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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 1487 by ph10, Wed Jun 18 17:17:03 2014 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Version 8.36 xx-xxx-2014
5    ------------------------
6    
7    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
8        -Wmissing-field-initializers and -Wunused-parameter.
9    
10    2.  The tests for quantifiers being too big (greater than 65535) were being
11        applied after reading the number, and stupidly assuming that integer
12        overflow would give a negative number. The tests are now applied as the
13        numbers are read.
14    
15    3.  Tidy code in pcre_exec.c where two branches that used to be different are
16        now the same.
17    
18    4.  The JIT compiler did not generate match limit checks for certain
19        bracketed expressions with quantifiers. This may lead to exponential
20        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21        issue should be resolved now.
22    
23    5.  Fixed an issue, which occures when nested alternatives are optimized
24        with table jumps.
25    
26    6.  Inserted two casts and changed some ints to size_t in the light of some
27        reported 64-bit compiler warnings (Bugzilla 1477).
28    
29    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
30        an empty string, which sometimes were behaving incorrectly in the
31        interpreter (though correctly in the JIT matcher). This pcretest input is
32        an example:
33    
34          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36    
37        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39        for an empty string was breaking the inner loop and carrying on at a lower
40        level, when possessive repeated groups should always return to a higher
41        level as they have no backtrack points in them. The empty string test now
42        occurs at the outer level.
43    
44    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46    
47    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48        doesn't).
49    
50    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51        to hold a pattern to be incorrectly computed (too small) when there were
52        named back references to duplicated names. This could cause "internal
53        error: code overflow" or "double free or corruption" or other memory
54        handling errors.
55    
56    11. When named subpatterns had the same prefixes, back references could be
57        confused. For example, in this pattern:
58    
59          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60    
61        the reference to 'Name' was incorrectly treated as a reference to a
62        duplicate name.
63    
64    12. A pattern such as /^s?c/mi8 where the optional character has more than
65        one "other case" was incorrectly compiled such that it would only try to
66        match starting at "c".
67    
68    13. When a pattern starting with \s was studied, VT was not included in the
69        list of possible starting characters; this should have been part of the
70        8.34/18 patch.
71    
72    14. If a character class started [\Qx]... where x is any character, the class
73        was incorrectly terminated at the ].
74    
75    
76    Version 8.35 04-April-2014
77    --------------------------
78    
79    1.  A new flag is set, when property checks are present in an XCLASS.
80        When this flag is not set, PCRE can perform certain optimizations
81        such as studying these XCLASS-es.
82    
83    2.  The auto-possessification of character sets were improved: a normal
84        and an extended character set can be compared now. Furthermore
85        the JIT compiler optimizes more character set checks.
86    
87    3.  Got rid of some compiler warnings for potentially uninitialized variables
88        that show up only when compiled with -O2.
89    
90    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
91        of a match later then the end of the match. The pcretest program was not
92        handling the case sensibly - it was outputting from the start to the next
93        binary zero. It now reports this situation in a message, and outputs the
94        text from the end to the start.
95    
96    5.  Fast forward search is improved in JIT. Instead of the first three
97        characters, any three characters with fixed position can be searched.
98        Search order: first, last, middle.
99    
100    6.  Improve character range checks in JIT. Characters are read by an inprecise
101        function now, which returns with an unknown value if the character code is
102        above a certain treshold (e.g: 256). The only limitation is that the value
103        must be bigger than the treshold as well. This function is useful, when
104        the characters above the treshold are handled in the same way.
105    
106    7.  The macros whose names start with RAWUCHAR are placeholders for a future
107        mode in which only the bottom 21 bits of 32-bit data items are used. To
108        make this more memorable for those maintaining the code, the names have
109        been changed to start with UCHAR21, and an extensive comment has been added
110        to their definition.
111    
112    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
113        to the export list in Makefile.am (they were accidentally omitted from the
114        8.34 tarball).
115    
116    9.  The informational output from pcretest used the phrase "starting byte set"
117        which is inappropriate for the 16-bit and 32-bit libraries. As the output
118        for "first char" and "need char" really means "non-UTF-char", I've changed
119        "byte" to "char", and slightly reworded the output. The documentation about
120        these values has also been (I hope) clarified.
121    
122    10. Another JIT related optimization: use table jumps for selecting the correct
123        backtracking path, when more than four alternatives are present inside a
124        bracket.
125    
126    11. Empty match is not possible, when the minimum length is greater than zero,
127        and there is no \K in the pattern. JIT should avoid empty match checks in
128        such cases.
129    
130    12. In a caseless character class with UCP support, when a character with more
131        than one alternative case was not the first character of a range, not all
132        the alternative cases were added to the class. For example, s and \x{17f}
133        are both alternative cases for S: the class [RST] was handled correctly,
134        but [R-T] was not.
135    
136    13. The configure.ac file always checked for pthread support when JIT was
137        enabled. This is not used in Windows, so I have put this test inside a
138        check for the presence of windows.h (which was already tested for).
139    
140    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
141        The algorithm provides a way to skip certain starting offsets, and usually
142        faster than linear prefix searches.
143    
144    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
145        as for 'fr_FR' and 'french'. For some reason, however, it then used the
146        Windows-specific input and output files, which have 'french' screwed in.
147        So this could never have worked. One of the problems with locales is that
148        they aren't always the same. I have now updated RunTest so that it checks
149        the output of the locale test (test 3) against three different output
150        files, and it allows the test to pass if any one of them matches. With luck
151        this should make the test pass on some versions of Solaris where it was
152        failing. Because of the uncertainty, the script did not used to stop if
153        test 3 failed; it now does. If further versions of a French locale ever
154        come to light, they can now easily be added.
155    
156    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
157        there was a message during ./configure, but it did not stop. This now
158        provokes an error. The invalid example in README has been corrected.
159        If a value less than the minimum is given, the minimum value has always
160        been used, but now a warning is given.
161    
162    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
163        was a bug in the test system, which is now fixed. Also, the list of various
164        configurations that are tested for each release did not have one with both
165        16/32 bits and --enable-bar-anycrlf. It now does.
166    
167    18. pcretest was missing "-C bsr" for displaying the \R default setting.
168    
169    19. Little endian PowerPC systems are supported now by the JIT compiler.
170    
171    20. The fast forward newline mechanism could enter to an infinite loop on
172        certain invalid UTF-8 input. Although we don't support these cases
173        this issue can be fixed by a performance optimization.
174    
175    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
176        not take account if existing stack usage. There is now a new global
177        variable called pcre_stack_guard that can be set to point to an external
178        function to check stack availability. It is called at the start of
179        processing every parenthesized group.
180    
181    22. A typo in the code meant that in ungreedy mode the max/min qualifier
182        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
183        not match "ab".
184    
185    23. When UTF was disabled, the JIT program reported some incorrect compile
186        errors. These messages are silenced now.
187    
188    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
189        compiler.
190    
191    25. Change all the temporary files used in RunGrepTest to be different to those
192        used by RunTest so that the tests can be run simultaneously, for example by
193        "make -j check".
194    
195    
196    Version 8.34 15-December-2013
197    -----------------------------
198    
199    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
200        executable memory. Patch inspired by Carsten Klein.
201    
202    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
203        this macro is never tested and has no effect, because the work to support
204        coverage involves only compiling and linking options and special targets in
205        the Makefile. The comment in config.h implied that defining the macro would
206        enable coverage support, which is totally false. There was also support for
207        setting this macro in the CMake files (my fault, I just copied it from
208        configure). SUPPORT_GCOV has now been removed.
209    
210    3.  Make a small performance improvement in strlen16() and strlen32() in
211        pcretest.
212    
213    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
214        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
215        these cases). There was also one in pcretest.c.
216    
217    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
218    
219    6.  In UTF mode, the code for checking whether a group could match an empty
220        string (which is used for indefinitely repeated groups to allow for
221        breaking an infinite loop) was broken when the group contained a repeated
222        negated single-character class with a character that occupied more than one
223        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
224        UTF-8 mode). The effect was undefined: the group might or might not be
225        deemed as matching an empty string, or the program might have crashed.
226    
227    7.  The code for checking whether a group could match an empty string was not
228        recognizing that \h, \H, \v, \V, and \R must match a character.
229    
230    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
231        an empty string. If it can, pcretest shows this in its information output.
232    
233    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
234        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
235        matched by pcre_exec() without using JIT:
236    
237        (a) If the rest of the pattern did not match after a maximal run of
238            grapheme clusters, the code for backing up to try with fewer of them
239            did not always back up over a full grapheme when characters that do not
240            have the modifier quality were involved, e.g. Hangul syllables.
241    
242        (b) If the match point in a subject started with modifier character, and
243            there was no match, the code could incorrectly back up beyond the match
244            point, and potentially beyond the first character in the subject,
245            leading to a segfault or an incorrect match result.
246    
247    10. A conditional group with an assertion condition could lead to PCRE
248        recording an incorrect first data item for a match if no other first data
249        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
250        first data item, and therefore matched "ca" after "c" instead of at the
251        start.
252    
253    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
254        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
255    
256    12. The source of pcregrep now includes z/OS-specific code so that it can be
257        compiled for z/OS as part of the special z/OS distribution.
258    
259    13. Added the -T and -TM options to pcretest.
260    
261    14. The code in pcre_compile.c for creating the table of named capturing groups
262        has been refactored. Instead of creating the table dynamically during the
263        actual compiling pass, the information is remembered during the pre-compile
264        pass (on the stack unless there are more than 20 named groups, in which
265        case malloc() is used) and the whole table is created before the actual
266        compile happens. This has simplified the code (it is now nearly 150 lines
267        shorter) and prepared the way for better handling of references to groups
268        with duplicate names.
269    
270    15. A back reference to a named subpattern when there is more than one of the
271        same name now checks them in the order in which they appear in the pattern.
272        The first one that is set is used for the reference. Previously only the
273        first one was inspected. This change makes PCRE more compatible with Perl.
274    
275    16. Unicode character properties were updated from Unicode 6.3.0.
276    
277    17. The compile-time code for auto-possessification has been refactored, based
278        on a patch by Zoltan Herczeg. It now happens after instead of during
279        compilation. The code is cleaner, and more cases are handled. The option
280        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
281        options in pcretest are provided to set it. It can also be set by
282        (*NO_AUTO_POSSESS) at the start of a pattern.
283    
284    18. The character VT has been added to the default ("C" locale) set of
285        characters that match \s and are generally treated as white space,
286        following this same change in Perl 5.18. There is now no difference between
287        "Perl space" and "POSIX space". Whether VT is treated as white space in
288        other locales depends on the locale.
289    
290    19. The code for checking named groups as conditions, either for being set or
291        for being recursed, has been refactored (this is related to 14 and 15
292        above). Processing unduplicated named groups should now be as fast at
293        numerical groups, and processing duplicated groups should be faster than
294        before.
295    
296    20. Two patches to the CMake build system, by Alexander Barkov:
297    
298          (1) Replace the "source" command by "." in CMakeLists.txt because
299              "source" is a bash-ism.
300    
301          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
302              without these the CMake build does not work on Solaris.
303    
304    21. Perl has changed its handling of \8 and \9. If there is no previously
305        encountered capturing group of those numbers, they are treated as the
306        literal characters 8 and 9 instead of a binary zero followed by the
307        literals. PCRE now does the same.
308    
309    22. Following Perl, added \o{} to specify codepoints in octal, making it
310        possible to specify values greater than 0777 and also making them
311        unambiguous.
312    
313    23. Perl now gives an error for missing closing braces after \x{... instead of
314        treating the string as literal. PCRE now does the same.
315    
316    24. RunTest used to grumble if an inappropriate test was selected explicitly,
317        but just skip it when running all tests. This make it awkward to run ranges
318        of tests when one of them was inappropriate. Now it just skips any
319        inappropriate tests, as it always did when running all tests.
320    
321    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
322        character types such as \d or \w, too many callouts were inserted, and the
323        data that they returned was rubbish.
324    
325    26. In UCP mode, \s was not matching two of the characters that Perl matches,
326        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
327        were matched by \h. The code has now been refactored so that the lists of
328        the horizontal and vertical whitespace characters used for \h and \v (which
329        are defined only in one place) are now also used for \s.
330    
331    27. Add JIT support for the 64 bit TileGX architecture.
332        Patch by Jiong Wang (Tilera Corporation).
333    
334    28. Possessive quantifiers for classes (both explicit and automatically
335        generated) now use special opcodes instead of wrapping in ONCE brackets.
336    
337    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
338        (because it's meaningless), this was not happening when PCRE_CASELESS was
339        set. Not wrong, but inefficient.
340    
341    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
342        properties for \w, \d, etc) is present in a test regex. Otherwise if the
343        test contains no characters greater than 255, Perl doesn't realise it
344        should be using Unicode semantics.
345    
346    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
347        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
348        does in Unicode mode.
349    
350    32. Added the "forbid" facility to pcretest so that putting tests into the
351        wrong test files can sometimes be quickly detected.
352    
353    33. There is now a limit (default 250) on the depth of nesting of parentheses.
354        This limit is imposed to control the amount of system stack used at compile
355        time. It can be changed at build time by --with-parens-nest-limit=xxx or
356        the equivalent in CMake.
357    
358    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
359        errors. Perl warns for these when in warning mode, but PCRE has no facility
360        for giving warnings.
361    
362    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
363        However, this was not working for (?!) because it is optimized to (*FAIL),
364        for which PCRE does not allow quantifiers. The optimization is now disabled
365        when a quantifier follows (?!). I can't see any use for this, but it makes
366        things uniform.
367    
368    36. Perl no longer allows group names to start with digits, so I have made this
369        change also in PCRE. It simplifies the code a bit.
370    
371    37. In extended mode, Perl ignores spaces before a + that indicates a
372        possessive quantifier. PCRE allowed a space before the quantifier, but not
373        before the possessive +. It now does.
374    
375    38. The use of \K (reset reported match start) within a repeated possessive
376        group such as (a\Kb)*+ was not working.
377    
378    40. Document that the same character tables must be used at compile time and
379        run time, and that the facility to pass tables to pcre_exec() and
380        pcre_dfa_exec() is for use only with saved/restored patterns.
381    
382    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
383        features for Builds with MSVC:
384    
385        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
386           builds). The .rc files can be used to set FileDescription and many other
387           attributes.
388    
389        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
390           This allows higher-level build scripts which want .pdb files to avoid
391           hard-coding the exact files needed."
392    
393    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
394        mean "start of word" and "end of word", respectively, as a transition aid.
395    
396    43. A minimizing repeat of a class containing codepoints greater than 255 in
397        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
398        compiled to use the heap for recursion.
399    
400    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
401        is configured.
402    
403    
404    Version 8.33 28-May-2013
405    ------------------------
406    
407    1.  Added 'U' to some constants that are compared to unsigned integers, to
408        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
409        variables that are added to signed variables, to ensure the result is
410        signed and can be negated.
411    
412    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
413        CMake config files.
414    
415    3.  Revise the creation of config.h.generic so that all boolean macros are
416        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
417        overriding via -D on the command line possible.
418    
419    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
420        to unsigned int is reported to make a quite noticeable speed difference in
421        a specific Windows environment. Testing on Linux did also appear to show
422        some benefit (and it is clearly not harmful). Also fixed the definition of
423        Xop which should be unsigned.
424    
425    5.  Related to (4), changing the definition of the intermediate variable cc
426        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
427        improvements.
428    
429    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
430        unnecessary spaces.
431    
432    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
433        and later.
