668        there is a change of source arrangements: the pcre_searchfuncs.c module is
2669        replaced by pcre_ucd.c.
2670    
2671    
2672    Version 7.7 07-May-08
2673    ---------------------
2674    
2675    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2676        a string to a long long, pretend we don't even have a long long." This is
2677        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2678    
2679    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2680        pre-7.6 versions, which defined a global no_arg variable instead of putting
2681        it in the RE class. (See also #8 below.)
2682    
2683    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2684        Lopes.
2685    
2686    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2687    
2688        (1) The include/exclude patterns were being applied to the whole pathnames
2689            of files, instead of just to the final components.
2690    
2691        (2) If there was more than one level of directory, the subdirectories were
2692            skipped unless they satisfied the include/exclude conditions. This is
2693            inconsistent with GNU grep (and could even be seen as contrary to the
2694            pcregrep specification - which I improved to make it absolutely clear).
2695            The action now is always to scan all levels of directory, and just
2696            apply the include/exclude patterns to regular files.
2697    
2698    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2699        --exclude_dir in the tests to avoid scanning .svn directories.
2700    
2701    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2702        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2703        doesn't support NULs in patterns.
2704    
2705    7.  Added some missing "const"s to declarations of static tables in
2706        pcre_compile.c and pcre_dfa_exec.c.
2707    
2708    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2709        caused by fix #2  above. (Subsequently also a second patch to fix the
2710        first patch. And a third patch - this was a messy problem.)
2711    
2712    9.  Applied Craig's patch to remove the use of push_back().
2713    
2714    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2715        matching function regexec().
2716    
2717    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2718        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2719        references. PCRE supports relative numbers with this syntax (I don't think
2720        Oniguruma does).
2721    
2722    12. Previously, a group with a zero repeat such as (...){0} was completely
2723        omitted from the compiled regex. However, this means that if the group
2724        was called as a subroutine from elsewhere in the pattern, things went wrong
2725        (an internal error was given). Such groups are now left in the compiled
2726        pattern, with a new opcode that causes them to be skipped at execution
2727        time.
2728    
2729    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2730        to the way PCRE behaves:
2731    
2732        (a) A lone ] character is dis-allowed (Perl treats it as data).
2733    
2734        (b) A back reference to an unmatched subpattern matches an empty string
2735            (Perl fails the current match path).
2736    
2737        (c) A data ] in a character class must be notated as \] because if the
2738            first data character in a class is ], it defines an empty class. (In
2739            Perl it is not possible to have an empty class.) The empty class []
2740            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2741            The negative empty class [^] matches any one character, independently
2742            of the DOTALL setting.
2743    
2744    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2745        non-existent subpattern following a character class starting with ']' and
2746        containing () gave an internal compiling error instead of "reference to
2747        non-existent subpattern". Fortunately, when the pattern did exist, the
2748        compiled code was correct. (When scanning forwards to check for the
2749        existence of the subpattern, it was treating the data ']' as terminating
2750        the class, so got the count wrong. When actually compiling, the reference
2751        was subsequently set up correctly.)
2752    
2753    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2754        it was being rejected as not supported by pcre_dfa_exec(), even though
2755        other assertions are supported. I have made pcre_dfa_exec() support
2756        (*FAIL).
2757    
2758    16. The implementation of 13c above involved the invention of a new opcode,
2759        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2760        cannot be changed at match time, I realized I could make a small
2761        improvement to matching performance by compiling OP_ALLANY instead of
2762        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2763        on the OP_ANY path.
2764    
2765    17. Compiling pcretest on Windows with readline support failed without the
2766        following two fixes: (1) Make the unistd.h include conditional on
2767        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2768    
2769    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2770        ncurses library to be included for pcretest when ReadLine support is
2771        requested, but also to allow for it to be overridden. This patch came from
2772        Daniel Bergström.
2773    
2774    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2775        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2776        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2777        spotting this.
2778    
2779    
2780    Version 7.6 28-Jan-08
2781    ---------------------
2782    
2783    1.  A character class containing a very large number of characters with
2784        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2785        overflow.
