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

Diff of /code/trunk/ChangeLog

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

revision 89 by nigel, Sat Feb 24 21:41:27 2007 UTC revision 418 by ph10, Fri Apr 24 09:17:18 2009 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.00 ??-???-??
5    ----------------------
6    
7    1.  The table for translating pcre_compile() error codes into POSIX error codes
8        was out-of-date, and there was no check on the pcre_compile() error code
9        being within the table. This could lead to an OK return being given in
10        error.
11    
12    
13    Version 7.9 11-Apr-09
14    ---------------------
15    
16    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
17        (pcretest), all targets were linked against these libraries. This included
18        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
19        libraries. This caused unwanted dependencies to be created. This problem
20        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
21        pcretest is linked with readline.
22    
23    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
24        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
25        moved outside it again, because FALSE and TRUE are already defined in AIX,
26        but BOOL is not.
27    
28    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
29        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
30    
31    4.  The pcregrep documentation said spaces were inserted as well as colons (or
32        hyphens) following file names and line numbers when outputting matching
33        lines. This is not true; no spaces are inserted. I have also clarified the
34        wording for the --colour (or --color) option.
35    
36    5.  In pcregrep, when --colour was used with -o, the list of matching strings
37        was not coloured; this is different to GNU grep, so I have changed it to be
38        the same.
39    
40    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
41        each matching line was coloured. Now it goes on to look for further matches
42        of any of the test patterns, which is the same behaviour as GNU grep.
43    
44    7.  A pattern that could match an empty string could cause pcregrep to loop; it
45        doesn't make sense to accept an empty string match in pcregrep, so I have
46        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
47        seems to be how GNU grep behaves.
48    
49    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
50        start or after a newline", because the conditional assertion was not being
51        correctly handled. The rule now is that both the assertion and what follows
52        in the first alternative must satisfy the test.
53    
54    9.  If auto-callout was enabled in a pattern with a conditional group whose
55        condition was an assertion, PCRE could crash during matching, both with
56        pcre_exec() and pcre_dfa_exec().
57    
58    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
59        used for matching.
60    
61    11. Unicode property support in character classes was not working for
62        characters (bytes) greater than 127 when not in UTF-8 mode.
63    
64    12. Added the -M command line option to pcretest.
65    
66    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
67    
68    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
69    
70    16. Added comments and documentation about mis-use of no_arg in the C++
71        wrapper.
72    
73    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
74        from Martin Jerabek that uses macro names for all relevant character and
75        string constants.
76    
77    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
78        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
79        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
80        these, but not everybody uses configure.
81    
82    19. A conditional group that had only one branch was not being correctly
83        recognized as an item that could match an empty string. This meant that an
84        enclosing group might also not be so recognized, causing infinite looping
85        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
86        with the subject "ab", where knowledge that the repeated group can match
87        nothing is needed in order to break the loop.
88    
89    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
90        exec(), but without supplying a callout function, matching went wrong.
91    
92    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
93        leak if the size of the offset vector was greater than 30. When the vector
94        is smaller, the saved offsets during recursion go onto a local stack
95        vector, but for larger vectors malloc() is used. It was failing to free
96        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
97        error, in fact).
98    
99    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
100        heapframe that is used only when UTF-8 support is enabled. This caused no
101        problem, but was untidy.
102    
103    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
104        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
105        included within another project.
106    
107    24. Steven Van Ingelgem's patches to add more options to the CMake support,
108        slightly modified by me:
109    
110          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
111              not building pcregrep.
112    
113          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
114              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
115    
116    25. Forward references, both numeric and by name, in patterns that made use of
117        duplicate group numbers, could behave incorrectly or give incorrect errors,
118        because when scanning forward to find the reference group, PCRE was not
119        taking into account the duplicate group numbers. A pattern such as
120        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
121    
122    26. Changed a few more instances of "const unsigned char *" to USPTR, making
123        the feature of a custom pointer more persuasive (as requested by a user).
124    
125    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
126        pcretest.c, inside #ifndefs, because it seems they are sometimes already
127        pre-defined.
128    
129    28. Added support for (*UTF8) at the start of a pattern.
130    
131    29. Arrange for flags added by the "release type" setting in CMake to be shown
132        in the configuration summary.
133    
134    
135    Version 7.8 05-Sep-08
136    ---------------------
137    
138    1.  Replaced UCP searching code with optimized version as implemented for Ad
139        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
140        stage table and inline lookup instead of a function, giving speed ups of 2
141        to 5 times on some simple patterns that I tested. Permission was given to
142        distribute the MultiStage2.py script that generates the tables (it's not in
143        the tarball, but is in the Subversion repository).
144    
145    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
146        scripts.
147    
148    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
149        a group with a zero qualifier. The result of the study could be incorrect,
150        or the function might crash, depending on the pattern.
151    
152    4.  Caseless matching was not working for non-ASCII characters in back
153        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
154        It now works when Unicode Property Support is available.
155    
156    5.  In pcretest, an escape such as \x{de} in the data was always generating
157        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
158        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
159        truncation.
160    
161    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
162    
163    7.  Added two (int) casts to pcregrep when printing the difference of two
164        pointers, in case they are 64-bit values.
165    
166    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
167        test 2 if it fails.
168    
169    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
170        and a #define of that name to empty if it is not externally set. This is to
171        allow users of MSVC to set it if necessary.
172    
173    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
174        the convenience functions in the pcre_get.c source file.
175    
176    11. An option change at the start of a pattern that had top-level alternatives
177        could cause overwriting and/or a crash. This command provoked a crash in
178        some environments:
179    
180          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
181    
182        This potential security problem was recorded as CVE-2008-2371.
183    
184    12. For a pattern where the match had to start at the beginning or immediately
185        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
186        pcre_dfa_exec() could read past the end of the passed subject if there was
187        no match. To help with detecting such bugs (e.g. with valgrind), I modified
188        pcretest so that it places the subject at the end of its malloc-ed buffer.
189    
190    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
191        exec() might read past the end of the data buffer in UTF-8 mode.
192    
193    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
194        the data contained the byte 0x85 as part of a UTF-8 character within its
195        first line. This applied both to normal and DFA matching.
196    
197    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
198        /^[^d]*?$/8 failed to match "abc".
199    
200    16. Added a missing copyright notice to pcrecpp_internal.h.