434    
435    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
436    
437    9.  Optimizing fast_forward_start_bits in JIT.
438    
439    10. Adding support for callouts in JIT, and fixing some issues revealed
440        during this work. Namely:
441    
442        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
443    
444        (b) Minimum length was not checked before the matching is started.
445    
446    11. The value of capture_last that is passed to callouts was incorrect in some
447        cases when there was a capture on one path that was subsequently abandoned
448        after a backtrack. Also, the capture_last value is now reset after a
449        recursion, since all captures are also reset in this case.
450    
451    12. The interpreter no longer returns the "too many substrings" error in the
452        case when an overflowing capture is in a branch that is subsequently
453        abandoned after a backtrack.
454    
455    13. In the pathological case when an offset vector of size 2 is used, pcretest
456        now prints out the matched string after a yield of 0 or 1.
457    
458    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
459        Only supported by the JIT compiler at the moment.
460    
461    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
462    
463    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
464        offset of the starting point of the matching process, provided the offsets
465        vector is large enough.
466    
467    17. The \A escape now records a lookbehind value of 1, though its execution
468        does not actually inspect the previous character. This is to ensure that,
469        in partial multi-segment matching, at least one character from the old
470        segment is retained when a new segment is processed. Otherwise, if there
471        are no lookbehinds in the pattern, \A might match incorrectly at the start
472        of a new segment.
473    
474    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
475    
476    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
477        gives some modest performance improvement in 8-bit mode.
478    
479    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
480        be expressed in certain programming languages using Universal Character
481        Names.
482    
483    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
484        which points out that "non characters" are not "characters that may not
485        appear in Unicode strings" but rather "characters that are reserved for
486        internal use and have only local meaning".
487    
488    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
489        there was a conditional group that depended on an assertion, if the
490        assertion was false, the callout that immediately followed the alternation
491        in the condition was skipped when pcre_exec() was used for matching.
492    
493    23. Allow an explicit callout to be inserted before an assertion that is the
494        condition for a conditional group, for compatibility with automatic
495        callouts, which always insert a callout at this point.
496    
497    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
498        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
499    
500    25. (*PRUNE) is now supported by the JIT compiler.
501    
502    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
503    
504    27. Fix the case where there are two or more SKIPs with arguments that may be
505        ignored.
506    
507    28. (*SKIP) is now supported by the JIT compiler.
508    
509    29. (*THEN) is now supported by the JIT compiler.
510    
511    30. Update RunTest with additional test selector options.
512    
513    31. The way PCRE handles backtracking verbs has been changed in two ways.
514    
515        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
516        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
517        backtracking. In some cases this makes it more Perl-compatible, but Perl's
518        rather obscure rules do not always do the same thing.
519    
520        (2) Previously, backtracking verbs were confined within assertions. This is
521        no longer the case for positive assertions, except for (*ACCEPT). Again,
522        this sometimes improves Perl compatibility, and sometimes does not.
523    
524    32. A number of tests that were in test 2 because Perl did things differently
525        have been moved to test 1, because either Perl or PCRE has changed, and
526        these tests are now compatible.
527    
528    32. Backtracking control verbs are now handled in the same way in JIT and
529        interpreter.
530    
531    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
532        contained a forward subroutine reference caused a compile error.
533    
534    34. Auto-detect and optimize limited repetitions in JIT.
535    
536    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
537        blocking (*UTF) etc.
538    
539    36. In the interpreter, maximizing pattern repetitions for characters and
540        character types now use tail recursion, which reduces stack usage.
541    
542    37. The value of the max lookbehind was not correctly preserved if a compiled
543        and saved regex was reloaded on a host of different endianness.
544    
545    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
546        of the compiled pattern block, expand the flags field from 16 to 32 bits
547        because it was almost full.
548    
549    39. Try madvise first before posix_madvise.
550    
551    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
552        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
553        original change locked out all matches of empty strings. This has been
554        changed so that one match of an empty string per line is recognized.
555        Subsequent searches on the same line (for colouring or for --only-matching,
556        for example) do not recognize empty strings.
557    
558    41. Applied a user patch to fix a number of spelling mistakes in comments.
559    
560    42. Data lines longer than 65536 caused pcretest to crash.
561    
562    43. Clarified the data type for length and startoffset arguments for pcre_exec
563        and pcre_dfa_exec in the function-specific man pages, where they were
564        explicitly stated to be in bytes, never having been updated. I also added
565        some clarification to the pcreapi man page.
566    
567    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
568        a segmentation fault.
569    
570    
571    Version 8.32 30-November-2012
572    -----------------------------
573    
574    1.  Improved JIT compiler optimizations for first character search and single
575        character iterators.
576    
577    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
578        Patch by Daniel Richard G.
579    
580    3.  Single character iterator optimizations in the JIT compiler.
581    
582    4.  Improved JIT compiler optimizations for character ranges.
583    
584    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
585        Reported by Giuseppe D'Angelo.
586    
587    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
588        of a line, was being set incorrectly in cases where .* appeared inside
589        atomic brackets at the start of a pattern, or where there was a subsequent
590        *PRUNE or *SKIP.
591    
592    7.  Improved instruction cache flush for POWER/PowerPC.
593        Patch by Daniel Richard G.
594    
595    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
596        grep:
597    
598        (a) There is now no limit to the number of patterns to be matched.
599    
600        (b) An error is given if a pattern is too long.
601    
602        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
603            are now supported.
604    
605        (d) --exclude-from and --include-from (multiple use) have been added.
606    
607        (e) Exclusions and inclusions now apply to all files and directories, not
608            just to those obtained from scanning a directory recursively.
609    
610        (f) Multiple uses of -f and --file-list are now supported.
611    
612        (g) In a Windows environment, the default for -d has been changed from
613            "read" (the GNU grep default) to "skip", because otherwise the presence
614            of a directory in the file list provokes an error.
615    
616        (h) The documentation has been revised and clarified in places.
617    
618    9.  Improve the matching speed of capturing brackets.
619    
620    10. Changed the meaning of \X so that it now matches a Unicode extended
621        grapheme cluster.
622    
623    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
624        out POSIX threads when JIT support is configured.
625    
626    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
627    
628    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
629        structure could go wrong in environments where size_t is not the same size
630        as int.
631    
632    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
633        set.
634    
635    15. The EBCDIC support had decayed; later updates to the code had included
636        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
637        general tidy up of EBCDIC-related issues, and the documentation was also
638        not quite right. There is now a test that can be run on ASCII systems to
639        check some of the EBCDIC-related things (but is it not a full test).
640    
641    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
642        in a small tidy to the code.
643    
644    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
645    
646    18. If the --only-matching (-o) option in pcregrep is specified multiple
647        times, each one causes appropriate output. For example, -o1 -o2 outputs the
648        substrings matched by the 1st and 2nd capturing parentheses. A separating
649        string can be specified by --om-separator (default empty).
650    
651    19. Improving the first n character searches.
652    
653    20. Turn case lists for horizontal and vertical white space into macros so that
654        they are defined only once.
655    
656    21. This set of changes together give more compatible Unicode case-folding
657        behaviour for characters that have more than one other case when UCP
658        support is available.
659    
660        (a) The Unicode property table now has offsets into a new table of sets of
661            three or more characters that are case-equivalent. The MultiStage2.py
662            script that generates these tables (the pcre_ucd.c file) now scans
663            CaseFolding.txt instead of UnicodeData.txt for character case
664            information.
665    
666        (b) The code for adding characters or ranges of characters to a character
667            class has been abstracted into a generalized function that also handles
668            case-independence. In UTF-mode with UCP support, this uses the new data
669            to handle characters with more than one other case.
670    
671        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
672            whose other case is greater than 256 are now correctly matched
673            caselessly. Previously, the high codepoint matched the low one, but not
674            vice versa.
675    
676        (d) The processing of \h, \H, \v, and \ in character classes now makes use
677            of the new class addition function, using character lists defined as
678            macros alongside the case definitions of 20 above.
679    
680        (e) Caseless back references now work with characters that have more than
681            one other case.
682    
683        (f) General caseless matching of characters with more than one other case
684            is supported.
685    
686    22. Unicode character properties were updated from Unicode 6.2.0
687    
688    23. Improved CMake support under Windows. Patch by Daniel Richard G.
689    
690    24. Add support for 32-bit character strings, and UTF-32
691    
692    25. Major JIT compiler update (code refactoring and bugfixing).
693        Experimental Sparc 32 support is added.
694    
695    26. Applied a modified version of Daniel Richard G's patch to create
696        pcre.h.generic and config.h.generic by "make" instead of in the
697        PrepareRelease script.
698    
699    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
700        pcre_compile.c when checking for a zero character.
701    
702    28. Introducing a native interface for JIT. Through this interface, the compiled
703        machine code can be directly executed. The purpose of this interface is to
704        provide fast pattern matching, so several sanity checks are not performed.
705        However, feature tests are still performed. The new interface provides
706        1.4x speedup compared to the old one.
707    
708    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
709        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
710        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
711    
712    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
713        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
714        the "old" RFC 2279). Instead, it ended up passing a negative length to
715        pcre_exec().
716    
717    31. Add support for GCC's visibility feature to hide internal functions.
718    
719    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
720        "unknown -C option" after outputting 0 or 1.
721    
722    33. There is now support for generating a code coverage report for the test
723        suite in environments where gcc is the compiler and lcov is installed. This
724        is mainly for the benefit of the developers.
725    
726    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
727        unaddressable using valgrind annotations, allowing valgrind to detect
728        invalid memory accesses. This is mainly for the benefit of the developers.
729    
730    25. (*UTF) can now be used to start a pattern in any of the three libraries.
731    
732    26. Give configure error if --enable-cpp but no C++ compiler found.
733    
734    
735    Version 8.31 06-July-2012
736    -------------------------
737    
738    1.  Fixing a wrong JIT test case and some compiler warnings.
739    
740    2.  Removed a bashism from the RunTest script.
741    
742    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
743        to unsigned type, result still unsigned" that was given by an MS compiler
744        on encountering the code "-sizeof(xxx)".
745    
746    4.  Partial matching support is added to the JIT compiler.
747    
748    5.  Fixed several bugs concerned with partial matching of items that consist
749        of more than one character:
750    
751        (a) /^(..)\1/ did not partially match "aba" because checking references was
752            done on an "all or nothing" basis. This also applied to repeated
753            references.
754    
755        (b) \R did not give a hard partial match if \r was found at the end of the
756            subject.
757    
758        (c) \X did not give a hard partial match after matching one or more
759            characters at the end of the subject.
760    
761        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
762            a partial match for the string "\r".
763    
764        (e) When newline was set to CRLF, the metacharacter "." did not recognize
765            a partial match for a CR character at the end of the subject string.
766    
767    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
768        running pcretest, the text "(JIT)" added to the output whenever JIT is
769        actually used to run the match.
770    
771    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
772        or /S+[+] with a digit between 1 and 7.
773    
774    8.  OP_NOT now supports any UTF character not just single-byte ones.
775    
776    9.  (*MARK) control verb is now supported by the JIT compiler.
777    
778    10. The command "./RunTest list" lists the available tests without actually
779        running any of them. (Because I keep forgetting what they all are.)
780    
781    11. Add PCRE_INFO_MAXLOOKBEHIND.
782    
783    12. Applied a (slightly modified) user-supplied patch that improves performance
784        when the heap is used for recursion (compiled with --disable-stack-for-
785        recursion). Instead of malloc and free for each heap frame each time a
786        logical recursion happens, frames are retained on a chain and re-used where
787        possible. This sometimes gives as much as 30% improvement.
788    
789    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
790        call.
791    
792    14. As documented, (*COMMIT) is now confined to within a positive assertion.
793    
794    15. It is now possible to link pcretest with libedit as an alternative to
795        libreadline.
796    
797    16. (*COMMIT) control verb is now supported by the JIT compiler.
798    
799    17. The Unicode data tables have been updated to Unicode 6.1.0.
800    
801    18. Added --file-list option to pcregrep.
802    
803    19. Added binary file support to pcregrep, including the -a, --binary-files,
804        -I, and --text options.
805    
806    20. The madvise function is renamed for posix_madvise for QNX compatibility
807        reasons. Fixed by Giuseppe D'Angelo.
808    
809    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
810    
811    22. Changed the option for creating symbolic links for 16-bit man pages from
812        -s to -sf so that re-installing does not cause issues.
813    
814    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
815    
816    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
817        matches in certain environments (the workspace was not being correctly
818        retained). Also added to pcre_dfa_exec() a simple plausibility check on
819        some of the workspace data at the beginning of a restart.
820    
821    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
822        was not doing so when it should - probably a typo introduced by SVN 528
823        (change 8.10/14).
824    
825    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
826        \w+ when the character tables indicated that \x{c4} was a word character.
827        There were several related cases, all because the tests for doing a table
828        lookup were testing for characters less than 127 instead of 255.
829    
830    27. If a pattern contains capturing parentheses that are not used in a match,
831        their slots in the ovector are set to -1. For those that are higher than
832        any matched groups, this happens at the end of processing. In the case when
833        there were back references that the ovector was too small to contain
834        (causing temporary malloc'd memory to be used during matching), and the
835        highest capturing number was not used, memory off the end of the ovector
836        was incorrectly being set to -1. (It was using the size of the temporary
837        memory instead of the true size.)
838    
839    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
840        ovector size, it uses memory at the end of the block that it has got.
841    
842    29. Check for an overlong MARK name and give an error at compile time. The
843        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
844    
845    30. JIT compiler update.
846    
847    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
848        Rill for the patch.
849    
850    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
851    
852    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
853    
854    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
855        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
856        was enabled.
857    
858    35. Improve JIT code generation for greedy plus quantifier.
859    
860    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
861        1 to "aa" instead of to an empty string. The bug affected repeated groups
862        that could potentially match an empty string.
863    
864    37. Optimizing single character iterators in JIT.
865    
866    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
867        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
868        codepoints that are too big for the mode are faulted, and in a UTF mode,
869        disallowed codepoints are also faulted.
870    
871    39. If PCRE was compiled with UTF support, in three places in the DFA
872        matcher there was code that should only have been obeyed in UTF mode, but
873        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
874        processing when bytes with values greater than 127 were present. In 16-bit
875        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
876        both cases the values are those that cannot be the first data item in a UTF
877        character. The three items that might have provoked this were recursions,
878        possessively repeated groups, and atomic groups.
879    
880    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
881        and pcregrep, because some OS require shared objects to be explicitly
882        passed to ld, causing the link step to fail if they are not.
883    
884    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
885        mode, patterns that started with \h* or \R* might be incorrectly matched.
886    
887    
888    Version 8.30 04-February-2012
889    -----------------------------
890    
891    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
892        name is defined in ctype.h.
893    
894    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
895        only in quite long subpatterns.
896    
897    3.  Removed the function pcre_info(), which has been obsolete and deprecated
898        since it was replaced by pcre_fullinfo() in February 2000.
899    
900    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
901        match a (*MARK), and the match failed at the start of the subject, a
902        reference to memory before the start of the subject could occur. This bug
903        was introduced by fix 17 of release 8.21.
904    
905    5.  A reference to an unset group with zero minimum repetition was giving
906        totally wrong answers (in non-JavaScript-compatibility mode). For example,
907        /(another)?(\1?)test/ matched against "hello world test". This bug was
908        introduced in release 8.13.
909    
910    6.  Add support for 16-bit character strings (a large amount of work involving
911        many changes and refactorings).