2786    
2787    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2788        HAVE_LONG_LONG is not defined.
2789    
2790    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2791        bring them up to date and include new features. This patch includes:
2792    
2793        - Fixed PH's badly added libz and libbz2 support.
2794        - Fixed a problem with static linking.
2795        - Added pcredemo. [But later removed - see 7 below.]
2796        - Fixed dftables problem and added an option.
2797        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2798            HAVE_LONG_LONG.
2799        - Added readline support for pcretest.
2800        - Added an listing of the option settings after cmake has run.
2801    
2802    4.  A user submitted a patch to Makefile that makes it easy to create
2803        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2804        Makefile.am that cause it to include this special target, without
2805        affecting anything else. Note that the same mingw target plus all
2806        the other distribution libraries and programs are now supported
2807        when configuring with CMake (see 6 below) instead of with
2808        Configure/Make.
2809    
2810    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2811        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2812        exported in the Windows port". It has not yet been confirmed that the patch
2813        solves the problem, but it does no harm.
2814    
2815    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2816        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2817        with CMake, and also correct the comment about stack recursion.
2818    
2819    7.  Remove the automatic building of pcredemo from the ./configure system and
2820        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2821        of a program that users should build themselves after PCRE is installed, so
2822        building it automatically is not really right. What is more, it gave
2823        trouble in some build environments.
2824    
2825    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2826    
2827    
2828    Version 7.5 10-Jan-08
2829    ---------------------
2830    
2831    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2832        values in parens when parsing an RE using the C++ wrapper."
2833    
2834    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2835        Characters greater than 255 were excluded from the class instead of being
2836        included.
2837    
2838    3.  The same bug as (2) above applied to negated POSIX classes such as
2839        [:^space:].
2840    
2841    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2842        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2843        I have changed it.
2844    
2845    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2846        first named subpattern) and a construct such as (?&a) would reference the
2847        first named subpattern whose name started with "a" (in other words, the
2848        length check was missing). Both these problems are fixed. "Subpattern name
2849        expected" is now given for (?&) (a zero-length name), and this patch also
2850        makes it give the same error for \k'' (previously it complained that that
2851        was a reference to a non-existent subpattern).
2852    
2853    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2854        this is right because (?- can be followed by option settings as well as by
2855        digits. I have, however, made the messages clearer.
2856    
2857    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2858        than the number used in the conditional) now cause a compile-time error.
2859        This is actually not compatible with Perl, which accepts such patterns, but
2860        treats the conditional as always being FALSE (as PCRE used to), but it
2861        seems to me that giving a diagnostic is better.
2862    
2863    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2864        and messages.
2865    
2866    9.  Fix two occurrences of "backslash" in comments that should have been
2867        "backspace".
2868    
2869    10. Remove two redundant lines of code that can never be obeyed (their function
2870        was moved elsewhere).
2871    
2872    11. The program that makes PCRE's Unicode character property table had a bug
2873        which caused it to generate incorrect table entries for sequences of
2874        characters that have the same character type, but are in different scripts.
2875        It amalgamated them into a single range, with the script of the first of
2876        them. In other words, some characters were in the wrong script. There were
2877        thirteen such cases, affecting characters in the following ranges:
2878    
2879          U+002b0 - U+002c1
2880          U+0060c - U+0060d
2881          U+0061e - U+00612
2882          U+0064b - U+0065e
2883          U+0074d - U+0076d
2884          U+01800 - U+01805
2885          U+01d00 - U+01d77
2886          U+01d9b - U+01dbf
2887          U+0200b - U+0200f
2888          U+030fc - U+030fe
2889          U+03260 - U+0327f
2890          U+0fb46 - U+0fbb1
2891          U+10450 - U+1049d
2892    
2893    12. The -o option (show only the matching part of a line) for pcregrep was not
2894        compatible with GNU grep in that, if there was more than one match in a
2895        line, it showed only the first of them. It now behaves in the same way as
2896        GNU grep.
2897    
2898    13. If the -o and -v options were combined for pcregrep, it printed a blank
2899        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2900        does the same. The return code can be used to tell if there were any
2901        non-matching lines.