201    
202    17. Make it more clear in the documentation that values returned from
203        pcre_exec() in ovector are byte offsets, not character counts.
204    
205    18. Tidied a few places to stop certain compilers from issuing warnings.
206    
207    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
208        supplied by Stefan Weber. I made a further small update for 7.8 because
209        there is a change of source arrangements: the pcre_searchfuncs.c module is
210        replaced by pcre_ucd.c.
211    
212    
213    Version 7.7 07-May-08
214    ---------------------
215    
216    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
217        a string to a long long, pretend we don't even have a long long." This is
218        done by checking for the strtoq, strtoll, and _strtoi64 functions.
219    
220    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
221        pre-7.6 versions, which defined a global no_arg variable instead of putting
222        it in the RE class. (See also #8 below.)
223    
224    3.  Remove a line of dead code, identified by coverity and reported by Nuno
225        Lopes.
226    
227    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
228    
229        (1) The include/exclude patterns were being applied to the whole pathnames
230            of files, instead of just to the final components.
231    
232        (2) If there was more than one level of directory, the subdirectories were
233            skipped unless they satisfied the include/exclude conditions. This is
234            inconsistent with GNU grep (and could even be seen as contrary to the
235            pcregrep specification - which I improved to make it absolutely clear).
236            The action now is always to scan all levels of directory, and just
237            apply the include/exclude patterns to regular files.
238    
239    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
240        --exclude_dir in the tests to avoid scanning .svn directories.
241    
242    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
243        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
244        doesn't support NULs in patterns.
245    
246    7.  Added some missing "const"s to declarations of static tables in
247        pcre_compile.c and pcre_dfa_exec.c.
248    
249    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
250        caused by fix #2  above. (Subsequently also a second patch to fix the
251        first patch. And a third patch - this was a messy problem.)
252    
253    9.  Applied Craig's patch to remove the use of push_back().
254    
255    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
256        matching function regexec().
257    
258    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
259        which, however, unlike Perl's \g{...}, are subroutine calls, not back
260        references. PCRE supports relative numbers with this syntax (I don't think
261        Oniguruma does).
262    
263    12. Previously, a group with a zero repeat such as (...){0} was completely
264        omitted from the compiled regex. However, this means that if the group
265        was called as a subroutine from elsewhere in the pattern, things went wrong
266        (an internal error was given). Such groups are now left in the compiled
267        pattern, with a new opcode that causes them to be skipped at execution
268        time.
269    
270    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
271        to the way PCRE behaves:
272    
273        (a) A lone ] character is dis-allowed (Perl treats it as data).
274    
275        (b) A back reference to an unmatched subpattern matches an empty string
276            (Perl fails the current match path).
277    
278        (c) A data ] in a character class must be notated as \] because if the
279            first data character in a class is ], it defines an empty class. (In
280            Perl it is not possible to have an empty class.) The empty class []
281            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
282            The negative empty class [^] matches any one character, independently
283            of the DOTALL setting.
284    
285    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
286        non-existent subpattern following a character class starting with ']' and
287        containing () gave an internal compiling error instead of "reference to
288        non-existent subpattern". Fortunately, when the pattern did exist, the
289        compiled code was correct. (When scanning forwards to check for the
290        existencd of the subpattern, it was treating the data ']' as terminating
291        the class, so got the count wrong. When actually compiling, the reference
292        was subsequently set up correctly.)
293    
294    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
295        it was being rejected as not supported by pcre_dfa_exec(), even though
296        other assertions are supported. I have made pcre_dfa_exec() support
297        (*FAIL).
298    
299    16. The implementation of 13c above involved the invention of a new opcode,
300        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
301        cannot be changed at match time, I realized I could make a small
302        improvement to matching performance by compiling OP_ALLANY instead of
303        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
304        on the OP_ANY path.
305    
306    17. Compiling pcretest on Windows with readline support failed without the
307        following two fixes: (1) Make the unistd.h include conditional on
308        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
309    
310    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
311        ncurses library to be included for pcretest when ReadLine support is
312        requested, but also to allow for it to be overridden. This patch came from
313        Daniel Bergström.
314    
315    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
316        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
317        any errors with the current Unicode tables. Thanks to Peter Kankowski for
318        spotting this.
319    
320    
321    Version 7.6 28-Jan-08
322    ---------------------
323    
324    1.  A character class containing a very large number of characters with
325        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
326        overflow.
327    
328    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
329        HAVE_LONG_LONG is not defined.
330    
331    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
332        bring them up to date and include new features. This patch includes:
333    
334        - Fixed PH's badly added libz and libbz2 support.
335        - Fixed a problem with static linking.
336        - Added pcredemo. [But later removed - see 7 below.]
337        - Fixed dftables problem and added an option.
338        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
339            HAVE_LONG_LONG.
340        - Added readline support for pcretest.
341        - Added an listing of the option settings after cmake has run.
342    
343    4.  A user submitted a patch to Makefile that makes it easy to create
344        "pcre.dll" under mingw when using Configure/Make. I added stuff to
345        Makefile.am that cause it to include this special target, without
346        affecting anything else. Note that the same mingw target plus all
347        the other distribution libraries and programs are now supported
348        when configuring with CMake (see 6 below) instead of with
349        Configure/Make.
350    
351    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
352        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
353        exported in the Windows port". It has not yet been confirmed that the patch
354        solves the problem, but it does no harm.
355    
356    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
357        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
358        with CMake, and also correct the comment about stack recursion.
359    
360    7.  Remove the automatic building of pcredemo from the ./configure system and
361        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
362        of a program that users should build themselves after PCRE is installed, so
363        building it automatically is not really right. What is more, it gave
364        trouble in some build environments.
365    
366    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
367    
368    
369    Version 7.5 10-Jan-08
370    ---------------------
371    
372    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
373        values in parens when parsing an RE using the C++ wrapper."
374    
375    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
376        Characters greater than 255 were excluded from the class instead of being
377        included.
378    
379    3.  The same bug as (2) above applied to negated POSIX classes such as
380        [:^space:].
381    
382    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
383        defined or documented. It seems to have been a typo for PCRE_STATIC, so
384        I have changed it.
385    
386    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
387        first named subpattern) and a construct such as (?&a) would reference the
388        first named subpattern whose name started with "a" (in other words, the
389        length check was missing). Both these problems are fixed. "Subpattern name
390        expected" is now given for (?&) (a zero-length name), and this patch also
391        makes it give the same error for \k'' (previously it complained that that
392        was a reference to a non-existent subpattern).