912    
913    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
914        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
915        from a file.
916    
917    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
918        rounding is not applied in this particular case).
919    
920    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
921        if they appear, or are escaped, in patterns.
922    
923    10. Get rid of a number of -Wunused-but-set-variable warnings.
924    
925    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
926        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
927        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
928        also returns the mark "x". This bug applied to capturing parentheses,
929        non-capturing parentheses, and atomic parentheses. It also applied to some
930        assertions.
931    
932    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
933        information out of configure.ac instead of relying on pcre.h.generic, which
934        is not stored in the repository.
935    
936    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
937        -lreadline.
938    
939    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
940    
941    15. Applied Graycode's patch to put the top-level frame on the stack rather
942        than the heap when not using the stack for recursion. This gives a
943        performance improvement in many cases when recursion is not deep.
944    
945    16. Experimental code added to "pcretest -C" to output the stack frame size.
946    
947    
948    Version 8.21 12-Dec-2011
949    ------------------------
950    
951    1.  Updating the JIT compiler.
952    
953    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
954        are added as well.
955    
956    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
957        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
958        calling _pcre_jit_exec. Some extra comments are added.
959    
960    4.  (*MARK) settings inside atomic groups that do not contain any capturing
961        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
962        was introduced by change 18 for 8.20.
963    
964    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
965        ECMA-262 standard.
966    
967    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
968        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
969        This bug was probably introduced by change 9 of 8.13.
970    
971    7.  While fixing 6 above, I noticed that a number of other items were being
972        incorrectly rejected as "not fixed length". This arose partly because newer
973        opcodes had not been added to the fixed-length checking code. I have (a)
974        corrected the bug and added tests for these items, and (b) arranged for an
975        error to occur if an unknown opcode is encountered while checking for fixed
976        length instead of just assuming "not fixed length". The items that were
977        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
978        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
979        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
980    
981    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
982        being incorrectly compiled and would have given unpredicatble results.
983    
984    9.  A possessively repeated subpattern with minimum repeat count greater than
985        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
986        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
987        the first (A) could occur when it should not.
988    
989    10. Add a cast and remove a redundant test from the code.
990    
991    11. JIT should use pcre_malloc/pcre_free for allocation.
992    
993    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
994        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
995        is anything other than /usr, -L is still shown).
996    
997    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
998    
999    14. Perl does not support \N without a following name in a [] class; PCRE now
1000        also gives an error.
1001    
1002    15. If a forward reference was repeated with an upper limit of around 2000,
1003        it caused the error "internal error: overran compiling workspace". The
1004        maximum number of forward references (including repeats) was limited by the
1005        internal workspace, and dependent on the LINK_SIZE. The code has been
1006        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1007        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1008        of around 200,000 forward references. While doing this, I also speeded up
1009        the filling in of repeated forward references.
1010    
1011    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1012        incorrectly expecting the subject to contain another "a" after the start.
1013    
1014    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1015        in the match, the SKIP should be ignored. This was not happening; instead
1016        the SKIP was being treated as NOMATCH. For patterns such as
1017        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1018        tested.
1019    
1020    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1021        now much more compatible with Perl, in particular in cases where the result
1022        is a non-match for a non-anchored pattern. For example, if
1023        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1024        "m", where previously it did not return a name. A side effect of this
1025        change is that for partial matches, the last encountered mark name is
1026        returned, as for non matches. A number of tests that were previously not
1027        Perl-compatible have been moved into the Perl-compatible test files. The
1028        refactoring has had the pleasing side effect of removing one argument from
1029        the match() function, thus reducing its stack requirements.
1030    
1031    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1032        subsequent uses of /S (without +) incorrectly behaved like /S+.
1033    
1034    21. Retrieve executable code size support for the JIT compiler and fixing
1035        some warnings.
1036    
1037    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1038        not work when the shorter character appeared right at the end of the
1039        subject string.
1040    
1041    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1042        systems.
1043    
1044    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1045        output it when the /M option is used in pcretest.
1046    
1047    25. The CheckMan script was not being included in the distribution. Also, added
1048        an explicit "perl" to run Perl scripts from the PrepareRelease script
1049        because this is reportedly needed in Windows.
1050    
1051    26. If study data was being save in a file and studying had not found a set of
1052        "starts with" bytes for the pattern, the data written to the file (though
1053        never used) was taken from uninitialized memory and so caused valgrind to
1054        complain.
1055    
1056    27. Updated RunTest.bat as provided by Sheri Pierce.
1057    
1058    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1059    
1060    29. Computation of memory usage for the table of capturing group names was
1061        giving an unnecessarily large value.
1062    
1063    
1064    Version 8.20 21-Oct-2011
1065    ------------------------
1066    
1067    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1068        a POSIX class. After further experiments with Perl, which convinced me that
1069        Perl has bugs and confusions, a closing square bracket is no longer allowed
1070        in a POSIX name. This bug also affected patterns with classes that started
1071        with full stops.
1072    
1073    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1074        captured substring, but while checking the failing first alternative,
1075        substring 1 is temporarily captured. If the output vector supplied to
1076        pcre_exec() was not big enough for this capture, the yield of the function
1077        was still zero ("insufficient space for captured substrings"). This cannot
1078        be totally fixed without adding another stack variable, which seems a lot
1079        of expense for a edge case. However, I have improved the situation in cases
1080        such as /(a)(b)x|abc/ matched against "abc", where the return code
1081        indicates that fewer than the maximum number of slots in the ovector have
1082        been set.
1083    
1084    3.  Related to (2) above: when there are more back references in a pattern than
1085        slots in the output vector, pcre_exec() uses temporary memory during
1086        matching, and copies in the captures as far as possible afterwards. It was
1087        using the entire output vector, but this conflicts with the specification
1088        that only 2/3 is used for passing back captured substrings. Now it uses
1089        only the first 2/3, for compatibility. This is, of course, another edge
1090        case.
1091    
1092    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1093        main code base, and can be used by building with --enable-jit. When this is
1094        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1095        runtime --no-jit option is given.
1096    
1097    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1098        ovector, the return from the function was zero, implying that there were
1099        other matches that did not fit. The correct "exactly full" value is now
1100        returned.
1101    
1102    6.  If a subpattern that was called recursively or as a subroutine contained
1103        (*PRUNE) or any other control that caused it to give a non-standard return,
1104        invalid errors such as "Error -26 (nested recursion at the same subject
1105        position)" or even infinite loops could occur.
1106    
1107    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1108        computing the minimum length on reaching *ACCEPT, and so ended up with the
1109        wrong value of 1 rather than 0. Further investigation indicates that
1110        computing a minimum subject length in the presence of *ACCEPT is difficult
1111        (think back references, subroutine calls), and so I have changed the code
1112        so that no minimum is registered for a pattern that contains *ACCEPT.
1113    
1114    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1115        it was not handled as intended. [But see 16 below.]
1116    
1117    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1118        Sheri Pierce.
1119    
1120    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1121        the first byte in a match must be "a".
1122    
1123    11. Change 17 for 8.13 increased the recursion depth for patterns like
1124        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1125        pattern contains any instances of (*THEN). If it does not, the old
1126        optimizations are restored. It would be nice to do this on a per-group
1127        basis, but at the moment that is not feasible.
1128    
1129    12. In some environments, the output of pcretest -C is CRLF terminated. This
1130        broke RunTest's code that checks for the link size. A single white space
1131        character after the value is now allowed for.
1132    
1133    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1134        For "fr", it uses the Windows-specific input and output files.
1135    
1136    14. If (*THEN) appeared in a group that was called recursively or as a
1137        subroutine, it did not work as intended. [But see next item.]
1138    
1139    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1140        pattern fragments (but not containing any | characters). If A and B are
1141        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1142        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1143        D. In other words, Perl considers parentheses that do not contain any |
1144        characters to be part of a surrounding alternative, whereas PCRE was
1145        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1146        differently. PCRE now behaves in the same way as Perl, except in the case
1147        of subroutine/recursion calls such as (?1) which have in any case always
1148        been different (but PCRE had them first :-).
1149    
1150    16. Related to 15 above: Perl does not treat the | in a conditional group as
1151        creating alternatives. Such a group is treated in the same way as an
1152        ordinary group without any | characters when processing (*THEN). PCRE has
1153        been changed to match Perl's behaviour.
1154    
1155    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1156        RunGrepTest script failed.
1157    
1158    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1159        inevitable for groups that contain captures, but it can lead to a lot of
1160        stack use in large patterns. The old behaviour has been restored for atomic
1161        groups that do not contain any capturing parentheses.
1162    
1163    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1164        suppress the check for a minimum subject length at run time. (If it was
1165        given to pcre_exec() or pcre_dfa_exec() it did work.)
1166    
1167    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1168        fail to work when decoding hex characters in data strings in EBCDIC
1169        environments.
1170    
1171    21. It appears that in at least one Mac OS environment, the isxdigit() function
1172        is implemented as a macro that evaluates to its argument more than once,
1173        contravening the C 90 Standard (I haven't checked a later standard). There
1174        was an instance in pcretest which caused it to go wrong when processing
1175        \x{...} escapes in subject strings. The has been rewritten to avoid using
1176        things like p++ in the argument of isxdigit().
1177    
1178    
1179    Version 8.13 16-Aug-2011
1180    ------------------------
1181    
1182    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1183    
1184    2.  Two minor typos in pcre_internal.h have been fixed.
1185    
1186    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1187        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1188        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1189    
1190    4.  There were a number of related bugs in the code for matching backrefences
1191        caselessly in UTF-8 mode when codes for the characters concerned were
1192        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1193        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1194        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1195        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1196        2-byte code at the end of the subject (it thought there wasn't enough data
1197        left).
1198    
1199    5.  Comprehensive information about what went wrong is now returned by
1200        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1201        as the output vector has at least 2 elements. The offset of the start of
1202        the failing character and a reason code are placed in the vector.
1203    
1204    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1205        now returned is for the first byte of the failing character, instead of the
1206        last byte inspected. This is an incompatible change, but I hope it is small
1207        enough not to be a problem. It makes the returned offset consistent with
1208        pcre_exec() and pcre_dfa_exec().
1209    
1210    7.  pcretest now gives a text phrase as well as the error number when
1211        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1212        failure, the offset and reason code are output.
1213    
1214    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1215        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1216        back over a single character (\n). This seems wrong (because it treated the
1217        two characters as a single entity when going forwards), conflicts with the
1218        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1219        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1220        has been changed.
1221    
1222    9.  Some internal refactoring has changed the processing so that the handling
1223        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1224        time (the PCRE_DOTALL option was changed this way some time ago: version
1225        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1226        which was always a bit of a fudge. It also means that there is one less
1227        argument for the match() function, which reduces its stack requirements
1228        slightly. This change also fixes an incompatibility with Perl: the pattern
1229        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1230    
1231    10. More internal refactoring has drastically reduced the number of recursive
1232        calls to match() for possessively repeated groups such as (abc)++ when
1233        using pcre_exec().
1234    
1235    11. While implementing 10, a number of bugs in the handling of groups were
1236        discovered and fixed:
1237    
1238        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1239        (a|)*(?1) gave a compile-time internal error.
1240        ((a|)+)+  did not notice that the outer group could match an empty string.
1241        (^a|^)+   was not marked as anchored.
1242        (.*a|.*)+ was not marked as matching at start or after a newline.
1243    
1244    12. Yet more internal refactoring has removed another argument from the match()
1245        function. Special calls to this function are now indicated by setting a
1246        value in a variable in the "match data" data block.
1247    
1248    13. Be more explicit in pcre_study() instead of relying on "default" for
1249        opcodes that mean there is no starting character; this means that when new
1250        ones are added and accidentally left out of pcre_study(), testing should
1251        pick them up.
1252    
1253    14. The -s option of pcretest has been documented for ages as being an old
1254        synonym of -m (show memory usage). I have changed it to mean "force study
1255        for every regex", that is, assume /S for every regex. This is similar to -i
1256        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1257        using it. It makes it easier to run collections of tests with and without
1258        study enabled, and thereby test pcre_study() more easily. All the standard
1259        tests are now run with and without -s (but some patterns can be marked as
1260        "never study" - see 20 below).
1261    
1262    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1263        restoration of the capturing data to the outer values was not happening
1264        correctly.
1265    
1266    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1267        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1268        pattern had matched an empty string, and so incorrectly returned a no
1269        match.
1270    
1271    17. There was optimizing code for the last branch of non-capturing parentheses,
1272        and also for the obeyed branch of a conditional subexpression, which used
1273        tail recursion to cut down on stack usage. Unfortunately, now that there is
1274        the possibility of (*THEN) occurring in these branches, tail recursion is
1275        no longer possible because the return has to be checked for (*THEN). These
1276        two optimizations have therefore been removed. [But see 8.20/11 above.]
1277    
1278    18. If a pattern containing \R was studied, it was assumed that \R always
1279        matched two bytes, thus causing the minimum subject length to be
1280        incorrectly computed because \R can also match just one byte.
1281    
1282    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1283        was incorrectly computed.
1284    
1285    20. If /S is present twice on a test pattern in pcretest input, it now
1286        *disables* studying, thereby overriding the use of -s on the command line
1287        (see 14 above). This is necessary for one or two tests to keep the output
1288        identical in both cases.
1289    
1290    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1291        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1292    
1293    22. When an atomic group that contained a capturing parenthesis was
1294        successfully matched, but the branch in which it appeared failed, the
1295        capturing was not being forgotten if a higher numbered group was later
1296        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1297        group 1 to "a", when in fact it should be unset. This applied to multi-
1298        branched capturing and non-capturing groups, repeated or not, and also to
1299        positive assertions (capturing in negative assertions does not happen
1300        in PCRE) and also to nested atomic groups.
1301    
1302    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1303        subject after a captured substring, to make it easier to tell which of a
1304        number of identical substrings has been captured.
1305    
1306    24. The way atomic groups are processed by pcre_exec() has been changed so that
1307        if they are repeated, backtracking one repetition now resets captured
1308        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1309        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1310        "aaa". Previously, it would have been "a". As part of this code
1311        refactoring, the way recursive calls are handled has also been changed.
1312    
1313    25. If an assertion condition captured any substrings, they were not passed
1314        back unless some other capturing happened later. For example, if
1315        (?(?=(a))a) was matched against "a", no capturing was returned.
1316    
1317    26. When studying a pattern that contained subroutine calls or assertions,
1318        the code for finding the minimum length of a possible match was handling
1319        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1320        group 1 called group 2 while simultaneously a separate group 2 called group
1321        1). A stack overflow occurred in this case. I have fixed this by limiting
1322        the recursion depth to 10.
1323    
1324    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1325        Fortmann. This supports explicit test numbers on the command line, and has
1326        argument validation and error reporting.
1327    
1328    28. An instance of \X with an unlimited repeat could fail if at any point the
1329        first character it looked at was a mark character.
1330    
1331    29. Some minor code refactoring concerning Unicode properties and scripts
1332        should reduce the stack requirement of match() slightly.
1333    
1334    30. Added the '=' option to pcretest to check the setting of unused capturing
1335        slots at the end of the pattern, which are documented as being -1, but are
1336        not included in the return count.
1337    
1338    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1339        compiled something random. Now it gives a compile-time error (as does
1340        Perl).
1341    
1342    32. A *MARK encountered during the processing of a positive assertion is now
1343        recorded and passed back (compatible with Perl).
1344    
1345    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1346        had alternative anchored branches, the search for a second match in a line
1347        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1348        matched the line "0102" twice. The same bug affected patterns that started
1349        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1350        twice.