2902    
2903    14. Added --file-offsets and --line-offsets to pcregrep.
2904    
2905    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2906        infinitely looping recursion. The bug was that positive lookaheads were not
2907        being skipped when checking for a possible empty match (negative lookaheads
2908        and both kinds of lookbehind were skipped).
2909    
2910    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2911        inclusion of <windows.h> to before rather than after the definition of
2912        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2913    
2914    17. Specifying a possessive quantifier with a specific limit for a Unicode
2915        character property caused pcre_compile() to compile bad code, which led at
2916        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2917        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2918        caused the error; without that there was no problem.
2919    
2920    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2921    
2922    19. Added --enable-pcretest-libreadline.
2923    
2924    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2925        RE::GlobalReplace(). As a result, the number of replacements returned was
2926        double what it should be. I removed one of the increments, but Craig sent a
2927        later patch that removed the other one (the right fix) and added unit tests
2928        that check the return values (which was not done before).
2929    
2930    21. Several CMake things:
2931    
2932        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2933            the names libpcre and libpcreposix, not just pcre and pcreposix.
2934    
2935        (2) The above change means that pcretest and pcregrep are now correctly
2936            linked with the newly-built libraries, not previously installed ones.
2937    
2938        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2939    
2940    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2941        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2942        UTF-8 newline character). The key issue is that the pattern starts .*;
2943        this means that the match must be either at the beginning, or after a
2944        newline. The bug was in the code for advancing after a failed match and
2945        checking that the new position followed a newline. It was not taking
2946        account of UTF-8 characters correctly.
2947    
2948    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2949        character classes. PCRE was not treating the sequence [:...:] as a
2950        character class unless the ... were all letters. Perl, however, seems to
2951        allow any characters between [: and :], though of course it rejects as
2952        unknown any "names" that contain non-letters, because all the known class
2953        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2954        for example, whereas PCRE did not - it did not recognize a POSIX character
2955        class. This seemed a bit dangerous, so the code has been changed to be
2956        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2957        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2958        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2959        Perl does, and where it didn't before.
2960    
2961    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2962        Windows environments %n is disabled by default.
2963    
2964    
2965    Version 7.4 21-Sep-07
2966    ---------------------
2967    
2968    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2969        means that a class such as [\s] counted as "explicit reference to CR or
2970        LF". That isn't really right - the whole point of the change was to try to
2971        help when there was an actual mention of one of the two characters. So now
2972        the change happens only if \r or \n (or a literal CR or LF) character is
2973        encountered.
2974    
2975    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2976        of both had grown to the point where there were only 3 bits left.
2977        Fortunately, there was spare space in the data structure, and so I have
2978        moved the internal flags into a new 16-bit field to free up more option
2979        bits.
2980    
2981    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2982        but did not set the internal JCHANGED flag - either of these is enough to
2983        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2984        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2985        start sets both bits.
2986    
2987    4.  Added options (at build time, compile time, exec time) to change \R from
2988        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2989    
2990    5.  doc/pcresyntax.html was missing from the distribution.
2991    
2992    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2993        compatibility, even though it is no longer used.
2994    
2995    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2996        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2997        windows.h file is present (where different names are used). [This was
2998        reversed later after testing - see 16 below.]
2999    
3000    8.  Changed all #include <config.h> to #include "config.h". There were also
3001        some further <pcre.h> cases that I changed to "pcre.h".
3002    
3003    9.  When pcregrep was used with the --colour option, it missed the line ending
3004        sequence off the lines that it output.