393    
394    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
395        this is right because (?- can be followed by option settings as well as by
396        digits. I have, however, made the messages clearer.
397    
398    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
399        than the number used in the conditional) now cause a compile-time error.
400        This is actually not compatible with Perl, which accepts such patterns, but
401        treats the conditional as always being FALSE (as PCRE used to), but it
402        seems to me that giving a diagnostic is better.
403    
404    8.  Change "alphameric" to the more common word "alphanumeric" in comments
405        and messages.
406    
407    9.  Fix two occurrences of "backslash" in comments that should have been
408        "backspace".
409    
410    10. Remove two redundant lines of code that can never be obeyed (their function
411        was moved elsewhere).
412    
413    11. The program that makes PCRE's Unicode character property table had a bug
414        which caused it to generate incorrect table entries for sequences of
415        characters that have the same character type, but are in different scripts.
416        It amalgamated them into a single range, with the script of the first of
417        them. In other words, some characters were in the wrong script. There were
418        thirteen such cases, affecting characters in the following ranges:
419    
420          U+002b0 - U+002c1
421          U+0060c - U+0060d
422          U+0061e - U+00612
423          U+0064b - U+0065e
424          U+0074d - U+0076d
425          U+01800 - U+01805
426          U+01d00 - U+01d77
427          U+01d9b - U+01dbf
428          U+0200b - U+0200f
429          U+030fc - U+030fe
430          U+03260 - U+0327f
431          U+0fb46 - U+0fbb1
432          U+10450 - U+1049d
433    
434    12. The -o option (show only the matching part of a line) for pcregrep was not
435        compatible with GNU grep in that, if there was more than one match in a
436        line, it showed only the first of them. It now behaves in the same way as
437        GNU grep.
438    
439    13. If the -o and -v options were combined for pcregrep, it printed a blank
440        line for every non-matching line. GNU grep prints nothing, and pcregrep now
441        does the same. The return code can be used to tell if there were any
442        non-matching lines.
443    
444    14. Added --file-offsets and --line-offsets to pcregrep.
445    
446    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
447        infinitely looping recursion. The bug was that positive lookaheads were not
448        being skipped when checking for a possible empty match (negative lookaheads
449        and both kinds of lookbehind were skipped).
450    
451    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
452        inclusion of <windows.h> to before rather than after the definition of
453        INVALID_FILE_ATTRIBUTES (patch from David Byron).
454    
455    17. Specifying a possessive quantifier with a specific limit for a Unicode
456        character property caused pcre_compile() to compile bad code, which led at
457        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
458        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
459        caused the error; without that there was no problem.
460    
461    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
462    
463    19. Added --enable-pcretest-libreadline.
464    
465    20. In pcrecpp.cc, the variable 'count' was incremented twice in
466        RE::GlobalReplace(). As a result, the number of replacements returned was
467        double what it should be. I removed one of the increments, but Craig sent a
468        later patch that removed the other one (the right fix) and added unit tests
469        that check the return values (which was not done before).
470    
471    21. Several CMake things:
472    
473        (1) Arranged that, when cmake is used on Unix, the libraries end up with
474            the names libpcre and libpcreposix, not just pcre and pcreposix.
475    
476        (2) The above change means that pcretest and pcregrep are now correctly
477            linked with the newly-built libraries, not previously installed ones.
478    
479        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
480    
481    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
482        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
483        UTF-8 newline character). The key issue is that the pattern starts .*;
484        this means that the match must be either at the beginning, or after a
485        newline. The bug was in the code for advancing after a failed match and
486        checking that the new position followed a newline. It was not taking
487        account of UTF-8 characters correctly.
488    
489    23. PCRE was behaving differently from Perl in the way it recognized POSIX
490        character classes. PCRE was not treating the sequence [:...:] as a
491        character class unless the ... were all letters. Perl, however, seems to
492        allow any characters between [: and :], though of course it rejects as
493        unknown any "names" that contain non-letters, because all the known class
494        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
495        for example, whereas PCRE did not - it did not recognize a POSIX character
496        class. This seemed a bit dangerous, so the code has been changed to be
497        closer to Perl. The behaviour is not identical to Perl, because PCRE will
498        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
499        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
500        Perl does, and where it didn't before.
501    
502    24. Rewrite so as to remove the single use of %n from pcregrep because in some
503        Windows environments %n is disabled by default.
504    
505    
506    Version 7.4 21-Sep-07
507    ---------------------
508    
509    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
510        means that a class such as [\s] counted as "explicit reference to CR or
511        LF". That isn't really right - the whole point of the change was to try to
512        help when there was an actual mention of one of the two characters. So now
513        the change happens only if \r or \n (or a literal CR or LF) character is
514        encountered.
515    
516    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
517        of both had grown to the point where there were only 3 bits left.
518        Fortunately, there was spare space in the data structure, and so I have
519        moved the internal flags into a new 16-bit field to free up more option
520        bits.
521    
522    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
523        but did not set the internal JCHANGED flag - either of these is enough to
524        control the way the "get" function works - but the PCRE_INFO_JCHANGED
525        facility is supposed to tell if (?J) was ever used, so now (?J) at the
526        start sets both bits.
527    
528    4.  Added options (at build time, compile time, exec time) to change \R from
529        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
530    
531    5.  doc/pcresyntax.html was missing from the distribution.
532    
533    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
534        compatibility, even though it is no longer used.
535    
536    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
537        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
538        windows.h file is present (where different names are used). [This was
539        reversed later after testing - see 16 below.]
540    
541    8.  Changed all #include <config.h> to #include "config.h". There were also
542        some further <pcre.h> cases that I changed to "pcre.h".
543    
544    9.  When pcregrep was used with the --colour option, it missed the line ending
545        sequence off the lines that it output.
546    
547    10. It was pointed out to me that arrays of string pointers cause lots of
548        relocations when a shared library is dynamically loaded. A technique of
549        using a single long string with a table of offsets can drastically reduce
550        these. I have refactored PCRE in four places to do this. The result is
551        dramatic:
552    
553          Originally:                          290
554          After changing UCP table:            187
555          After changing error message table:   43
556          After changing table of "verbs"       36
557          After changing table of Posix names   22
558    
559        Thanks to the folks working on Gregex for glib for this insight.
560    
561    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
562        unicode-properties was also set.