1351    
1352    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1353        does, and because of capturing effects, quantifying parenthesized
1354        assertions may at times be useful. Quantifiers are now allowed for
1355        parenthesized assertions.
1356    
1357    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1358    
1359    36. \g was being checked for fancy things in a character class, when it should
1360        just be a literal "g".
1361    
1362    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1363        appearance of a nested POSIX class supersedes an apparent external class.
1364        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1365        unescaped square brackets may also appear as part of class names. For
1366        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1367        more like Perl. (But see 8.20/1 above.)
1368    
1369    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1370        was because it thought it was \N{name}, which is not supported).
1371    
1372    39. Add minix to OS list not supporting the -S option in pcretest.
1373    
1374    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1375        cannot analyze patterns in sufficient detail to catch mutual recursions
1376        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1377        subgroup is called recursively as a subpattern for a second time at the
1378        same position in the subject string. In previous releases this might have
1379        been caught by the recursion limit, or it might have run out of stack.
1380    
1381    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1382        happen only once. PCRE was, however incorrectly giving a compile time error
1383        "recursive call could loop indefinitely" because it cannot analyze the
1384        pattern in sufficient detail. The compile time test no longer happens when
1385        PCRE is compiling a conditional subpattern, but actual runaway loops are
1386        now caught at runtime (see 40 above).
1387    
1388    42. It seems that Perl allows any characters other than a closing parenthesis
1389        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1390        has been changed to be the same.
1391    
1392    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1393        as not to get warnings when autogen.sh is called. Also changed
1394        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1395    
1396    44. To help people who use pcregrep to scan files containing exceedingly long
1397        lines, the following changes have been made:
1398    
1399        (a) The default value of the buffer size parameter has been increased from
1400            8K to 20K. (The actual buffer used is three times this size.)
1401    
1402        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1403            PCRE is built.
1404    
1405        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1406            to be set at run time.
1407    
1408        (d) Numerical values in pcregrep options can be followed by K or M, for
1409            example --buffer-size=50K.
1410    
1411        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1412            given and the return code is set to 2.
1413    
1414    45. Add a pointer to the latest mark to the callout data block.
1415    
1416    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1417        partial match of an empty string instead of no match. This was specific to
1418        the use of ".".
1419    
1420    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1421        complete match instead of a partial match. This bug was dependent on both
1422        the PCRE_UTF8 and PCRE_DOTALL options being set.
1423    
1424    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1425        starting byte set, because \b was not being ignored.
1426    
1427    
1428    Version 8.12 15-Jan-2011
1429    ------------------------
1430    
1431    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1432        checks for such things as part of the documentation building process.
1433    
1434    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1435        --match-limit and --recursion-limit options (added for 8.11). In
1436        particular, this made one of the standard tests fail. (The integer value
1437        went into the wrong half of a long int.)
1438    
1439    3.  If the --colour option was given to pcregrep with -v (invert match), it
1440        did strange things, either producing crazy output, or crashing. It should,
1441        of course, ignore a request for colour when reporting lines that do not
1442        match.
1443    
1444    4.  Another pcregrep bug caused similar problems if --colour was specified with
1445        -M (multiline) and the pattern match finished with a line ending.
1446    
1447    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1448        matched in multiline mode, the following line was shown as part of the
1449        match. This seems wrong, so I have changed it.
1450    
1451    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1452        the check for further matches in the same line (so they could be coloured)
1453        to overrun the end of the current line. If another match was found, it was
1454        incorrectly shown (and then shown again when found in the next line).
1455    
1456    7.  If pcregrep was compiled under Windows, there was a reference to the
1457        function pcregrep_exit() before it was defined. I am assuming this was
1458        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1459        reported by a user. I've moved the definition above the reference.
1460    
1461    
1462    Version 8.11 10-Dec-2010
1463    ------------------------
1464    
1465    1.  (*THEN) was not working properly if there were untried alternatives prior
1466        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1467        backtracked to try for "b" instead of moving to the next alternative branch
1468        at the same level (in this case, to look for "c"). The Perl documentation
1469        is clear that when (*THEN) is backtracked onto, it goes to the "next
1470        alternative in the innermost enclosing group".
1471    
1472    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1473        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1474        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1475        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1476        (*THEN).
1477    
1478    3.  If \s appeared in a character class, it removed the VT character from
1479        the class, even if it had been included by some previous item, for example
1480        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1481        of \s, but is part of the POSIX "space" class.)
1482    
1483    4.  A partial match never returns an empty string (because you can always
1484        match an empty string at the end of the subject); however the checking for
1485        an empty string was starting at the "start of match" point. This has been
1486        changed to the "earliest inspected character" point, because the returned
1487        data for a partial match starts at this character. This means that, for
1488        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1489        (previously it gave "no match").
1490    
1491    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1492        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1493        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1494        has an implication that the given string is incomplete (because a partial
1495        match is preferred over a full match). For this reason, these items now
1496        give a partial match in this situation. [Aside: previously, the one case
1497        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1498        match rather than a full match, which was wrong by the old rules, but is
1499        now correct.]
1500    
1501    6.  There was a bug in the handling of #-introduced comments, recognized when
1502        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1503        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1504        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1505        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1506        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1507        places in pcre_compile().
1508    
1509    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1510        comments when looking ahead for named forward references to subpatterns,
1511        the only newline sequence it recognized was NL. It now handles newlines
1512        according to the set newline convention.
1513    
1514    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1515        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1516        cater for a lack of strerror(). These oversights have been fixed.
1517    
1518    9.  Added --match-limit and --recursion-limit to pcregrep.
1519    
1520    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1521    
1522    11. When the -o option was used, pcregrep was setting a return code of 1, even
1523        when matches were found, and --line-buffered was not being honoured.
1524    
1525    12. Added an optional parentheses number to the -o and --only-matching options
1526        of pcregrep.
1527    
1528    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1529        can match an empty string. The code to do it in pcretest and pcredemo
1530        needed fixing:
1531    
1532        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1533            only one byte after an empty string match just before CRLF (this case
1534            just got forgotten; "any" and "anycrlf" were OK).
1535    
1536        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1537            mode when an empty string match preceded an ASCII character followed by
1538            a non-ASCII character. (The code for advancing by one character rather
1539            than one byte was nonsense.)
1540    
1541        (c) The pcredemo.c sample program did not have any code at all to handle
1542            the cases when CRLF is a valid newline sequence.
1543    
1544    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1545        as a starting offset was within the subject string. There is now a new
1546        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1547        negative or greater than the length of the string. In order to test this,
1548        pcretest is extended to allow the setting of negative starting offsets.
1549    
1550    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1551        starting offset points to the beginning of a UTF-8 character was
1552        unnecessarily clumsy. I tidied it up.
1553    
1554    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1555        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1556    
1557    17. Nobody had reported that the --include_dir option, which was added in
1558        release 7.7 should have been called --include-dir (hyphen, not underscore)
1559        for compatibility with GNU grep. I have changed it to --include-dir, but
1560        left --include_dir as an undocumented synonym, and the same for
1561        --exclude-dir, though that is not available in GNU grep, at least as of
1562        release 2.5.4.
1563    
1564    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1565        characters from a string of bytes) have been redefined so as not to use
1566        loops, in order to improve performance in some environments. At the same
1567        time, I abstracted some of the common code into auxiliary macros to save
1568        repetition (this should not affect the compiled code).
1569    
1570    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1571        compile-time error is now given if \c is not followed by an ASCII
1572        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1573        different, and any byte value is allowed.)
1574    
1575    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1576        START_OPTIMIZE option, which is now allowed at compile time - but just
1577        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1578        to pcregrep and other applications that have no direct access to PCRE
1579        options. The new /Y option in pcretest sets this option when calling
1580        pcre_compile().
1581    
1582    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1583        back references. Groups containing recursive back references were forced to
1584        be atomic by that change, but in the case of named groups, the amount of
1585        memory required was incorrectly computed, leading to "Failed: internal
1586        error: code overflow". This has been fixed.
1587    
1588    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1589        pcretest.c, to avoid build problems in some Borland environments.
1590    
1591    
1592    Version 8.10 25-Jun-2010
1593    ------------------------
1594    
1595    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1596        THEN.
1597    
1598    2.  (*ACCEPT) was not working when inside an atomic group.
1599    
1600    3.  Inside a character class, \B is treated as a literal by default, but
1601        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1602        causes the error). The code is unchanged, but I tidied the documentation.
1603    
1604    4.  Inside a character class, PCRE always treated \R and \X as literals,
1605        whereas Perl faults them if its -w option is set. I have changed PCRE so
1606        that it faults them when PCRE_EXTRA is set.
1607    
1608    5.  Added support for \N, which always matches any character other than
1609        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1610    
1611    6.  When compiling pcregrep with newer versions of gcc which may have
1612        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1613        declared with attribute warn_unused_result" were given. Just casting the
1614        result to (void) does not stop the warnings; a more elaborate fudge is
1615        needed. I've used a macro to implement this.
1616    
1617    7.  Minor change to pcretest.c to avoid a compiler warning.
1618    
1619    8.  Added four artifical Unicode properties to help with an option to make
1620        \s etc use properties (see next item). The new properties are: Xan
1621        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1622    
1623    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1624        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1625        this option. Modified pcretest to add /W to test this facility. Added
1626        REG_UCP to make it available via the POSIX interface.
1627    
1628    10. Added --line-buffered to pcregrep.
1629    
1630    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1631        studied, and the match started with a letter with a code point greater than
1632        127 whose first byte was different to the first byte of the other case of
1633        the letter, the other case of this starting letter was not recognized
1634        (#976).
1635    
1636    12. If a pattern that was studied started with a repeated Unicode property
1637        test, for example, \p{Nd}+, there was the theoretical possibility of
1638        setting up an incorrect bitmap of starting bytes, but fortunately it could
1639        not have actually happened in practice until change 8 above was made (it
1640        added property types that matched character-matching opcodes).
1641    
1642    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1643        possible starting bytes for non-anchored patterns.
1644    
1645    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1646        \R, and also a number of cases that involve Unicode properties, both
1647        explicit and implicit when PCRE_UCP is set.
1648    
1649    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1650        input, it could crash or give wrong results if characters with values
1651        greater than 0xc0 were present in the subject string. (Detail: it assumed
1652        UTF-8 input when processing these items.)
1653    
1654    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1655        size_t is 64-bit (#991).
1656    
1657    17. Added a check for running out of memory when PCRE is compiled with
1658        --disable-stack-for-recursion (#990).
1659    
1660    18. If the last data line in a file for pcretest does not have a newline on
1661        the end, a newline was missing in the output.
1662    
1663    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1664        less than 128) in its various bitmaps. However, there is a facility for
1665        generating tables according to the current locale when PCRE is compiled. It
1666        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1667        characters, are recognized by isspace() and therefore were getting set in
1668        these tables, and indeed these tables seem to approximate to ISO 8859. This
1669        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1670        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1671        which of course cannot start UTF-8 characters. I have changed the code so
1672        that only real ASCII characters (less than 128) and the correct starting
1673        bytes for UTF-8 encodings are set for characters greater than 127 when in
1674        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1675        altogether.)
1676    
1677    20. Added the /T option to pcretest so as to be able to run tests with non-
1678        standard character tables, thus making it possible to include the tests
1679        used for 19 above in the standard set of tests.
1680    
1681    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1682        reference to a subpattern the other side of a comment that contains an
1683        opening parenthesis caused either an internal compiling error, or a
1684        reference to the wrong subpattern.
1685    
1686    
1687    Version 8.02 19-Mar-2010
1688    ------------------------
1689    
1690    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1691    
1692    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1693        configured.
1694    
1695    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1696        original author of that file, following a query about its status.
1697    
1698    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1699        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1700    
1701    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1702        quantifier applied to a forward-referencing subroutine call, could compile
1703        incorrect code or give the error "internal error: previously-checked
1704        referenced subpattern not found".
1705    
1706    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1707        variables to point to external functions. For these systems, therefore,
1708        pcre_malloc etc. are now initialized to local functions that call the
1709        relevant global functions.
1710    
1711    7.  There were two entries missing in the vectors called coptable and poptable
1712        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1713        I've fixed the data, and added a kludgy way of testing at compile time that
1714        the lengths are correct (equal to the number of opcodes).
1715    
1716    8.  Following on from 7, I added a similar kludge to check the length of the
1717        eint vector in pcreposix.c.
1718    
1719    9.  Error texts for pcre_compile() are held as one long string to avoid too
1720        much relocation at load time. To find a text, the string is searched,
1721        counting zeros. There was no check for running off the end of the string,
1722        which could happen if a new error number was added without updating the
1723        string.
1724    
1725    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1726    
1727    11. \K was not working if it appeared in an atomic group or in a group that
1728        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1729        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1730        the assertion is positive, but not if it is negative.
1731    
1732    12. Change 11 fortuitously reduced the size of the stack frame used in the
1733        "match()" function of pcre_exec.c by one pointer. Forthcoming
1734        implementation of support for (*MARK) will need an extra pointer on the
1735        stack; I have reserved it now, so that the stack frame size does not
1736        decrease.
1737    
1738    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1739        item in branch that calls a recursion is a subroutine call - as in the
1740        second branch in the above example - was incorrectly given the compile-
1741        time error "recursive call could loop indefinitely" because pcre_compile()
1742        was not correctly checking the subroutine for matching a non-empty string.
1743    
1744    14. The checks for overrunning compiling workspace could trigger after an
1745        overrun had occurred. This is a "should never occur" error, but it can be
1746        triggered by pathological patterns such as hundreds of nested parentheses.
1747        The checks now trigger 100 bytes before the end of the workspace.
1748    
1749    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1750    
1751    
1752    Version 8.01 19-Jan-2010
1753    ------------------------
1754    
1755    1.  If a pattern contained a conditional subpattern with only one branch (in
1756        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1757        computed the wrong minimum data length (which is of course zero for such
1758        subpatterns). This could cause incorrect "no match" results.
1759    
1760    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1761        the pattern is reset in the first branch, pcre_compile() failed with
1762        "internal error: code overflow at offset...". This happened only when
1763        the reset was to the original external option setting. (An optimization
1764        abstracts leading options settings into an external setting, which was the
1765        cause of this.)
1766    
1767    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1768        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1769        assertion pattern did not match (meaning that the assertion was true), it
1770        was incorrectly treated as false if the SKIP had been reached during the
1771        matching. This also applied to assertions used as conditions.
1772    
1773    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1774        assertion subpattern, including such a pattern used as a condition,
1775        unpredictable results occurred, instead of the error return
1776        PCRE_ERROR_DFA_UITEM.
1777    
1778    5.  The C++ GlobalReplace function was not working like Perl for the special
1779        situation when an empty string is matched. It now does the fancy magic
1780        stuff that is necessary.
1781    
1782    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1783        removed. (These were left over from very, very early versions of PCRE.)
1784    
1785    7.  Some cosmetic changes to the code to make life easier when compiling it
1786        as part of something else:
1787    
1788        (a) Change DEBUG to PCRE_DEBUG.
1789    
1790        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1791            called "current" as "current_branch", to prevent a collision with the
1792            Linux macro when compiled as a kernel module.
1793    
1794        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1795            prevent a collision with the Linux macro when compiled as a kernel
1796            module.
1797    
1798    8.  In pcre_compile() there are some checks for integer overflows that used to
1799        cast potentially large values to (double). This has been changed to that
1800        when building, a check for int64_t is made, and if it is found, it is used
1801        instead, thus avoiding the use of floating point arithmetic. (There is no
1802        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1803        double.
