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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.158  
changed lines
  Added in v.1547

  ViewVC Help
Powered by ViewVC 1.1.5