3005    
3006    10. It was pointed out to me that arrays of string pointers cause lots of
3007        relocations when a shared library is dynamically loaded. A technique of
3008        using a single long string with a table of offsets can drastically reduce
3009        these. I have refactored PCRE in four places to do this. The result is
3010        dramatic:
3011    
3012          Originally:                          290
3013          After changing UCP table:            187
3014          After changing error message table:   43
3015          After changing table of "verbs"       36
3016          After changing table of Posix names   22
3017    
3018        Thanks to the folks working on Gregex for glib for this insight.
3019    
3020    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3021        unicode-properties was also set.
3022    
3023    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3024    
3025    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3026        checked only for CRLF.
3027    
3028    14. Added casts to pcretest.c to avoid compiler warnings.
3029    
3030    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3031    
3032    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3033        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3034        entirely. This removes changes made in 7 above.
3035    
3036    17. The CMake files have been updated, and there is now more information about
3037        building with CMake in the NON-UNIX-USE document.
3038    
3039    
3040    Version 7.3 28-Aug-07
3041    ---------------------
3042    
3043     1. In the rejigging of the build system that eventually resulted in 7.1, the
3044        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3045        brackets there is not right, since it causes compilers to look for an
3046        installed pcre.h, not the version that is in the source that is being
3047        compiled (which of course may be different). I have changed it back to:
3048    
3049          #include "pcre.h"
3050    
3051        I have a vague recollection that the change was concerned with compiling in
3052        different directories, but in the new build system, that is taken care of
3053        by the VPATH setting the Makefile.
3054    
3055     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3056        when the subject happened to end in the byte 0x85 (e.g. if the last
3057        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3058        characters but of course it shouldn't be taken as a newline when it is part
3059        of another character. The bug was that, for an unlimited repeat of . in
3060        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3061        characters when looking for a newline.
3062    
3063     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3064    
3065     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3066        in debug output.
3067    
3068     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3069        long printing in the pcrecpp unittest when running under MinGW.
3070    
3071     6. ESC_K was left out of the EBCDIC table.
3072    
3073     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3074        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3075        limit also applies to "virtual nesting" when a pattern is recursive, and in
3076        this case 1000 isn't so big. I have been able to remove this limit at the
3077        expense of backing off one optimization in certain circumstances. Normally,
3078        when pcre_exec() would call its internal match() function recursively and
3079        immediately return the result unconditionally, it uses a "tail recursion"
3080        feature to save stack. However, when a subpattern that can match an empty
3081        string has an unlimited repetition quantifier, it no longer makes this
3082        optimization. That gives it a stack frame in which to save the data for
3083        checking that an empty string has been matched. Previously this was taken
3084        from the 1000-entry workspace that had been reserved. So now there is no
3085        explicit limit, but more stack is used.
3086    
3087     8. Applied Daniel's patches to solve problems with the import/export magic
3088        syntax that is required for Windows, and which was going wrong for the
3089        pcreposix and pcrecpp parts of the library. These were overlooked when this
3090        problem was solved for the main library.
3091    
3092     9. There were some crude static tests to avoid integer overflow when computing
3093        the size of patterns that contain repeated groups with explicit upper
3094        limits. As the maximum quantifier is 65535, the maximum group length was
3095        set at 30,000 so that the product of these two numbers did not overflow a
3096        32-bit integer. However, it turns out that people want to use groups that
3097        are longer than 30,000 bytes (though not repeat them that many times).
3098        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3099        made it possible to implement the integer overflow checks in a much more
3100        dynamic way, which I have now done. The artificial limitation on group
3101        length has been removed - we now have only the limit on the total length of
3102        the compiled pattern, which depends on the LINK_SIZE setting.
3103    
3104    10. Fixed a bug in the documentation for get/copy named substring when
3105        duplicate names are permitted. If none of the named substrings are set, the
3106        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3107        empty string.
3108    
3109    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3110        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3111        because the ] is interpreted as the first data character and the
3112        terminating ] is not found. PCRE has been made compatible with Perl in this
3113        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3114        cause memory overwriting.
3115    
3116    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3117        string has been matched (to stop an infinite loop). It was not recognizing
3118        a conditional subpattern that could match an empty string if that
3119        subpattern was within another subpattern. For example, it looped when
3120        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3121        condition was not nested. This bug has been fixed.