1804    
1805    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1806        2005 (difference between two addresses compared to an unsigned value).
1807    
1808    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1809        custom one, because of the following reported problem in Windows:
1810    
1811          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1812              under Win32.
1813          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1814              therefore missing the function definition.
1815          - The compiler thus generates a "C" signature for the test function.
1816          - The linker fails to find the "C" function.
1817          - PCRE fails to configure if asked to do so against libbz2.
1818    
1819    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1820        messages were output:
1821    
1822          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1823          rerunning libtoolize, to keep the correct libtool macros in-tree.
1824          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1825    
1826        I have done both of these things.
1827    
1828    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1829        most of the time, it *can* run out if it is given a pattern that contains a
1830        runaway infinite recursion. I updated the discussion in the pcrestack man
1831        page.
1832    
1833    13. Now that we have gone to the x.xx style of version numbers, the minor
1834        version may start with zero. Using 08 or 09 is a bad idea because users
1835        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1836        interpreted as invalid octal numbers. I've updated the previous comment in
1837        configure.ac, and also added a check that gives an error if 08 or 09 are
1838        used.
1839    
1840    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1841        causing partial matching to fail when the end of the subject matched \W
1842        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1843    
1844    15. There were some discrepancies between the declarations in pcre_internal.h
1845        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1846        their definitions. The declarations used "const uschar *" and the
1847        definitions used USPTR. Even though USPTR is normally defined as "const
1848        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1849        reported that: "This difference in casting confuses some C++ compilers, for
1850        example, SunCC recognizes above declarations as different functions and
1851        generates broken code for hbpcre." I have changed the declarations to use
1852        USPTR.
1853    
1854    16. GNU libtool is named differently on some systems. The autogen.sh script now
1855        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1856        (FreeBSD).
1857    
1858    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1859        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1860        comment: "Figure out how to create a longlong from a string: strtoll and
1861        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1862        instance, but it only takes 2 args instead of 3!"
1863    
1864    18. A subtle bug concerned with back references has been fixed by a change of
1865        specification, with a corresponding code fix. A pattern such as
1866        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1867        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1868        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1869        same bug. Such groups have to be quantified to be useful, or contained
1870        inside another quantified group. (If there's no repetition, the reference
1871        can never match.) The problem arises because, having left the group and
1872        moved on to the rest of the pattern, a later failure that backtracks into
1873        the group uses the captured value from the final iteration of the group
1874        rather than the correct earlier one. I have fixed this in PCRE by forcing
1875        any group that contains a reference to itself to be an atomic group; that
1876        is, there cannot be any backtracking into it once it has completed. This is
1877        similar to recursive and subroutine calls.
1878    
1879    
1880    Version 8.00 19-Oct-09
1881    ----------------------
1882    
1883    1.  The table for translating pcre_compile() error codes into POSIX error codes
1884        was out-of-date, and there was no check on the pcre_compile() error code
1885        being within the table. This could lead to an OK return being given in
1886        error.
1887    
1888    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1889        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1890        in a Windows environment.
1891    
1892    3.  The pcregrep --count option prints the count for each file even when it is
1893        zero, as does GNU grep. However, pcregrep was also printing all files when
1894        --files-with-matches was added. Now, when both options are given, it prints
1895        counts only for those files that have at least one match. (GNU grep just
1896        prints the file name in this circumstance, but including the count seems
1897        more useful - otherwise, why use --count?) Also ensured that the
1898        combination -clh just lists non-zero counts, with no names.
1899    
1900    4.  The long form of the pcregrep -F option was incorrectly implemented as
1901        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1902        but it seems right to fix it, and I didn't think it was worth preserving
1903        the old behaviour.
1904    
1905    5.  The command line items --regex=pattern and --regexp=pattern were not
1906        recognized by pcregrep, which required --regex pattern or --regexp pattern
1907        (with a space rather than an '='). The man page documented the '=' forms,
1908        which are compatible with GNU grep; these now work.
1909    
1910    6.  No libpcreposix.pc file was created for pkg-config; there was just
1911        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1912    
1913    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1914        when UCP support is not needed, by modifying the Python script that
1915        generates it from Unicode data files. This should not matter if the module
1916        is correctly used as a library, but I received one complaint about 50K of
1917        unwanted data. My guess is that the person linked everything into his
1918        program rather than using a library. Anyway, it does no harm.
1919    
1920    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1921        was a minimum greater than 1 for a wide character in a possessive
1922        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1923        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1924        character. Chaos in the form of incorrect output or a compiling loop could
1925        result.
1926    
1927    9.  The restrictions on what a pattern can contain when partial matching is
1928        requested for pcre_exec() have been removed. All patterns can now be
1929        partially matched by this function. In addition, if there are at least two
1930        slots in the offset vector, the offset of the earliest inspected character
1931        for the match and the offset of the end of the subject are set in them when
1932        PCRE_ERROR_PARTIAL is returned.
1933    
1934    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1935        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1936        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1937        and may be more useful for multi-segment matching.
1938    
1939    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1940        used to be given if ever the end of the subject was reached; now it is
1941        given only if matching could not proceed because another character was
1942        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1943        string "Z", which now yields "no match" instead of "partial match". In the
1944        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1945        final character ended with (*FAIL).
1946    
1947    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1948        if the pattern had a "must contain" character that was already found in the
1949        earlier partial match, unless partial matching was again requested. For
1950        example, with the pattern /dog.(body)?/, the "must contain" character is
1951        "g". If the first part-match was for the string "dog", restarting with
1952        "sbody" failed. This bug has been fixed.
1953    
1954    13. The string returned by pcre_dfa_exec() after a partial match has been
1955        changed so that it starts at the first inspected character rather than the
1956        first character of the match. This makes a difference only if the pattern
1957        starts with a lookbehind assertion or \b or \B (\K is not supported by
1958        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1959        matching functions compatible, and I think it's the right thing to do.
1960    
1961    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1962        so that the demonstration program is easily available in environments where
1963        PCRE has not been installed from source.
1964    
1965    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1966        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1967        library.
1968    
1969    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1970        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1971        is not the first non-POSIX option to be added. Clearly some people find
1972        these options useful.
1973    
1974    17. If a caller to the POSIX matching function regexec() passes a non-zero
1975        value for nmatch with a NULL value for pmatch, the value of
1976        nmatch is forced to zero.
1977    
1978    18. RunGrepTest did not have a test for the availability of the -u option of
1979        the diff command, as RunTest does. It now checks in the same way as
1980        RunTest, and also checks for the -b option.
1981    
1982    19. If an odd number of negated classes containing just a single character
1983        interposed, within parentheses, between a forward reference to a named
1984        subpattern and the definition of the subpattern, compilation crashed with
1985        an internal error, complaining that it could not find the referenced
1986        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1987        [The bug was that it was starting one character too far in when skipping
1988        over the character class, thus treating the ] as data rather than
1989        terminating the class. This meant it could skip too much.]
1990    
1991    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1992        /g option in pcretest when the pattern contains \K, which makes it possible
1993        to have an empty string match not at the start, even when the pattern is
1994        anchored. Updated pcretest and pcredemo to use this option.
1995    
1996    21. If the maximum number of capturing subpatterns in a recursion was greater
1997        than the maximum at the outer level, the higher number was returned, but
1998        with unset values at the outer level. The correct (outer level) value is
1999        now given.
2000    
2001    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2002        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2003        make it do so. The string so far is captured, making this feature
2004        compatible with Perl.
2005    
2006    23. The tests have been re-organized, adding tests 11 and 12, to make it
2007        possible to check the Perl 5.10 features against Perl 5.10.
2008    
2009    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2010        pattern matches a fixed length string. PCRE did not allow this; now it
2011        does. Neither allows recursion.
2012    
2013    25. I finally figured out how to implement a request to provide the minimum
2014        length of subject string that was needed in order to match a given pattern.
2015        (It was back references and recursion that I had previously got hung up
2016        on.) This code has now been added to pcre_study(); it finds a lower bound
2017        to the length of subject needed. It is not necessarily the greatest lower
2018        bound, but using it to avoid searching strings that are too short does give
2019        some useful speed-ups. The value is available to calling programs via
2020        pcre_fullinfo().
2021    
2022    26. While implementing 25, I discovered to my embarrassment that pcretest had
2023        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2024        study optimizations had never been tested with that matching function.
2025        Oops. What is worse, even when it was passed study data, there was a bug in
2026        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2027        were also very few tests of studied patterns with pcre_dfa_exec().
2028    
2029    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2030        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2031        on the other side of the coin, they are no longer allowed to have different
2032        names, because these cannot be distinguished in PCRE, and this has caused
2033        confusion. (This is a difference from Perl.)
2034    
2035    28. When duplicate subpattern names are present (necessarily with different
2036        numbers, as required by 27 above), and a test is made by name in a
2037        conditional pattern, either for a subpattern having been matched, or for
2038        recursion in such a pattern, all the associated numbered subpatterns are
2039        tested, and the overall condition is true if the condition is true for any
2040        one of them. This is the way Perl works, and is also more like the way
2041        testing by number works.
2042    
2043    
2044    Version 7.9 11-Apr-09
2045    ---------------------
2046    
2047    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2048        (pcretest), all targets were linked against these libraries. This included
2049        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2050        libraries. This caused unwanted dependencies to be created. This problem
2051        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2052        pcretest is linked with readline.
2053    
2054    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2055        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2056        moved outside it again, because FALSE and TRUE are already defined in AIX,
2057        but BOOL is not.
2058    
2059    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2060        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2061    
2062    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2063        hyphens) following file names and line numbers when outputting matching
2064        lines. This is not true; no spaces are inserted. I have also clarified the
2065        wording for the --colour (or --color) option.
2066    
2067    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2068        was not coloured; this is different to GNU grep, so I have changed it to be
2069        the same.
2070    
2071    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2072        each matching line was coloured. Now it goes on to look for further matches
2073        of any of the test patterns, which is the same behaviour as GNU grep.
2074    
2075    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2076        doesn't make sense to accept an empty string match in pcregrep, so I have
2077        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2078        seems to be how GNU grep behaves. [But see later change 40 for release
2079        8.33.]
2080    
2081    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2082        start or after a newline", because the conditional assertion was not being
2083        correctly handled. The rule now is that both the assertion and what follows
2084        in the first alternative must satisfy the test.
2085    
2086    9.  If auto-callout was enabled in a pattern with a conditional group whose
2087        condition was an assertion, PCRE could crash during matching, both with
2088        pcre_exec() and pcre_dfa_exec().
2089    
2090    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2091        used for matching.
2092    
2093    11. Unicode property support in character classes was not working for
2094        characters (bytes) greater than 127 when not in UTF-8 mode.
2095    
2096    12. Added the -M command line option to pcretest.
2097    
2098    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2099    
2100    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2101    
2102    16. Added comments and documentation about mis-use of no_arg in the C++
2103        wrapper.
2104    
2105    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2106        from Martin Jerabek that uses macro names for all relevant character and
2107        string constants.
2108    
2109    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2110        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2111        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2112        these, but not everybody uses configure.
2113    
2114    19. A conditional group that had only one branch was not being correctly
2115        recognized as an item that could match an empty string. This meant that an
2116        enclosing group might also not be so recognized, causing infinite looping
2117        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2118        with the subject "ab", where knowledge that the repeated group can match
2119        nothing is needed in order to break the loop.
2120    
2121    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2122        exec(), but without supplying a callout function, matching went wrong.
2123    
2124    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2125        leak if the size of the offset vector was greater than 30. When the vector
2126        is smaller, the saved offsets during recursion go onto a local stack
2127        vector, but for larger vectors malloc() is used. It was failing to free
2128        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2129        error, in fact).
2130    
2131    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2132        heapframe that is used only when UTF-8 support is enabled. This caused no
2133        problem, but was untidy.
2134    
2135    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2136        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2137        included within another project.
2138    
2139    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2140        slightly modified by me:
2141    
2142          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2143              not building pcregrep.
2144    
2145          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2146              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2147    
2148    25. Forward references, both numeric and by name, in patterns that made use of
2149        duplicate group numbers, could behave incorrectly or give incorrect errors,
2150        because when scanning forward to find the reference group, PCRE was not
2151        taking into account the duplicate group numbers. A pattern such as
2152        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2153    
2154    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2155        the feature of a custom pointer more persuasive (as requested by a user).
2156    
2157    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2158        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2159        pre-defined.
2160    
2161    28. Added support for (*UTF8) at the start of a pattern.
2162    
2163    29. Arrange for flags added by the "release type" setting in CMake to be shown
2164        in the configuration summary.
2165    
2166    
2167    Version 7.8 05-Sep-08
2168    ---------------------
2169    
2170    1.  Replaced UCP searching code with optimized version as implemented for Ad
2171        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2172        stage table and inline lookup instead of a function, giving speed ups of 2
2173        to 5 times on some simple patterns that I tested. Permission was given to
2174        distribute the MultiStage2.py script that generates the tables (it's not in
2175        the tarball, but is in the Subversion repository).
2176    
2177    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2178        scripts.
2179    
2180    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2181        a group with a zero qualifier. The result of the study could be incorrect,
2182        or the function might crash, depending on the pattern.
2183    
2184    4.  Caseless matching was not working for non-ASCII characters in back
2185        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2186        It now works when Unicode Property Support is available.
2187    
2188    5.  In pcretest, an escape such as \x{de} in the data was always generating
2189        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2190        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2191        truncation.
2192    
2193    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2194    
2195    7.  Added two (int) casts to pcregrep when printing the difference of two
2196        pointers, in case they are 64-bit values.
2197    
2198    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2199        test 2 if it fails.
2200    
2201    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2202        and a #define of that name to empty if it is not externally set. This is to
2203        allow users of MSVC to set it if necessary.
2204    
2205    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2206        the convenience functions in the pcre_get.c source file.
2207    
2208    11. An option change at the start of a pattern that had top-level alternatives
2209        could cause overwriting and/or a crash. This command provoked a crash in
2210        some environments:
2211    
2212          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2213    
2214        This potential security problem was recorded as CVE-2008-2371.
2215    
2216    12. For a pattern where the match had to start at the beginning or immediately
2217        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2218        pcre_dfa_exec() could read past the end of the passed subject if there was
2219        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2220        pcretest so that it places the subject at the end of its malloc-ed buffer.
2221    
2222    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2223        exec() might read past the end of the data buffer in UTF-8 mode.
2224    
2225    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2226        the data contained the byte 0x85 as part of a UTF-8 character within its
2227        first line. This applied both to normal and DFA matching.
2228    
2229    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2230        /^[^d]*?$/8 failed to match "abc".
2231    
2232    16. Added a missing copyright notice to pcrecpp_internal.h.
2233    
2234    17. Make it more clear in the documentation that values returned from
2235        pcre_exec() in ovector are byte offsets, not character counts.
2236    
2237    18. Tidied a few places to stop certain compilers from issuing warnings.
2238    
2239    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2240        supplied by Stefan Weber. I made a further small update for 7.8 because
2241        there is a change of source arrangements: the pcre_searchfuncs.c module is
2242        replaced by pcre_ucd.c.
2243    
2244    
2245    Version 7.7 07-May-08
2246    ---------------------
2247    
2248    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2249        a string to a long long, pretend we don't even have a long long." This is
2250        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2251    
2252    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2253        pre-7.6 versions, which defined a global no_arg variable instead of putting
2254        it in the RE class. (See also #8 below.)