563    
564    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
565    
566    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
567        checked only for CRLF.
568    
569    14. Added casts to pcretest.c to avoid compiler warnings.
570    
571    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
572    
573    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
574        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
575        entirely. This removes changes made in 7 above.
576    
577    17. The CMake files have been updated, and there is now more information about
578        building with CMake in the NON-UNIX-USE document.
579    
580    
581    Version 7.3 28-Aug-07
582    ---------------------
583    
584     1. In the rejigging of the build system that eventually resulted in 7.1, the
585        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
586        brackets there is not right, since it causes compilers to look for an
587        installed pcre.h, not the version that is in the source that is being
588        compiled (which of course may be different). I have changed it back to:
589    
590          #include "pcre.h"
591    
592        I have a vague recollection that the change was concerned with compiling in
593        different directories, but in the new build system, that is taken care of
594        by the VPATH setting the Makefile.
595    
596     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
597        when the subject happened to end in the byte 0x85 (e.g. if the last
598        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
599        characters but of course it shouldn't be taken as a newline when it is part
600        of another character. The bug was that, for an unlimited repeat of . in
601        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
602        characters when looking for a newline.
603    
604     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
605    
606     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
607        in debug output.
608    
609     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
610        long printing in the pcrecpp unittest when running under MinGW.
611    
612     6. ESC_K was left out of the EBCDIC table.
613    
614     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
615        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
616        limit also applies to "virtual nesting" when a pattern is recursive, and in
617        this case 1000 isn't so big. I have been able to remove this limit at the
618        expense of backing off one optimization in certain circumstances. Normally,
619        when pcre_exec() would call its internal match() function recursively and
620        immediately return the result unconditionally, it uses a "tail recursion"
621        feature to save stack. However, when a subpattern that can match an empty
622        string has an unlimited repetition quantifier, it no longer makes this
623        optimization. That gives it a stack frame in which to save the data for
624        checking that an empty string has been matched. Previously this was taken
625        from the 1000-entry workspace that had been reserved. So now there is no
626        explicit limit, but more stack is used.
627    
628     8. Applied Daniel's patches to solve problems with the import/export magic
629        syntax that is required for Windows, and which was going wrong for the
630        pcreposix and pcrecpp parts of the library. These were overlooked when this
631        problem was solved for the main library.
632    
633     9. There were some crude static tests to avoid integer overflow when computing
634        the size of patterns that contain repeated groups with explicit upper
635        limits. As the maximum quantifier is 65535, the maximum group length was
636        set at 30,000 so that the product of these two numbers did not overflow a
637        32-bit integer. However, it turns out that people want to use groups that
638        are longer than 30,000 bytes (though not repeat them that many times).
639        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
640        made it possible to implement the integer overflow checks in a much more
641        dynamic way, which I have now done. The artificial limitation on group
642        length has been removed - we now have only the limit on the total length of
643        the compiled pattern, which depends on the LINK_SIZE setting.
644    
645    10. Fixed a bug in the documentation for get/copy named substring when
646        duplicate names are permitted. If none of the named substrings are set, the
647        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
648        empty string.
649    
650    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
651        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
652        because the ] is interpreted as the first data character and the
653        terminating ] is not found. PCRE has been made compatible with Perl in this
654        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
655        cause memory overwriting.
656    
657    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
658        string has been matched (to stop an infinite loop). It was not recognizing
659        a conditional subpattern that could match an empty string if that
660        subpattern was within another subpattern. For example, it looped when
661        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
662        condition was not nested. This bug has been fixed.
663    
664    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
665        past the start of the subject in the presence of bytes with the top bit
666        set, for example "\x8aBCD".
667    
668    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
669        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
670    
671    14. Optimized (?!) to (*FAIL).
672    
673    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
674        This restricts code points to be within the range 0 to 0x10FFFF, excluding
675        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
676        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
677        does: it's just the validity check that is more restrictive.
678    
679    16. Inserted checks for integer overflows during escape sequence (backslash)
680        processing, and also fixed erroneous offset values for syntax errors during
681        backslash processing.
682    
683    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
684        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
685    
686    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
687        caused an overrun.
688    
689    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
690        something other than just ASCII characters) inside a group that had an
691        unlimited repeat caused a loop at compile time (while checking to see
692        whether the group could match an empty string).
693    
694    20. Debugging a pattern containing \p or \P could cause a crash. For example,
695        [\P{Any}] did so. (Error in the code for printing property names.)
696    
697    21. An orphan \E inside a character class could cause a crash.
698    
699    22. A repeated capturing bracket such as (A)? could cause a wild memory
700        reference during compilation.
701    
702    23. There are several functions in pcre_compile() that scan along a compiled
703        expression for various reasons (e.g. to see if it's fixed length for look
704        behind). There were bugs in these functions when a repeated \p or \P was
705        present in the pattern. These operators have additional parameters compared
706        with \d, etc, and these were not being taken into account when moving along
707        the compiled data. Specifically:
708    
709        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
710            length.
711    
712        (b) An item such as \pL+ within a repeated group could cause crashes or
713            loops.
714    
715        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
716            "reference to non-existent subpattern" error.
717    
718        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
719    
720    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
721        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
722    
723    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
724    
725    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
726        character were causing crashes (broken optimization).
727    
728    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
729        \p or \P) caused a compile-time loop.
730    
731    28. More problems have arisen in unanchored patterns when CRLF is a valid line
732        break. For example, the unstudied pattern [\r\n]A does not match the string
733        "\r\nA" because change 7.0/46 below moves the current point on by two
734        characters after failing to match at the start. However, the pattern \nA
735        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
736        the same is true. There doesn't seem any very clean way out of this, but
737        what I have chosen to do makes the common cases work: PCRE now takes note
738        of whether there can be an explicit match for \r or \n anywhere in the
739        pattern, and if so, 7.0/46 no longer applies. As part of this change,
740        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
741        pattern has explicit CR or LF references.
742    
743    29. Added (*CR) etc for changing newline setting at start of pattern.
744    
745    
746    Version 7.2 19-Jun-07
747    ---------------------
748    
749     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
750        which is apparently normally available under Windows.
751    
752     2. Re-jig the pcregrep tests with different newline settings in an attempt
753        to make them independent of the local environment's newline setting.