3122    
3123    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3124        past the start of the subject in the presence of bytes with the top bit
3125        set, for example "\x8aBCD".
3126    
3127    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3128        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3129    
3130    14. Optimized (?!) to (*FAIL).
3131    
3132    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3133        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3134        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3135        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3136        does: it's just the validity check that is more restrictive.
3137    
3138    16. Inserted checks for integer overflows during escape sequence (backslash)
3139        processing, and also fixed erroneous offset values for syntax errors during
3140        backslash processing.
3141    
3142    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3143        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3144    
3145    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3146        caused an overrun.
3147    
3148    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3149        something other than just ASCII characters) inside a group that had an
3150        unlimited repeat caused a loop at compile time (while checking to see
3151        whether the group could match an empty string).
3152    
3153    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3154        [\P{Any}] did so. (Error in the code for printing property names.)
3155    
3156    21. An orphan \E inside a character class could cause a crash.
3157    
3158    22. A repeated capturing bracket such as (A)? could cause a wild memory
3159        reference during compilation.
3160    
3161    23. There are several functions in pcre_compile() that scan along a compiled
3162        expression for various reasons (e.g. to see if it's fixed length for look
3163        behind). There were bugs in these functions when a repeated \p or \P was
3164        present in the pattern. These operators have additional parameters compared
3165        with \d, etc, and these were not being taken into account when moving along
3166        the compiled data. Specifically:
3167    
3168        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3169            length.
3170    
3171        (b) An item such as \pL+ within a repeated group could cause crashes or
3172            loops.
3173    
3174        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3175            "reference to non-existent subpattern" error.
3176    
3177        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3178    
3179    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3180        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3181    
3182    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3183    
3184    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3185        character were causing crashes (broken optimization).
3186    
3187    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3188        \p or \P) caused a compile-time loop.
3189    
3190    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3191        break. For example, the unstudied pattern [\r\n]A does not match the string
3192        "\r\nA" because change 7.0/46 below moves the current point on by two
3193        characters after failing to match at the start. However, the pattern \nA
3194        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3195        the same is true. There doesn't seem any very clean way out of this, but
3196        what I have chosen to do makes the common cases work: PCRE now takes note
3197        of whether there can be an explicit match for \r or \n anywhere in the
3198        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3199        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3200        pattern has explicit CR or LF references.
3201    
3202    29. Added (*CR) etc for changing newline setting at start of pattern.
3203    
3204    
3205    Version 7.2 19-Jun-07
3206  ---------------------  ---------------------
3207    
3208   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 3222  Version 7.2 01-May-07
3222      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3223      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
3224      in all cases.      in all cases.
3225    
3226   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3227    
3228      (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
3229          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3230    
3231      (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
3232          to be opened parentheses.          to be opened parentheses.
3233    
3234      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3235          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3236    
3237        (d) \K resets the start of the current match so that everything before
3238            is not part of it.
3239    
3240        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3241    
3242        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3243            reference syntax.
3244    
3245        (g) (?| introduces a group in which the numbering of parentheses in each
3246            alternative starts with the same number.
3247    
3248        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3249    
3250     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3251        PCRE_INFO_JCHANGED.
3252    
3253     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3254        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3255        for detecting groups that can match an empty string.
3256    
3257     9. A pattern with a very large number of alternatives (more than several
3258        hundred) was running out of internal workspace during the pre-compile
3259        phase, where pcre_compile() figures out how much memory will be needed. A
3260        bit of new cunning has reduced the workspace needed for groups with
3261        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3262        workspace instead of running out of the 4096 that are available.
3263    
3264    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3265    
3266    11. Applied patch from Google to remove an optimization that didn't quite work.
3267        The report of the bug said:
3268    
3269          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3270          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3271          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3272    
3273    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3274        it matched the wrong number of bytes.
3275    
3276    
3277  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3278  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1636

  ViewVC Help
Powered by ViewVC 1.1.5