2255    
2256    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2257        Lopes.
2258    
2259    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2260    
2261        (1) The include/exclude patterns were being applied to the whole pathnames
2262            of files, instead of just to the final components.
2263    
2264        (2) If there was more than one level of directory, the subdirectories were
2265            skipped unless they satisfied the include/exclude conditions. This is
2266            inconsistent with GNU grep (and could even be seen as contrary to the
2267            pcregrep specification - which I improved to make it absolutely clear).
2268            The action now is always to scan all levels of directory, and just
2269            apply the include/exclude patterns to regular files.
2270    
2271    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2272        --exclude_dir in the tests to avoid scanning .svn directories.
2273    
2274    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2275        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2276        doesn't support NULs in patterns.
2277    
2278    7.  Added some missing "const"s to declarations of static tables in
2279        pcre_compile.c and pcre_dfa_exec.c.
2280    
2281    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2282        caused by fix #2  above. (Subsequently also a second patch to fix the
2283        first patch. And a third patch - this was a messy problem.)
2284    
2285    9.  Applied Craig's patch to remove the use of push_back().
2286    
2287    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2288        matching function regexec().
2289    
2290    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2291        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2292        references. PCRE supports relative numbers with this syntax (I don't think
2293        Oniguruma does).
2294    
2295    12. Previously, a group with a zero repeat such as (...){0} was completely
2296        omitted from the compiled regex. However, this means that if the group
2297        was called as a subroutine from elsewhere in the pattern, things went wrong
2298        (an internal error was given). Such groups are now left in the compiled
2299        pattern, with a new opcode that causes them to be skipped at execution
2300        time.
2301    
2302    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2303        to the way PCRE behaves:
2304    
2305        (a) A lone ] character is dis-allowed (Perl treats it as data).
2306    
2307        (b) A back reference to an unmatched subpattern matches an empty string
2308            (Perl fails the current match path).
2309    
2310        (c) A data ] in a character class must be notated as \] because if the
2311            first data character in a class is ], it defines an empty class. (In
2312            Perl it is not possible to have an empty class.) The empty class []
2313            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2314            The negative empty class [^] matches any one character, independently
2315            of the DOTALL setting.
2316    
2317    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2318        non-existent subpattern following a character class starting with ']' and
2319        containing () gave an internal compiling error instead of "reference to
2320        non-existent subpattern". Fortunately, when the pattern did exist, the
2321        compiled code was correct. (When scanning forwards to check for the
2322        existence of the subpattern, it was treating the data ']' as terminating
2323        the class, so got the count wrong. When actually compiling, the reference
2324        was subsequently set up correctly.)
2325    
2326    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2327        it was being rejected as not supported by pcre_dfa_exec(), even though
2328        other assertions are supported. I have made pcre_dfa_exec() support
2329        (*FAIL).
2330    
2331    16. The implementation of 13c above involved the invention of a new opcode,
2332        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2333        cannot be changed at match time, I realized I could make a small
2334        improvement to matching performance by compiling OP_ALLANY instead of
2335        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2336        on the OP_ANY path.
2337    
2338    17. Compiling pcretest on Windows with readline support failed without the
2339        following two fixes: (1) Make the unistd.h include conditional on
2340        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2341    
2342    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2343        ncurses library to be included for pcretest when ReadLine support is
2344        requested, but also to allow for it to be overridden. This patch came from
2345        Daniel Bergström.
2346    
2347    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2348        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2349        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2350        spotting this.
2351    
2352    
2353    Version 7.6 28-Jan-08
2354    ---------------------
2355    
2356    1.  A character class containing a very large number of characters with
2357        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2358        overflow.
2359    
2360    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2361        HAVE_LONG_LONG is not defined.
2362    
2363    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2364        bring them up to date and include new features. This patch includes:
2365    
2366        - Fixed PH's badly added libz and libbz2 support.
2367        - Fixed a problem with static linking.
2368        - Added pcredemo. [But later removed - see 7 below.]
2369        - Fixed dftables problem and added an option.
2370        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2371            HAVE_LONG_LONG.
2372        - Added readline support for pcretest.
2373        - Added an listing of the option settings after cmake has run.
2374    
2375    4.  A user submitted a patch to Makefile that makes it easy to create
2376        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2377        Makefile.am that cause it to include this special target, without
2378        affecting anything else. Note that the same mingw target plus all
2379        the other distribution libraries and programs are now supported
2380        when configuring with CMake (see 6 below) instead of with
2381        Configure/Make.
2382    
2383    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2384        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2385        exported in the Windows port". It has not yet been confirmed that the patch
2386        solves the problem, but it does no harm.
2387    
2388    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2389        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2390        with CMake, and also correct the comment about stack recursion.
2391    
2392    7.  Remove the automatic building of pcredemo from the ./configure system and
2393        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2394        of a program that users should build themselves after PCRE is installed, so
2395        building it automatically is not really right. What is more, it gave
2396        trouble in some build environments.
2397    
2398    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2399    
2400    
2401    Version 7.5 10-Jan-08
2402    ---------------------
2403    
2404    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2405        values in parens when parsing an RE using the C++ wrapper."
2406    
2407    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2408        Characters greater than 255 were excluded from the class instead of being
2409        included.
2410    
2411    3.  The same bug as (2) above applied to negated POSIX classes such as
2412        [:^space:].
2413    
2414    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2415        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2416        I have changed it.
2417    
2418    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2419        first named subpattern) and a construct such as (?&a) would reference the
2420        first named subpattern whose name started with "a" (in other words, the
2421        length check was missing). Both these problems are fixed. "Subpattern name
2422        expected" is now given for (?&) (a zero-length name), and this patch also
2423        makes it give the same error for \k'' (previously it complained that that
2424        was a reference to a non-existent subpattern).
2425    
2426    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2427        this is right because (?- can be followed by option settings as well as by
2428        digits. I have, however, made the messages clearer.
2429    
2430    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2431        than the number used in the conditional) now cause a compile-time error.
2432        This is actually not compatible with Perl, which accepts such patterns, but
2433        treats the conditional as always being FALSE (as PCRE used to), but it
2434        seems to me that giving a diagnostic is better.
2435    
2436    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2437        and messages.
2438    
2439    9.  Fix two occurrences of "backslash" in comments that should have been
2440        "backspace".
2441    
2442    10. Remove two redundant lines of code that can never be obeyed (their function
2443        was moved elsewhere).
2444    
2445    11. The program that makes PCRE's Unicode character property table had a bug
2446        which caused it to generate incorrect table entries for sequences of
2447        characters that have the same character type, but are in different scripts.
2448        It amalgamated them into a single range, with the script of the first of
2449        them. In other words, some characters were in the wrong script. There were
2450        thirteen such cases, affecting characters in the following ranges:
2451    
2452          U+002b0 - U+002c1
2453          U+0060c - U+0060d
2454          U+0061e - U+00612
2455          U+0064b - U+0065e
2456          U+0074d - U+0076d
2457          U+01800 - U+01805
2458          U+01d00 - U+01d77
2459          U+01d9b - U+01dbf
2460          U+0200b - U+0200f
2461          U+030fc - U+030fe
2462          U+03260 - U+0327f
2463          U+0fb46 - U+0fbb1
2464          U+10450 - U+1049d
2465    
2466    12. The -o option (show only the matching part of a line) for pcregrep was not
2467        compatible with GNU grep in that, if there was more than one match in a
2468        line, it showed only the first of them. It now behaves in the same way as
2469        GNU grep.
2470    
2471    13. If the -o and -v options were combined for pcregrep, it printed a blank
2472        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2473        does the same. The return code can be used to tell if there were any
2474        non-matching lines.
2475    
2476    14. Added --file-offsets and --line-offsets to pcregrep.
2477    
2478    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2479        infinitely looping recursion. The bug was that positive lookaheads were not
2480        being skipped when checking for a possible empty match (negative lookaheads
2481        and both kinds of lookbehind were skipped).
2482    
2483    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2484        inclusion of <windows.h> to before rather than after the definition of
2485        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2486    
2487    17. Specifying a possessive quantifier with a specific limit for a Unicode
2488        character property caused pcre_compile() to compile bad code, which led at
2489        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2490        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2491        caused the error; without that there was no problem.
2492    
2493    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2494    
2495    19. Added --enable-pcretest-libreadline.
2496    
2497    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2498        RE::GlobalReplace(). As a result, the number of replacements returned was
2499        double what it should be. I removed one of the increments, but Craig sent a
2500        later patch that removed the other one (the right fix) and added unit tests
2501        that check the return values (which was not done before).
2502    
2503    21. Several CMake things:
2504    
2505        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2506            the names libpcre and libpcreposix, not just pcre and pcreposix.
2507    
2508        (2) The above change means that pcretest and pcregrep are now correctly
2509            linked with the newly-built libraries, not previously installed ones.
2510    
2511        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2512    
2513    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2514        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2515        UTF-8 newline character). The key issue is that the pattern starts .*;
2516        this means that the match must be either at the beginning, or after a
2517        newline. The bug was in the code for advancing after a failed match and
2518        checking that the new position followed a newline. It was not taking
2519        account of UTF-8 characters correctly.
2520    
2521    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2522        character classes. PCRE was not treating the sequence [:...:] as a
2523        character class unless the ... were all letters. Perl, however, seems to
2524        allow any characters between [: and :], though of course it rejects as
2525        unknown any "names" that contain non-letters, because all the known class
2526        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2527        for example, whereas PCRE did not - it did not recognize a POSIX character
2528        class. This seemed a bit dangerous, so the code has been changed to be
2529        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2530        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2531        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2532        Perl does, and where it didn't before.
2533    
2534    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2535        Windows environments %n is disabled by default.
2536    
2537    
2538    Version 7.4 21-Sep-07
2539    ---------------------
2540    
2541    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2542        means that a class such as [\s] counted as "explicit reference to CR or
2543        LF". That isn't really right - the whole point of the change was to try to
2544        help when there was an actual mention of one of the two characters. So now
2545        the change happens only if \r or \n (or a literal CR or LF) character is
2546        encountered.
2547    
2548    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2549        of both had grown to the point where there were only 3 bits left.
2550        Fortunately, there was spare space in the data structure, and so I have
2551        moved the internal flags into a new 16-bit field to free up more option
2552        bits.
2553    
2554    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2555        but did not set the internal JCHANGED flag - either of these is enough to
2556        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2557        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2558        start sets both bits.
2559    
2560    4.  Added options (at build time, compile time, exec time) to change \R from
2561        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2562    
2563    5.  doc/pcresyntax.html was missing from the distribution.
2564    
2565    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2566        compatibility, even though it is no longer used.
2567    
2568    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2569        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2570        windows.h file is present (where different names are used). [This was
2571        reversed later after testing - see 16 below.]
2572    
2573    8.  Changed all #include <config.h> to #include "config.h". There were also
2574        some further <pcre.h> cases that I changed to "pcre.h".
2575    
2576    9.  When pcregrep was used with the --colour option, it missed the line ending
2577        sequence off the lines that it output.
2578    
2579    10. It was pointed out to me that arrays of string pointers cause lots of
2580        relocations when a shared library is dynamically loaded. A technique of
2581        using a single long string with a table of offsets can drastically reduce
2582        these. I have refactored PCRE in four places to do this. The result is
2583        dramatic:
2584    
2585          Originally:                          290
2586          After changing UCP table:            187
2587          After changing error message table:   43
2588          After changing table of "verbs"       36
2589          After changing table of Posix names   22
2590    
2591        Thanks to the folks working on Gregex for glib for this insight.
2592    
2593    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2594        unicode-properties was also set.
2595    
2596    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2597    
2598    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2599        checked only for CRLF.
2600    
2601    14. Added casts to pcretest.c to avoid compiler warnings.
2602    
2603    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2604    
2605    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2606        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2607        entirely. This removes changes made in 7 above.
2608    
2609    17. The CMake files have been updated, and there is now more information about
2610        building with CMake in the NON-UNIX-USE document.
2611    
2612    
2613    Version 7.3 28-Aug-07
2614    ---------------------
2615    
2616     1. In the rejigging of the build system that eventually resulted in 7.1, the
2617        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2618        brackets there is not right, since it causes compilers to look for an
2619        installed pcre.h, not the version that is in the source that is being
2620        compiled (which of course may be different). I have changed it back to:
2621    
2622          #include "pcre.h"
2623    
2624        I have a vague recollection that the change was concerned with compiling in
2625        different directories, but in the new build system, that is taken care of
2626        by the VPATH setting the Makefile.
2627    
2628     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2629        when the subject happened to end in the byte 0x85 (e.g. if the last
2630        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2631        characters but of course it shouldn't be taken as a newline when it is part
2632        of another character. The bug was that, for an unlimited repeat of . in
2633        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2634        characters when looking for a newline.
2635    
2636     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2637    
2638     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2639        in debug output.
2640    
2641     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2642        long printing in the pcrecpp unittest when running under MinGW.
2643    
2644     6. ESC_K was left out of the EBCDIC table.
2645    
2646     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2647        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2648        limit also applies to "virtual nesting" when a pattern is recursive, and in
2649        this case 1000 isn't so big. I have been able to remove this limit at the
2650        expense of backing off one optimization in certain circumstances. Normally,
2651        when pcre_exec() would call its internal match() function recursively and
2652        immediately return the result unconditionally, it uses a "tail recursion"
2653        feature to save stack. However, when a subpattern that can match an empty
2654        string has an unlimited repetition quantifier, it no longer makes this
2655        optimization. That gives it a stack frame in which to save the data for
2656        checking that an empty string has been matched. Previously this was taken
2657        from the 1000-entry workspace that had been reserved. So now there is no
2658        explicit limit, but more stack is used.
2659    
2660     8. Applied Daniel's patches to solve problems with the import/export magic
2661        syntax that is required for Windows, and which was going wrong for the
2662        pcreposix and pcrecpp parts of the library. These were overlooked when this
2663        problem was solved for the main library.
2664    
2665     9. There were some crude static tests to avoid integer overflow when computing
2666        the size of patterns that contain repeated groups with explicit upper
2667        limits. As the maximum quantifier is 65535, the maximum group length was
2668        set at 30,000 so that the product of these two numbers did not overflow a
2669        32-bit integer. However, it turns out that people want to use groups that
2670        are longer than 30,000 bytes (though not repeat them that many times).
2671        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2672        made it possible to implement the integer overflow checks in a much more
2673        dynamic way, which I have now done. The artificial limitation on group
2674        length has been removed - we now have only the limit on the total length of
2675        the compiled pattern, which depends on the LINK_SIZE setting.
2676    
2677    10. Fixed a bug in the documentation for get/copy named substring when
2678        duplicate names are permitted. If none of the named substrings are set, the
2679        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2680        empty string.
2681    
2682    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2683        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2684        because the ] is interpreted as the first data character and the
2685        terminating ] is not found. PCRE has been made compatible with Perl in this
2686        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2687        cause memory overwriting.
2688    
2689    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2690        string has been matched (to stop an infinite loop). It was not recognizing
2691        a conditional subpattern that could match an empty string if that
2692        subpattern was within another subpattern. For example, it looped when
2693        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2694        condition was not nested. This bug has been fixed.
2695    
2696    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2697        past the start of the subject in the presence of bytes with the top bit
2698        set, for example "\x8aBCD".
2699    
2700    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2701        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2702    
2703    14. Optimized (?!) to (*FAIL).