754    
755     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
756    
757     4. Some of the "internals" tests were previously cut out when the link size
758        was not 2, because the output contained actual offsets. The recent new
759        "Z" feature of pcretest means that these can be cut out, making the tests
760        usable with all link sizes.
761    
762     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
763        stack recursion. This gives a massive performance boost under BSD, but just
764        a small improvement under Linux. However, it saves one field in the frame
765        in all cases.
766    
767     6. Added more features from the forthcoming Perl 5.10:
768    
769        (a) (?-n) (where n is a string of digits) is a relative subroutine or
770            recursion call. It refers to the nth most recently opened parentheses.
771    
772        (b) (?+n) is also a relative subroutine call; it refers to the nth next
773            to be opened parentheses.
774    
775        (c) Conditions that refer to capturing parentheses can be specified
776            relatively, for example, (?(-2)... or (?(+3)...
777    
778        (d) \K resets the start of the current match so that everything before
779            is not part of it.
780    
781        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
782    
783        (f) \g{name} is another synonym - part of Perl 5.10's unification of
784            reference syntax.
785    
786        (g) (?| introduces a group in which the numbering of parentheses in each
787            alternative starts with the same number.
788    
789        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
790    
791     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
792        PCRE_INFO_JCHANGED.
793    
794     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
795        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
796        for detecting groups that can match an empty string.
797    
798     9. A pattern with a very large number of alternatives (more than several
799        hundred) was running out of internal workspace during the pre-compile
800        phase, where pcre_compile() figures out how much memory will be needed. A
801        bit of new cunning has reduced the workspace needed for groups with
802        alternatives. The 1000-alternative test pattern now uses 12 bytes of
803        workspace instead of running out of the 4096 that are available.
804    
805    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
806    
807    11. Applied patch from Google to remove an optimization that didn't quite work.
808        The report of the bug said:
809    
810          pcrecpp::RE("a*").FullMatch("aaa") matches, while
811          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
812          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
813    
814    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
815        it matched the wrong number of bytes.
816    
817    
818    Version 7.1 24-Apr-07
819    ---------------------
820    
821     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
822        that is more "standard", making use of automake and other Autotools. There
823        is some re-arrangement of the files and adjustment of comments consequent
824        on this.
825    
826     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
827        for recursive directory scanning broke on some systems because the files
828        are not scanned in any specific order and on different systems the order
829        was different. A call to "sort" has been inserted into RunGrepTest for the
830        approprate test as a short-term fix. In the longer term there may be an
831        alternative.
832    
833     3. I had an email from Eric Raymond about problems translating some of PCRE's
834        man pages to HTML (despite the fact that I distribute HTML pages, some
835        people do their own conversions for various reasons). The problems
836        concerned the use of low-level troff macros .br and .in. I have therefore
837        removed all such uses from the man pages (some were redundant, some could
838        be replaced by .nf/.fi pairs). The 132html script that I use to generate
839        HTML has been updated to handle .nf/.fi and to complain if it encounters
840        .br or .in.
841    
842     4. Updated comments in configure.ac that get placed in config.h.in and also
843        arranged for config.h to be included in the distribution, with the name
844        config.h.generic, for the benefit of those who have to compile without
845        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
846    
847     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
848        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
849        makevp.bat for the current PCRE, using the additional files
850        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
851    
852     6. A Windows user reported a minor discrepancy with test 2, which turned out
853        to be caused by a trailing space on an input line that had got lost in his
854        copy. The trailing space was an accident, so I've just removed it.
855    
856     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
857        that is needed.
858    
859     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
860        as "const" (a) because they are and (b) because it helps the PHP
861        maintainers who have recently made a script to detect big data structures
862        in the php code that should be moved to the .rodata section. I remembered
863        to update Builducptable as well, so it won't revert if ucptable.h is ever
864        re-created.
865    
866     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
867        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
868        order to be able to cut out the UTF-8 tables in the latter when UTF-8
869        support is not required. This saves 1.5-2K of code, which is important in
870        some applications.
871    
872        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
873        so as not to refer to the tables, even though these functions will never be
874        called when UTF-8 support is disabled. Otherwise there are problems with a
875        shared library.
876    
877    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
878    
879        (a) It was defining its arguments as char * instead of void *.
880    
881        (b) It was assuming that all moves were upwards in memory; this was true
882            a long time ago when I wrote it, but is no longer the case.
883    
884        The emulated memove() is provided for those environments that have neither
885        memmove() nor bcopy(). I didn't think anyone used it these days, but that
886        is clearly not the case, as these two bugs were recently reported.
887    
888    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
889        and Detrail to create the HTML documentation, the .txt form of the man
890        pages, and it removes trailing spaces from listed files. It also creates
891        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
892        case, it wraps all the #defines with #ifndefs. This script should be run
893        before "make dist".
894    
895    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
896        with Unicode property support.
897    
898        (a) For a maximizing quantifier, if the two different cases of the
899            character were of different lengths in their UTF-8 codings (there are
900            some cases like this - I found 11), and the matching function had to
901            back up over a mixture of the two cases, it incorrectly assumed they
902            were both the same length.
903    
904        (b) When PCRE was configured to use the heap rather than the stack for
905            recursion during matching, it was not correctly preserving the data for
906            the other case of a UTF-8 character when checking ahead for a match
907            while processing a minimizing repeat. If the check also involved
908            matching a wide character, but failed, corruption could cause an
909            erroneous result when trying to check for a repeat of the original
910            character.
911    
912    13. Some tidying changes to the testing mechanism:
913    
914        (a) The RunTest script now detects the internal link size and whether there
915            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
916            values substituted by "configure". (The RunGrepTest script already did
917            this for UTF-8.) The configure.ac script no longer substitutes the
918            relevant variables.
919    
920        (b) The debugging options /B and /D in pcretest show the compiled bytecode
921            with length and offset values. This means that the output is different
922            for different internal link sizes. Test 2 is skipped for link sizes
923            other than 2 because of this, bypassing the problem. Unfortunately,
924            there was also a test in test 3 (the locale tests) that used /B and
925            failed for link sizes other than 2. Rather than cut the whole test out,
926            I have added a new /Z option to pcretest that replaces the length and
927            offset values with spaces. This is now used to make test 3 independent
928            of link size. (Test 2 will be tidied up later.)
929    
930    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
931        segmentation fault instead of returning the appropriate error message.