2704    
2705    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2706        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2707        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2708        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2709        does: it's just the validity check that is more restrictive.
2710    
2711    16. Inserted checks for integer overflows during escape sequence (backslash)
2712        processing, and also fixed erroneous offset values for syntax errors during
2713        backslash processing.
2714    
2715    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2716        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2717    
2718    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2719        caused an overrun.
2720    
2721    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2722        something other than just ASCII characters) inside a group that had an
2723        unlimited repeat caused a loop at compile time (while checking to see
2724        whether the group could match an empty string).
2725    
2726    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2727        [\P{Any}] did so. (Error in the code for printing property names.)
2728    
2729    21. An orphan \E inside a character class could cause a crash.
2730    
2731    22. A repeated capturing bracket such as (A)? could cause a wild memory
2732        reference during compilation.
2733    
2734    23. There are several functions in pcre_compile() that scan along a compiled
2735        expression for various reasons (e.g. to see if it's fixed length for look
2736        behind). There were bugs in these functions when a repeated \p or \P was
2737        present in the pattern. These operators have additional parameters compared
2738        with \d, etc, and these were not being taken into account when moving along
2739        the compiled data. Specifically:
2740    
2741        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2742            length.
2743    
2744        (b) An item such as \pL+ within a repeated group could cause crashes or
2745            loops.
2746    
2747        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2748            "reference to non-existent subpattern" error.
2749    
2750        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2751    
2752    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2753        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2754    
2755    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2756    
2757    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2758        character were causing crashes (broken optimization).
2759    
2760    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2761        \p or \P) caused a compile-time loop.
2762    
2763    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2764        break. For example, the unstudied pattern [\r\n]A does not match the string
2765        "\r\nA" because change 7.0/46 below moves the current point on by two
2766        characters after failing to match at the start. However, the pattern \nA
2767        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2768        the same is true. There doesn't seem any very clean way out of this, but
2769        what I have chosen to do makes the common cases work: PCRE now takes note
2770        of whether there can be an explicit match for \r or \n anywhere in the
2771        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2772        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2773        pattern has explicit CR or LF references.
2774    
2775    29. Added (*CR) etc for changing newline setting at start of pattern.
2776    
2777    
2778    Version 7.2 19-Jun-07
2779    ---------------------
2780    
2781     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2782        which is apparently normally available under Windows.
2783    
2784     2. Re-jig the pcregrep tests with different newline settings in an attempt
2785        to make them independent of the local environment's newline setting.
2786    
2787     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2788    
2789     4. Some of the "internals" tests were previously cut out when the link size
2790        was not 2, because the output contained actual offsets. The recent new
2791        "Z" feature of pcretest means that these can be cut out, making the tests
2792        usable with all link sizes.
2793    
2794     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2795        stack recursion. This gives a massive performance boost under BSD, but just
2796        a small improvement under Linux. However, it saves one field in the frame
2797        in all cases.
2798    
2799     6. Added more features from the forthcoming Perl 5.10:
2800    
2801        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2802            recursion call. It refers to the nth most recently opened parentheses.
2803    
2804        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2805            to be opened parentheses.
2806    
2807        (c) Conditions that refer to capturing parentheses can be specified
2808            relatively, for example, (?(-2)... or (?(+3)...
2809    
2810        (d) \K resets the start of the current match so that everything before
2811            is not part of it.
2812    
2813        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2814    
2815        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2816            reference syntax.
2817    
2818        (g) (?| introduces a group in which the numbering of parentheses in each
2819            alternative starts with the same number.
2820    
2821        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2822    
2823     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2824        PCRE_INFO_JCHANGED.
2825    
2826     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2827        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2828        for detecting groups that can match an empty string.
2829    
2830     9. A pattern with a very large number of alternatives (more than several
2831        hundred) was running out of internal workspace during the pre-compile
2832        phase, where pcre_compile() figures out how much memory will be needed. A
2833        bit of new cunning has reduced the workspace needed for groups with
2834        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2835        workspace instead of running out of the 4096 that are available.
2836    
2837    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2838    
2839    11. Applied patch from Google to remove an optimization that didn't quite work.
2840        The report of the bug said:
2841    
2842          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2843          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2844          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2845    
2846    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2847        it matched the wrong number of bytes.
2848    
2849    
2850    Version 7.1 24-Apr-07
2851    ---------------------
2852    
2853     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2854        that is more "standard", making use of automake and other Autotools. There
2855        is some re-arrangement of the files and adjustment of comments consequent
2856        on this.
2857    
2858     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2859        for recursive directory scanning broke on some systems because the files
2860        are not scanned in any specific order and on different systems the order
2861        was different. A call to "sort" has been inserted into RunGrepTest for the
2862        approprate test as a short-term fix. In the longer term there may be an
2863        alternative.
2864    
2865     3. I had an email from Eric Raymond about problems translating some of PCRE's
2866        man pages to HTML (despite the fact that I distribute HTML pages, some
2867        people do their own conversions for various reasons). The problems
2868        concerned the use of low-level troff macros .br and .in. I have therefore
2869        removed all such uses from the man pages (some were redundant, some could
2870        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2871        HTML has been updated to handle .nf/.fi and to complain if it encounters
2872        .br or .in.
2873    
2874     4. Updated comments in configure.ac that get placed in config.h.in and also
2875        arranged for config.h to be included in the distribution, with the name
2876        config.h.generic, for the benefit of those who have to compile without
2877        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2878    
2879     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2880        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2881        makevp.bat for the current PCRE, using the additional files
2882        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2883    
2884     6. A Windows user reported a minor discrepancy with test 2, which turned out
2885        to be caused by a trailing space on an input line that had got lost in his
2886        copy. The trailing space was an accident, so I've just removed it.
2887    
2888     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2889        that is needed.
2890    
2891     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2892        as "const" (a) because they are and (b) because it helps the PHP
2893        maintainers who have recently made a script to detect big data structures
2894        in the php code that should be moved to the .rodata section. I remembered
2895        to update Builducptable as well, so it won't revert if ucptable.h is ever
2896        re-created.
2897    
2898     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2899        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2900        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2901        support is not required. This saves 1.5-2K of code, which is important in
2902        some applications.
2903    
2904        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2905        so as not to refer to the tables, even though these functions will never be
2906        called when UTF-8 support is disabled. Otherwise there are problems with a
2907        shared library.
2908    
2909    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2910    
2911        (a) It was defining its arguments as char * instead of void *.
2912    
2913        (b) It was assuming that all moves were upwards in memory; this was true
2914            a long time ago when I wrote it, but is no longer the case.
2915    
2916        The emulated memove() is provided for those environments that have neither
2917        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2918        is clearly not the case, as these two bugs were recently reported.
2919    
2920    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2921        and Detrail to create the HTML documentation, the .txt form of the man
2922        pages, and it removes trailing spaces from listed files. It also creates
2923        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2924        case, it wraps all the #defines with #ifndefs. This script should be run
2925        before "make dist".
2926    
2927    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2928        with Unicode property support.
2929    
2930        (a) For a maximizing quantifier, if the two different cases of the
2931            character were of different lengths in their UTF-8 codings (there are
2932            some cases like this - I found 11), and the matching function had to
2933            back up over a mixture of the two cases, it incorrectly assumed they
2934            were both the same length.
2935    
2936        (b) When PCRE was configured to use the heap rather than the stack for
2937            recursion during matching, it was not correctly preserving the data for
2938            the other case of a UTF-8 character when checking ahead for a match
2939            while processing a minimizing repeat. If the check also involved
2940            matching a wide character, but failed, corruption could cause an
2941            erroneous result when trying to check for a repeat of the original
2942            character.
2943    
2944    13. Some tidying changes to the testing mechanism:
2945    
2946        (a) The RunTest script now detects the internal link size and whether there
2947            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2948            values substituted by "configure". (The RunGrepTest script already did
2949            this for UTF-8.) The configure.ac script no longer substitutes the
2950            relevant variables.
2951    
2952        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2953            with length and offset values. This means that the output is different
2954            for different internal link sizes. Test 2 is skipped for link sizes
2955            other than 2 because of this, bypassing the problem. Unfortunately,
2956            there was also a test in test 3 (the locale tests) that used /B and
2957            failed for link sizes other than 2. Rather than cut the whole test out,
2958            I have added a new /Z option to pcretest that replaces the length and
2959            offset values with spaces. This is now used to make test 3 independent
2960            of link size. (Test 2 will be tidied up later.)
2961    
2962    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2963        segmentation fault instead of returning the appropriate error message.
2964    
2965    15. In multiline mode when the newline sequence was set to "any", the pattern
2966        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2967        This doesn't seem right; it now treats the CRLF combination as the line
2968        ending, and so does not match in that case. It's only a pattern such as ^$
2969        that would hit this one: something like ^ABC$ would have failed after \r
2970        and then tried again after \r\n.
2971    
2972    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2973        in an attempt to make files that differ only in their line terminators
2974        compare equal. This works on Linux.
2975    
2976    17. Under certain error circumstances pcregrep might try to free random memory
2977        as it exited. This is now fixed, thanks to valgrind.
2978    
2979    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2980        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2981        was because its rules for how to advance for /g after matching an empty
2982        string at the end of a line did not allow for this case. They now check for
2983        it specially.
2984    
2985    20. pcretest is supposed to handle patterns and data of any length, by
2986        extending its buffers when necessary. It was getting this wrong when the
2987        buffer for a data line had to be extended.
2988    
2989    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2990        CRLF as a newline sequence.
2991    
2992    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2993        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2994        I have nevertheless tidied it up.
2995    
2996    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2997    
2998    24. Added a man page for pcre-config.
2999    
3000    
3001    Version 7.0 19-Dec-06
3002    ---------------------
3003    
3004     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3005        moving to gcc 4.1.1.
3006    
3007     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3008        sys/time.h, which is documented as needed for this function. It doesn't
3009        seem to matter on Linux, but it showed up on some releases of OS X.
3010    
3011     3. It seems that there are systems where bytes whose values are greater than
3012        127 match isprint() in the "C" locale. The "C" locale should be the
3013        default when a C program starts up. In most systems, only ASCII printing
3014        characters match isprint(). This difference caused the output from pcretest
3015        to vary, making some of the tests fail. I have changed pcretest so that:
3016    
3017        (a) When it is outputting text in the compiled version of a pattern, bytes
3018            other than 32-126 are always shown as hex escapes.
3019    
3020        (b) When it is outputting text that is a matched part of a subject string,
3021            it does the same, unless a different locale has been set for the match
3022            (using the /L modifier). In this case, it uses isprint() to decide.
3023    
3024     4. Fixed a major bug that caused incorrect computation of the amount of memory
3025        required for a compiled pattern when options that changed within the
3026        pattern affected the logic of the preliminary scan that determines the
3027        length. The relevant options are -x, and -i in UTF-8 mode. The result was
3028        that the computed length was too small. The symptoms of this bug were
3029        either the PCRE error "internal error: code overflow" from pcre_compile(),
3030        or a glibc crash with a message such as "pcretest: free(): invalid next
3031        size (fast)". Examples of patterns that provoked this bug (shown in
3032        pcretest format) are:
3033    
3034          /(?-x: )/x
3035          /(?x)(?-x: \s*#\s*)/
3036          /((?i)[\x{c0}])/8
3037          /(?i:[\x{c0}])/8
3038    
3039        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3040        is now done differently.
3041    
3042     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3043        wrapper classes; (b) implement a new function in the C++ scanner that is
3044        more efficient than the old way of doing things because it avoids levels of
3045        recursion in the regex matching; (c) add a paragraph to the documentation
3046        for the FullMatch() function.
3047    
3048     6. The escape sequence \n was being treated as whatever was defined as
3049        "newline". Not only was this contrary to the documentation, which states
3050        that \n is character 10 (hex 0A), but it also went horribly wrong when
3051        "newline" was defined as CRLF. This has been fixed.
3052    
3053     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3054        was being set to -1 for the "end of line" case (supposedly a value that no
3055        character can have). Though this value is never used (the check for end of
3056        line is "zero bytes in current character"), it caused compiler complaints.
3057        I've changed it to 0xffffffff.
3058    
3059     8. In pcre_version.c, the version string was being built by a sequence of
3060        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3061        string (as it is for production releases) called a macro with an empty
3062        argument. The C standard says the result of this is undefined. The gcc
3063        compiler treats it as an empty string (which was what was wanted) but it is
3064        reported that Visual C gives an error. The source has been hacked around to
3065        avoid this problem.
3066    
3067     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3068        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3069        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3070        of them did).
3071    
3072    10. Originally, pcretest opened its input and output without "b"; then I was
3073        told that "b" was needed in some environments, so it was added for release
3074        5.0 to both the input and output. (It makes no difference on Unix-like
3075        systems.) Later I was told that it is wrong for the input on Windows. I've
3076        now abstracted the modes into two macros, to make it easier to fiddle with
3077        them, and removed "b" from the input mode under Windows.
3078    
3079    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3080    
3081    12. Added -help and --help to pcretest as an official way of being reminded
3082        of the options.
3083    
3084    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3085        and pcrecpp.cc because they annoy compilers at high warning levels.
3086    
3087    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3088    
3089    15. Fixed an occurrence of == in configure.ac that should have been = (shell
3090        scripts are not C programs :-) and which was not noticed because it works
3091        on Linux.
3092    
3093    16. pcretest is supposed to handle any length of pattern and data line (as one
3094        line or as a continued sequence of lines) by extending its input buffer if
3095        necessary. This feature was broken for very long pattern lines, leading to
3096        a string of junk being passed to pcre_compile() if the pattern was longer
3097        than about 50K.
3098    
3099    17. I have done a major re-factoring of the way pcre_compile() computes the
3100        amount of memory needed for a compiled pattern. Previously, there was code
3101        that made a preliminary scan of the pattern in order to do this. That was
3102        OK when PCRE was new, but as the facilities have expanded, it has become
3103        harder and harder to keep it in step with the real compile phase, and there
3104        have been a number of bugs (see for example, 4 above). I have now found a
3105        cunning way of running the real compile function in a "fake" mode that
3106        enables it to compute how much memory it would need, while actually only
3107        ever using a few hundred bytes of working memory and without too many
3108        tests of the mode. This should make future maintenance and development
3109        easier. A side effect of this work is that the limit of 200 on the nesting
3110        depth of parentheses has been removed (though this was never a serious
3111        limitation, I suspect). However, there is a downside: pcre_compile() now
3112        runs more slowly than before (30% or more, depending on the pattern). I
3113        hope this isn't a big issue. There is no effect on runtime performance.
3114    
3115    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3116        newline (only possible for the last line of a file) and it was a
3117        pattern that set a locale (followed by /Lsomething), pcretest crashed.
3118    
3119    19. Added additional timing features to pcretest. (1) The -tm option now times
3120        matching only, not compiling. (2) Both -t and -tm can be followed, as a
3121        separate command line item, by a number that specifies the number of
3122        repeats to use when timing. The default is 50000; this gives better
3123        precision, but takes uncomfortably long for very large patterns.
3124    
3125    20. Extended pcre_study() to be more clever in cases where a branch of a
3126        subpattern has no definite first character. For example, (a*|b*)[cd] would
3127        previously give no result from pcre_study(). Now it recognizes that the
3128        first character must be a, b, c, or d.
3129    
3130    21. There was an incorrect error "recursive call could loop indefinitely" if
3131        a subpattern (or the entire pattern) that was being tested for matching an
3132        empty string contained only one non-empty item after a nested subpattern.