932    
933    15. In multiline mode when the newline sequence was set to "any", the pattern
934        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
935        This doesn't seem right; it now treats the CRLF combination as the line
936        ending, and so does not match in that case. It's only a pattern such as ^$
937        that would hit this one: something like ^ABC$ would have failed after \r
938        and then tried again after \r\n.
939    
940    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
941        in an attempt to make files that differ only in their line terminators
942        compare equal. This works on Linux.
943    
944    17. Under certain error circumstances pcregrep might try to free random memory
945        as it exited. This is now fixed, thanks to valgrind.
946    
947    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
948        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
949        was because its rules for how to advance for /g after matching an empty
950        string at the end of a line did not allow for this case. They now check for
951        it specially.
952    
953    20. pcretest is supposed to handle patterns and data of any length, by
954        extending its buffers when necessary. It was getting this wrong when the
955        buffer for a data line had to be extended.
956    
957    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
958        CRLF as a newline sequence.
959    
960    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
961        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
962        I have nevertheless tidied it up.
963    
964    23. Added some casts to kill warnings from HP-UX ia64 compiler.
965    
966    24. Added a man page for pcre-config.
967    
968    
969    Version 7.0 19-Dec-06
970    ---------------------
971    
972     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
973        moving to gcc 4.1.1.
974    
975     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
976        sys/time.h, which is documented as needed for this function. It doesn't
977        seem to matter on Linux, but it showed up on some releases of OS X.
978    
979     3. It seems that there are systems where bytes whose values are greater than
980        127 match isprint() in the "C" locale. The "C" locale should be the
981        default when a C program starts up. In most systems, only ASCII printing
982        characters match isprint(). This difference caused the output from pcretest
983        to vary, making some of the tests fail. I have changed pcretest so that:
984    
985        (a) When it is outputting text in the compiled version of a pattern, bytes
986            other than 32-126 are always shown as hex escapes.
987    
988        (b) When it is outputting text that is a matched part of a subject string,
989            it does the same, unless a different locale has been set for the match
990            (using the /L modifier). In this case, it uses isprint() to decide.
991    
992     4. Fixed a major bug that caused incorrect computation of the amount of memory
993        required for a compiled pattern when options that changed within the
994        pattern affected the logic of the preliminary scan that determines the
995        length. The relevant options are -x, and -i in UTF-8 mode. The result was
996        that the computed length was too small. The symptoms of this bug were
997        either the PCRE error "internal error: code overflow" from pcre_compile(),
998        or a glibc crash with a message such as "pcretest: free(): invalid next
999        size (fast)". Examples of patterns that provoked this bug (shown in
1000        pcretest format) are:
1001    
1002          /(?-x: )/x
1003          /(?x)(?-x: \s*#\s*)/
1004          /((?i)[\x{c0}])/8
1005          /(?i:[\x{c0}])/8
1006    
1007        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1008        is now done differently.
1009    
1010     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1011        wrapper classes; (b) implement a new function in the C++ scanner that is
1012        more efficient than the old way of doing things because it avoids levels of
1013        recursion in the regex matching; (c) add a paragraph to the documentation
1014        for the FullMatch() function.
1015    
1016     6. The escape sequence \n was being treated as whatever was defined as
1017        "newline". Not only was this contrary to the documentation, which states
1018        that \n is character 10 (hex 0A), but it also went horribly wrong when
1019        "newline" was defined as CRLF. This has been fixed.
1020    
1021     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1022        was being set to -1 for the "end of line" case (supposedly a value that no
1023        character can have). Though this value is never used (the check for end of
1024        line is "zero bytes in current character"), it caused compiler complaints.
1025        I've changed it to 0xffffffff.
1026    
1027     8. In pcre_version.c, the version string was being built by a sequence of
1028        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1029        string (as it is for production releases) called a macro with an empty
1030        argument. The C standard says the result of this is undefined. The gcc
1031        compiler treats it as an empty string (which was what was wanted) but it is
1032        reported that Visual C gives an error. The source has been hacked around to
1033        avoid this problem.
1034    
1035     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1036        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1037        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1038        of them did).
1039    
1040    10. Originally, pcretest opened its input and output without "b"; then I was
1041        told that "b" was needed in some environments, so it was added for release
1042        5.0 to both the input and output. (It makes no difference on Unix-like
1043        systems.) Later I was told that it is wrong for the input on Windows. I've
1044        now abstracted the modes into two macros, to make it easier to fiddle with
1045        them, and removed "b" from the input mode under Windows.
1046    
1047    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1048    
1049    12. Added -help and --help to pcretest as an official way of being reminded
1050        of the options.
1051    
1052    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1053        and pcrecpp.cc because they annoy compilers at high warning levels.
1054    
1055    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1056    
1057    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1058        scripts are not C programs :-) and which was not noticed because it works
1059        on Linux.
1060    
1061    16. pcretest is supposed to handle any length of pattern and data line (as one
1062        line or as a continued sequence of lines) by extending its input buffer if
1063        necessary. This feature was broken for very long pattern lines, leading to
1064        a string of junk being passed to pcre_compile() if the pattern was longer
1065        than about 50K.
1066    
1067    17. I have done a major re-factoring of the way pcre_compile() computes the
1068        amount of memory needed for a compiled pattern. Previously, there was code
1069        that made a preliminary scan of the pattern in order to do this. That was
1070        OK when PCRE was new, but as the facilities have expanded, it has become
1071        harder and harder to keep it in step with the real compile phase, and there
1072        have been a number of bugs (see for example, 4 above). I have now found a
1073        cunning way of running the real compile function in a "fake" mode that
1074        enables it to compute how much memory it would need, while actually only
1075        ever using a few hundred bytes of working memory and without too many
1076        tests of the mode. This should make future maintenance and development
1077        easier. A side effect of this work is that the limit of 200 on the nesting
1078        depth of parentheses has been removed (though this was never a serious
1079        limitation, I suspect). However, there is a downside: pcre_compile() now
1080        runs more slowly than before (30% or more, depending on the pattern). I
1081        hope this isn't a big issue. There is no effect on runtime performance.
1082    
1083    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1084        newline (only possible for the last line of a file) and it was a
1085        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1086    
1087    19. Added additional timing features to pcretest. (1) The -tm option now times
1088        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1089        separate command line item, by a number that specifies the number of
1090        repeats to use when timing. The default is 50000; this gives better
1091        precision, but takes uncomfortably long for very large patterns.
1092    
1093    20. Extended pcre_study() to be more clever in cases where a branch of a
1094        subpattern has no definite first character. For example, (a*|b*)[cd] would
1095        previously give no result from pcre_study(). Now it recognizes that the
1096        first character must be a, b, c, or d.
1097    
1098    21. There was an incorrect error "recursive call could loop indefinitely" if
1099        a subpattern (or the entire pattern) that was being tested for matching an
1100        empty string contained only one non-empty item after a nested subpattern.
1101        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1102        incorrectly, because the \d was being skipped in the check.
1103    
1104    22. The pcretest program now has a new pattern option /B and a command line
1105        option -b, which is equivalent to adding /B to every pattern. This causes
1106        it to show the compiled bytecode, without the additional information that
1107        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1108        is the same as /B/I).
1109    
1110    23. A new optimization is now able automatically to treat some sequences such
1111        as a*b as a*+b. More specifically, if something simple (such as a character
1112        or a simple class like \d) has an unlimited quantifier, and is followed by
1113        something that cannot possibly match the quantified thing, the quantifier
1114        is automatically "possessified".
1115    
1116    24. A recursive reference to a subpattern whose number was greater than 39
1117        went wrong under certain circumstances in UTF-8 mode. This bug could also
1118        have affected the operation of pcre_study().
1119    
1120    25. Realized that a little bit of performance could be had by replacing
1121        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1122    
1123    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1124    
1125    27. Possessive quantifiers such as a++ were previously implemented by turning
1126        them into atomic groups such as ($>a+). Now they have their own opcodes,
1127        which improves performance. This includes the automatically created ones
1128        from 23 above.
1129    
1130    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1131        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1132        the first matched character to be a colon. This applied both to named and
1133        numbered groups.
1134    
1135    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1136    
1137    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1138        building PCRE on Windows easier, so I have included it in the distribution.
1139    
1140    31. There is now a check in pcretest against a ridiculously large number being
1141        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1142        loop, the loop is abandoned.
1143    
1144    32. Forward references to subpatterns in conditions such as (?(2)...) where
1145        subpattern 2 is defined later cause pcre_compile() to search forwards in
1146        the pattern for the relevant set of parentheses. This search went wrong
1147        when there were unescaped parentheses in a character class, parentheses
1148        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1149    
1150    33. "Subroutine" calls and backreferences were previously restricted to
1151        referencing subpatterns earlier in the regex. This restriction has now
1152        been removed.
1153    
1154    34. Added a number of extra features that are going to be in Perl 5.10. On the
1155        whole, these are just syntactic alternatives for features that PCRE had
1156        previously implemented using the Python syntax or my own invention. The
1157        other formats are all retained for compatibility.
1158    
1159        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1160            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1161            also .NET compatible.
1162    
1163        (b) A recursion or subroutine call to a named group can now be defined as
1164            (?&name) as well as (?P>name).
1165    
1166        (c) A backreference to a named group can now be defined as \k<name> or
1167            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1168            5.10, are also .NET compatible.
1169    
1170        (d) A conditional reference to a named group can now use the syntax
1171            (?(<name>) or (?('name') as well as (?(name).
1172    
1173        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1174            groups (named and numbered) that are never evaluated inline, but can be
1175            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1176            is always false. There may be only one alternative in such a group.
1177    
1178        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1179            as the simple (?(R). The condition is true only if the most recent
1180            recursion is that of the given number or name. It does not search out
1181            through the entire recursion stack.
1182    
1183        (g) The escape \gN or \g{N} has been added, where N is a positive or
1184            negative number, specifying an absolute or relative reference.
1185    
1186    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1187        some "unreachable code" warnings.
1188    
1189    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1190        things, this adds five new scripts.
1191    
1192    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1193        There were also incompatibilities regarding the handling of \Q..\E inside
1194        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1195        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1196    
1197    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1198        matches an empty string, and forcibly breaks the loop. There were bugs in
1199        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1200        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1201        separate and independent bugs (that affected different cases) have been
1202        fixed.
1203    
1204    39. Refactored the code to abolish the use of different opcodes for small
1205        capturing bracket numbers. This is a tidy that I avoided doing when I
1206        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1207        The new approach is not only tidier, it makes it possible to reduce the
1208        memory needed to fix the previous bug (38).
1209    
1210    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1211        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1212        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1213        mode.
1214    
1215    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1216        report.
1217    
1218    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1219        copy construction and assignment in the C++ wrapper.
1220    
1221    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1222        couple of bugs that could have given wrong results in the "--newline=crlf"
1223        case.
1224    
1225    44. Added a number of casts and did some reorganization of signed/unsigned int
1226        variables following suggestions from Dair Grant. Also renamed the variable
1227        "this" as "item" because it is a C++ keyword.
1228    
1229    45. Arranged for dftables to add
1230    
1231          #include "pcre_internal.h"
1232    
1233        to pcre_chartables.c because without it, gcc 4.x may remove the array
1234        definition from the final binary if PCRE is built into a static library and
1235        dead code stripping is activated.
1236    
1237    46. For an unanchored pattern, if a match attempt fails at the start of a
1238        newline sequence, and the newline setting is CRLF or ANY, and the next two
1239        characters are CRLF, advance by two characters instead of one.
1240    
1241    
1242    Version 6.7 04-Jul-06
1243    ---------------------
1244    
1245     1. In order to handle tests when input lines are enormously long, pcretest has
1246        been re-factored so that it automatically extends its buffers when
1247        necessary. The code is crude, but this _is_ just a test program. The
1248        default size has been increased from 32K to 50K.
1249    
1250     2. The code in pcre_study() was using the value of the re argument before
1251        testing it for NULL. (Of course, in any sensible call of the function, it
1252        won't be NULL.)
1253    
1254     3. The memmove() emulation function in pcre_internal.h, which is used on
1255        systems that lack both memmove() and bcopy() - that is, hardly ever -
1256        was missing a "static" storage class specifier.
1257    
1258     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1259        containing an extended class (one that cannot be represented by a bitmap
1260        because it contains high-valued characters or Unicode property items, e.g.
1261        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1262        pattern, but PCRE should not loop if you do not (it no longer does).