3133        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3134        incorrectly, because the \d was being skipped in the check.
3135    
3136    22. The pcretest program now has a new pattern option /B and a command line
3137        option -b, which is equivalent to adding /B to every pattern. This causes
3138        it to show the compiled bytecode, without the additional information that
3139        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3140        is the same as /B/I).
3141    
3142    23. A new optimization is now able automatically to treat some sequences such
3143        as a*b as a*+b. More specifically, if something simple (such as a character
3144        or a simple class like \d) has an unlimited quantifier, and is followed by
3145        something that cannot possibly match the quantified thing, the quantifier
3146        is automatically "possessified".
3147    
3148    24. A recursive reference to a subpattern whose number was greater than 39
3149        went wrong under certain circumstances in UTF-8 mode. This bug could also
3150        have affected the operation of pcre_study().
3151    
3152    25. Realized that a little bit of performance could be had by replacing
3153        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3154    
3155    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3156    
3157    27. Possessive quantifiers such as a++ were previously implemented by turning
3158        them into atomic groups such as ($>a+). Now they have their own opcodes,
3159        which improves performance. This includes the automatically created ones
3160        from 23 above.
3161    
3162    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3163        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3164        the first matched character to be a colon. This applied both to named and
3165        numbered groups.
3166    
3167    29. The ucpinternal.h header file was missing its idempotency #ifdef.
3168    
3169    30. I was sent a "project" file called libpcre.a.dev which I understand makes
3170        building PCRE on Windows easier, so I have included it in the distribution.
3171    
3172    31. There is now a check in pcretest against a ridiculously large number being
3173        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3174        loop, the loop is abandoned.
3175    
3176    32. Forward references to subpatterns in conditions such as (?(2)...) where
3177        subpattern 2 is defined later cause pcre_compile() to search forwards in
3178        the pattern for the relevant set of parentheses. This search went wrong
3179        when there were unescaped parentheses in a character class, parentheses
3180        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3181    
3182    33. "Subroutine" calls and backreferences were previously restricted to
3183        referencing subpatterns earlier in the regex. This restriction has now
3184        been removed.
3185    
3186    34. Added a number of extra features that are going to be in Perl 5.10. On the
3187        whole, these are just syntactic alternatives for features that PCRE had
3188        previously implemented using the Python syntax or my own invention. The
3189        other formats are all retained for compatibility.
3190    
3191        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3192            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3193            also .NET compatible.
3194    
3195        (b) A recursion or subroutine call to a named group can now be defined as
3196            (?&name) as well as (?P>name).
3197    
3198        (c) A backreference to a named group can now be defined as \k<name> or
3199            \k'name' as well as (?P=name). The new forms, as well as being in Perl
3200            5.10, are also .NET compatible.
3201    
3202        (d) A conditional reference to a named group can now use the syntax
3203            (?(<name>) or (?('name') as well as (?(name).
3204    
3205        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3206            groups (named and numbered) that are never evaluated inline, but can be
3207            called as "subroutines" from elsewhere. In effect, the DEFINE condition
3208            is always false. There may be only one alternative in such a group.
3209    
3210        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3211            as the simple (?(R). The condition is true only if the most recent
3212            recursion is that of the given number or name. It does not search out
3213            through the entire recursion stack.
3214    
3215        (g) The escape \gN or \g{N} has been added, where N is a positive or
3216            negative number, specifying an absolute or relative reference.
3217    
3218    35. Tidied to get rid of some further signed/unsigned compiler warnings and
3219        some "unreachable code" warnings.
3220    
3221    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3222        things, this adds five new scripts.
3223    
3224    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3225        There were also incompatibilities regarding the handling of \Q..\E inside
3226        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3227        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3228    
3229    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3230        matches an empty string, and forcibly breaks the loop. There were bugs in
3231        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
3232        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
3233        separate and independent bugs (that affected different cases) have been
3234        fixed.
3235    
3236    39. Refactored the code to abolish the use of different opcodes for small
3237        capturing bracket numbers. This is a tidy that I avoided doing when I
3238        removed the limit on the number of capturing brackets for 3.5 back in 2001.
3239        The new approach is not only tidier, it makes it possible to reduce the
3240        memory needed to fix the previous bug (38).
3241    
3242    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3243        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3244        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3245        mode.
3246    
3247    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3248        report.
3249    
3250    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3251        copy construction and assignment in the C++ wrapper.
3252    
3253    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3254        couple of bugs that could have given wrong results in the "--newline=crlf"
3255        case.
3256    
3257    44. Added a number of casts and did some reorganization of signed/unsigned int
3258        variables following suggestions from Dair Grant. Also renamed the variable
3259        "this" as "item" because it is a C++ keyword.
3260    
3261    45. Arranged for dftables to add
3262    
3263          #include "pcre_internal.h"
3264    
3265        to pcre_chartables.c because without it, gcc 4.x may remove the array
3266        definition from the final binary if PCRE is built into a static library and
3267        dead code stripping is activated.
3268    
3269    46. For an unanchored pattern, if a match attempt fails at the start of a
3270        newline sequence, and the newline setting is CRLF or ANY, and the next two
3271        characters are CRLF, advance by two characters instead of one.
3272    
3273    
3274    Version 6.7 04-Jul-06
3275    ---------------------
3276    
3277     1. In order to handle tests when input lines are enormously long, pcretest has
3278        been re-factored so that it automatically extends its buffers when
3279        necessary. The code is crude, but this _is_ just a test program. The
3280        default size has been increased from 32K to 50K.
3281    
3282     2. The code in pcre_study() was using the value of the re argument before
3283        testing it for NULL. (Of course, in any sensible call of the function, it
3284        won't be NULL.)
3285    
3286     3. The memmove() emulation function in pcre_internal.h, which is used on
3287        systems that lack both memmove() and bcopy() - that is, hardly ever -
3288        was missing a "static" storage class specifier.
3289    
3290     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3291        containing an extended class (one that cannot be represented by a bitmap
3292        because it contains high-valued characters or Unicode property items, e.g.
3293        [\pZ]). Almost always one would set UTF-8 mode when processing such a
3294        pattern, but PCRE should not loop if you do not (it no longer does).
3295        [Detail: two cases were found: (a) a repeated subpattern containing an
3296        extended class; (b) a recursive reference to a subpattern that followed a
3297        previous extended class. It wasn't skipping over the extended class
3298        correctly when UTF-8 mode was not set.]
3299    
3300     5. A negated single-character class was not being recognized as fixed-length
3301        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3302        compile error "lookbehind assertion is not fixed length".
3303    
3304     6. The RunPerlTest auxiliary script was showing an unexpected difference
3305        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3306        write a Perl script that can interpret lines of an input file either as
3307        byte characters or as UTF-8, which is what "perltest" was being required to
3308        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3309        can't do is switch easily at run time between having the "use utf8;" pragma
3310        or not. In the end, I fudged it by using the RunPerlTest script to insert
3311        "use utf8;" explicitly for the UTF-8 tests.
3312    
3313     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3314        the end of the subject string, contrary to the documentation and to what
3315        Perl does. This was true of both matching functions. Now it matches only at
3316        the start of the subject and immediately after *internal* newlines.
3317    
3318     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3319        a pointer to an int instead of a pointer to an unsigned long int. This
3320        caused problems on 64-bit systems.
3321    
3322     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3323        instance of the 'standard' template library not being so standard".
3324    
3325    10. There was no check on the number of named subpatterns nor the maximum
3326        length of a subpattern name. The product of these values is used to compute
3327        the size of the memory block for a compiled pattern. By supplying a very
3328        long subpattern name and a large number of named subpatterns, the size
3329        computation could be caused to overflow. This is now prevented by limiting
3330        the length of names to 32 characters, and the number of named subpatterns
3331        to 10,000.
3332    
3333    11. Subpatterns that are repeated with specific counts have to be replicated in
3334        the compiled pattern. The size of memory for this was computed from the
3335        length of the subpattern and the repeat count. The latter is limited to
3336        65535, but there was no limit on the former, meaning that integer overflow
3337        could in principle occur. The compiled length of a repeated subpattern is
3338        now limited to 30,000 bytes in order to prevent this.
3339    
3340    12. Added the optional facility to have named substrings with the same name.
3341    
3342    13. Added the ability to use a named substring as a condition, using the
3343        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3344        are numbers (not recommended). Forward references are permitted.
3345    
3346    14. Added forward references in named backreferences (if you see what I mean).
3347    
3348    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3349        pattern could run off the end of the subject. For example, the pattern
3350        "(?s)(.{1,5})"8 did this with the subject "ab".
3351    
3352    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3353        PCRE_CASELESS was set when matching characters that were quantified with ?
3354        or *.
3355    
3356    17. A character class other than a single negated character that had a minimum
3357        but no maximum quantifier - for example [ab]{6,} - was not handled
3358        correctly by pce_dfa_exec(). It would match only one character.
3359    
3360    18. A valid (though odd) pattern that looked like a POSIX character
3361        class but used an invalid character after [ (for example [[,abc,]]) caused
3362        pcre_compile() to give the error "Failed: internal error: code overflow" or
3363        in some cases to crash with a glibc free() error. This could even happen if
3364        the pattern terminated after [[ but there just happened to be a sequence of
3365        letters, a binary zero, and a closing ] in the memory that followed.
3366    
3367    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3368        over the years. Originally (before any Unicode support), just the bottom 8
3369        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3370        output from "man perlunicode" includes this:
3371    
3372          The regular expression compiler produces polymorphic opcodes.  That
3373          is, the pattern adapts to the data and automatically switches to
3374          the Unicode character scheme when presented with Unicode data--or
3375          instead uses a traditional byte scheme when presented with byte
3376          data.
3377    
3378        Sadly, a wide octal escape does not cause a switch, and in a string with
3379        no other multibyte characters, these octal escapes are treated as before.
3380        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3381        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3382        Unicode string.
3383    
3384        I have not perpetrated such confusion in PCRE. Up till now, it took just
3385        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3386        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3387        translate to the appropriate multibyte character.
3388    
3389    29. Applied some refactoring to reduce the number of warnings from Microsoft
3390        and Borland compilers. This has included removing the fudge introduced
3391        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3392        a warning about an unused variable.
3393    
3394    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3395        characters since release 4.0, because Perl (from release 5.004) does not.
3396        [Or at least, is documented not to: some releases seem to be in conflict
3397        with the documentation.] However, when a pattern was studied with
3398        pcre_study() and all its branches started with \s, PCRE still included VT
3399        as a possible starting character. Of course, this did no harm; it just
3400        caused an unnecessary match attempt.
3401    
3402    22. Removed a now-redundant internal flag bit that recorded the fact that case
3403        dependency changed within the pattern. This was once needed for "required
3404        byte" processing, but is no longer used. This recovers a now-scarce options
3405        bit. Also moved the least significant internal flag bit to the most-
3406        significant bit of the word, which was not previously used (hangover from
3407        the days when it was an int rather than a uint) to free up another bit for
3408        the future.
3409    
3410    23. Added support for CRLF line endings as well as CR and LF. As well as the
3411        default being selectable at build time, it can now be changed at runtime
3412        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3413        specify that it is scanning data with non-default line endings.
3414    
3415    24. Changed the definition of CXXLINK to make it agree with the definition of
3416        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3417    
3418    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3419        recursions. This makes a big different to stack usage for some patterns.
3420    
3421    26. If a subpattern containing a named recursion or subroutine reference such
3422        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3423        the space required for the compiled pattern went wrong and gave too small a
3424        value. Depending on the environment, this could lead to "Failed: internal
3425        error: code overflow at offset 49" or "glibc detected double free or
3426        corruption" errors.
3427    
3428    27. Applied patches from Google (a) to support the new newline modes and (b) to
3429        advance over multibyte UTF-8 characters in GlobalReplace.
3430    
3431    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3432        difference for some implementation of PCRE in some Windows version.
3433    
3434    29. Added some extra testing facilities to pcretest:
3435    
3436        \q<number>   in a data line sets the "match limit" value
3437        \Q<number>   in a data line sets the "match recursion limt" value
3438        -S <number>  sets the stack size, where <number> is in megabytes
3439    
3440        The -S option isn't available for Windows.
3441    
3442    
3443    Version 6.6 06-Feb-06
3444    ---------------------
3445    
3446     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3447        in pcreposix.h. I have copied the definition from pcre.h.
3448    
3449     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3450        because pcre.h is no longer a built file.
3451    
3452     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3453        not normally included in the compiled code.
3454    
3455    
3456    Version 6.5 01-Feb-06
3457    ---------------------
3458    
3459     1. When using the partial match feature with pcre_dfa_exec(), it was not
3460        anchoring the second and subsequent partial matches at the new starting
3461        point. This could lead to incorrect results. For example, with the pattern
3462        /1234/, partially matching against "123" and then "a4" gave a match.
3463    
3464     2. Changes to pcregrep:
3465    
3466        (a) All non-match returns from pcre_exec() were being treated as failures
3467            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3468            error message is output. Some extra information is given for the
3469            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3470            probably the only errors that are likely to be caused by users (by
3471            specifying a regex that has nested indefinite repeats, for instance).
3472            If there are more than 20 of these errors, pcregrep is abandoned.
3473    
3474        (b) A binary zero was treated as data while matching, but terminated the
3475            output line if it was written out. This has been fixed: binary zeroes
3476            are now no different to any other data bytes.
3477    
3478        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3479            used to set a locale for matching. The --locale=xxxx long option has
3480            been added (no short equivalent) to specify a locale explicitly on the
3481            pcregrep command, overriding the environment variables.
3482    
3483        (d) When -B was used with -n, some line numbers in the output were one less
3484            than they should have been.
3485    
3486        (e) Added the -o (--only-matching) option.
3487    
3488        (f) If -A or -C was used with -c (count only), some lines of context were
3489            accidentally printed for the final match.
3490    
3491        (g) Added the -H (--with-filename) option.
3492    
3493        (h) The combination of options -rh failed to suppress file names for files
3494            that were found from directory arguments.
3495    
3496        (i) Added the -D (--devices) and -d (--directories) options.
3497    
3498        (j) Added the -F (--fixed-strings) option.
3499    
3500        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3501    
3502        (l) Added the --colo(u)r option.
3503    
3504        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3505            is not present by default.
3506    
3507     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3508        items such as (?R) or (?1), when the recursion could match a number of
3509        alternatives. If it matched one of the alternatives, but subsequently,
3510        outside the recursion, there was a failure, the code tried to back up into
3511        the recursion. However, because of the way PCRE is implemented, this is not
3512        possible, and the result was an incorrect result from the match.
3513    
3514        In order to prevent this happening, the specification of recursion has
3515        been changed so that all such subpatterns are automatically treated as
3516        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3517    
3518     4. I had overlooked the fact that, in some locales, there are characters for
3519        which isalpha() is true but neither isupper() nor islower() are true. In
3520        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3521        and ordfeminine) are like this. This affected the treatment of \w and \W
3522        when they appeared in character classes, but not when they appeared outside
3523        a character class. The bit map for "word" characters is now created
3524        separately from the results of isalnum() instead of just taking it from the
3525        upper, lower, and digit maps. (Plus the underscore character, of course.)
3526    
3527     5. The above bug also affected the handling of POSIX character classes such as
3528        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3529        permanent tables. Instead, the bit maps for such a class were previously
3530        created as the appropriate unions of the upper, lower, and digit bitmaps.
3531        Now they are created by subtraction from the [[:word:]] class, which has
3532        its own bitmap.
3533    
3534   &