1263        [Detail: two cases were found: (a) a repeated subpattern containing an
1264        extended class; (b) a recursive reference to a subpattern that followed a
1265        previous extended class. It wasn't skipping over the extended class
1266        correctly when UTF-8 mode was not set.]
1267    
1268     5. A negated single-character class was not being recognized as fixed-length
1269        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1270        compile error "lookbehind assertion is not fixed length".
1271    
1272     6. The RunPerlTest auxiliary script was showing an unexpected difference
1273        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1274        write a Perl script that can interpret lines of an input file either as
1275        byte characters or as UTF-8, which is what "perltest" was being required to
1276        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1277        can't do is switch easily at run time between having the "use utf8;" pragma
1278        or not. In the end, I fudged it by using the RunPerlTest script to insert
1279        "use utf8;" explicitly for the UTF-8 tests.
1280    
1281     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1282        the end of the subject string, contrary to the documentation and to what
1283        Perl does. This was true of both matching functions. Now it matches only at
1284        the start of the subject and immediately after *internal* newlines.
1285    
1286     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1287        a pointer to an int instead of a pointer to an unsigned long int. This
1288        caused problems on 64-bit systems.
1289    
1290     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1291        instance of the 'standard' template library not being so standard".
1292    
1293    10. There was no check on the number of named subpatterns nor the maximum
1294        length of a subpattern name. The product of these values is used to compute
1295        the size of the memory block for a compiled pattern. By supplying a very
1296        long subpattern name and a large number of named subpatterns, the size
1297        computation could be caused to overflow. This is now prevented by limiting
1298        the length of names to 32 characters, and the number of named subpatterns
1299        to 10,000.
1300    
1301    11. Subpatterns that are repeated with specific counts have to be replicated in
1302        the compiled pattern. The size of memory for this was computed from the
1303        length of the subpattern and the repeat count. The latter is limited to
1304        65535, but there was no limit on the former, meaning that integer overflow
1305        could in principle occur. The compiled length of a repeated subpattern is
1306        now limited to 30,000 bytes in order to prevent this.
1307    
1308    12. Added the optional facility to have named substrings with the same name.
1309    
1310    13. Added the ability to use a named substring as a condition, using the
1311        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1312        are numbers (not recommended). Forward references are permitted.
1313    
1314    14. Added forward references in named backreferences (if you see what I mean).
1315    
1316    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1317        pattern could run off the end of the subject. For example, the pattern
1318        "(?s)(.{1,5})"8 did this with the subject "ab".
1319    
1320    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1321        PCRE_CASELESS was set when matching characters that were quantified with ?
1322        or *.
1323    
1324    17. A character class other than a single negated character that had a minimum
1325        but no maximum quantifier - for example [ab]{6,} - was not handled
1326        correctly by pce_dfa_exec(). It would match only one character.
1327    
1328    18. A valid (though odd) pattern that looked like a POSIX character
1329        class but used an invalid character after [ (for example [[,abc,]]) caused
1330        pcre_compile() to give the error "Failed: internal error: code overflow" or
1331        in some cases to crash with a glibc free() error. This could even happen if
1332        the pattern terminated after [[ but there just happened to be a sequence of
1333        letters, a binary zero, and a closing ] in the memory that followed.
1334    
1335    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1336        over the years. Originally (before any Unicode support), just the bottom 8
1337        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1338        output from "man perlunicode" includes this:
1339    
1340          The regular expression compiler produces polymorphic opcodes.  That
1341          is, the pattern adapts to the data and automatically switches to
1342          the Unicode character scheme when presented with Unicode data--or
1343          instead uses a traditional byte scheme when presented with byte
1344          data.
1345    
1346        Sadly, a wide octal escape does not cause a switch, and in a string with
1347        no other multibyte characters, these octal escapes are treated as before.
1348        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1349        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1350        Unicode string.
1351    
1352        I have not perpetrated such confusion in PCRE. Up till now, it took just
1353        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1354        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1355        translate to the appropriate multibyte character.
1356    
1357    29. Applied some refactoring to reduce the number of warnings from Microsoft
1358        and Borland compilers. This has included removing the fudge introduced
1359        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1360        a warning about an unused variable.
1361    
1362    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1363        characters since release 4.0, because Perl (from release 5.004) does not.
1364        [Or at least, is documented not to: some releases seem to be in conflict
1365        with the documentation.] However, when a pattern was studied with
1366        pcre_study() and all its branches started with \s, PCRE still included VT
1367        as a possible starting character. Of course, this did no harm; it just
1368        caused an unnecessary match attempt.
1369    
1370    22. Removed a now-redundant internal flag bit that recorded the fact that case
1371        dependency changed within the pattern. This was once needed for "required
1372        byte" processing, but is no longer used. This recovers a now-scarce options
1373        bit. Also moved the least significant internal flag bit to the most-
1374        significant bit of the word, which was not previously used (hangover from
1375        the days when it was an int rather than a uint) to free up another bit for
1376        the future.
1377    
1378    23. Added support for CRLF line endings as well as CR and LF. As well as the
1379        default being selectable at build time, it can now be changed at runtime
1380        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1381        specify that it is scanning data with non-default line endings.
1382    
1383    24. Changed the definition of CXXLINK to make it agree with the definition of
1384        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1385    
1386    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1387        recursions. This makes a big different to stack usage for some patterns.
1388    
1389    26. If a subpattern containing a named recursion or subroutine reference such
1390        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1391        the space required for the compiled pattern went wrong and gave too small a
1392        value. Depending on the environment, this could lead to "Failed: internal
1393        error: code overflow at offset 49" or "glibc detected double free or
1394        corruption" errors.
1395    
1396    27. Applied patches from Google (a) to support the new newline modes and (b) to
1397        advance over multibyte UTF-8 characters in GlobalReplace.
1398    
1399    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1400        difference for some implementation of PCRE in some Windows version.
1401    
1402    29. Added some extra testing facilities to pcretest:
1403    
1404        \q<number>   in a data line sets the "match limit" value
1405        \Q<number>   in a data line sets the "match recursion limt" value
1406        -S <number>  sets the stack size, where <number> is in megabytes
1407    
1408        The -S option isn't available for Windows.
1409    
1410    
1411  Version 6.6 06-Feb-06  Version 6.6 06-Feb-06
1412  ---------------------  ---------------------
1413    

Legend:
Removed from v.89  
changed lines
  Added in v.418

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12