/[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 474 by ph10, Sat Jan 2 16:30:46 2010 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Version 8.01 11-Dec-09
5    ----------------------
6    
7    1.  If a pattern contained a conditional subpattern with only one branch (in
8        particular, this includes all (DEFINE) patterns), a call to pcre_study()
9        computed the wrong minimum data length (which is of course zero for such
10        subpatterns).
11    
12    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
13        the pattern is reset in the first branch, pcre_compile() failed with
14        "internal error: code overflow at offset...". This happened only when
15        the reset was to the original external option setting. (An optimization
16        abstracts leading options settings into an external setting, which was the
17        cause of this.)
18    
19    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
20        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
21        assertion pattern did not match (meaning that the assertion was true), it
22        was incorrectly treated as false if the SKIP had been reached during the
23        matching. This also applied to assertions used as conditions.
24    
25    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
26        assertion subpattern, including such a pattern used as a condition,
27        unpredictable results occurred, instead of the error return
28        PCRE_ERROR_DFA_UITEM.
29    
30    5.  The C++ GlobalReplace function was not working like Perl for the special
31        situation when an empty string is matched. It now does the fancy magic
32        stuff that is necessary.
33    
34    
35    Version 8.00 19-Oct-09
36    ----------------------
37    
38    1.  The table for translating pcre_compile() error codes into POSIX error codes
39        was out-of-date, and there was no check on the pcre_compile() error code
40        being within the table. This could lead to an OK return being given in
41        error.
42    
43    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
44        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
45        in a Windows environment.
46    
47    3.  The pcregrep --count option prints the count for each file even when it is
48        zero, as does GNU grep. However, pcregrep was also printing all files when
49        --files-with-matches was added. Now, when both options are given, it prints
50        counts only for those files that have at least one match. (GNU grep just
51        prints the file name in this circumstance, but including the count seems
52        more useful - otherwise, why use --count?) Also ensured that the
53        combination -clh just lists non-zero counts, with no names.
54    
55    4.  The long form of the pcregrep -F option was incorrectly implemented as
56        --fixed_strings instead of --fixed-strings. This is an incompatible change,
57        but it seems right to fix it, and I didn't think it was worth preserving
58        the old behaviour.
59    
60    5.  The command line items --regex=pattern and --regexp=pattern were not
61        recognized by pcregrep, which required --regex pattern or --regexp pattern
62        (with a space rather than an '='). The man page documented the '=' forms,
63        which are compatible with GNU grep; these now work.
64    
65    6.  No libpcreposix.pc file was created for pkg-config; there was just
66        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
67    
68    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
69        when UCP support is not needed, by modifying the Python script that
70        generates it from Unicode data files. This should not matter if the module
71        is correctly used as a library, but I received one complaint about 50K of
72        unwanted data. My guess is that the person linked everything into his
73        program rather than using a library. Anyway, it does no harm.
74    
75    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
76        was a minimum greater than 1 for a wide character in a possessive
77        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
78        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
79        character. Chaos in the form of incorrect output or a compiling loop could
80        result.
81    
82    9.  The restrictions on what a pattern can contain when partial matching is
83        requested for pcre_exec() have been removed. All patterns can now be
84        partially matched by this function. In addition, if there are at least two
85        slots in the offset vector, the offset of the earliest inspected character
86        for the match and the offset of the end of the subject are set in them when
87        PCRE_ERROR_PARTIAL is returned.
88    
89    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
90        synonymous with PCRE_PARTIAL, for backwards compatibility, and
91        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
92        and may be more useful for multi-segment matching.
93    
94    11. Partial matching with pcre_exec() is now more intuitive. A partial match
95        used to be given if ever the end of the subject was reached; now it is
96        given only if matching could not proceed because another character was
97        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
98        string "Z", which now yields "no match" instead of "partial match". In the
99        case of pcre_dfa_exec(), "no match" is given if every matching path for the
100        final character ended with (*FAIL).
101    
102    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
103        if the pattern had a "must contain" character that was already found in the
104        earlier partial match, unless partial matching was again requested. For
105        example, with the pattern /dog.(body)?/, the "must contain" character is
106        "g". If the first part-match was for the string "dog", restarting with
107        "sbody" failed. This bug has been fixed.
108    
109    13. The string returned by pcre_dfa_exec() after a partial match has been
110        changed so that it starts at the first inspected character rather than the
111        first character of the match. This makes a difference only if the pattern
112        starts with a lookbehind assertion or \b or \B (\K is not supported by
113        pcre_dfa_exec()). It's an incompatible change, but it makes the two
114        matching functions compatible, and I think it's the right thing to do.
115    
116    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
117        so that the demonstration program is easily available in environments where
118        PCRE has not been installed from source.
119    
120    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
121        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
122        library.
123    
124    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
125        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
126        is not the first non-POSIX option to be added. Clearly some people find
127        these options useful.
128    
129    17. If a caller to the POSIX matching function regexec() passes a non-zero
130        value for nmatch with a NULL value for pmatch, the value of
131        nmatch is forced to zero.
132    
133    18. RunGrepTest did not have a test for the availability of the -u option of
134        the diff command, as RunTest does. It now checks in the same way as
135        RunTest, and also checks for the -b option.
136    
137    19. If an odd number of negated classes containing just a single character
138        interposed, within parentheses, between a forward reference to a named
139        subpattern and the definition of the subpattern, compilation crashed with
140        an internal error, complaining that it could not find the referenced
141        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
142        [The bug was that it was starting one character too far in when skipping
143        over the character class, thus treating the ] as data rather than
144        terminating the class. This meant it could skip too much.]
145    
146    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
147        /g option in pcretest when the pattern contains \K, which makes it possible
148        to have an empty string match not at the start, even when the pattern is
149        anchored. Updated pcretest and pcredemo to use this option.
150    
151    21. If the maximum number of capturing subpatterns in a recursion was greater
152        than the maximum at the outer level, the higher number was returned, but
153        with unset values at the outer level. The correct (outer level) value is
154        now given.
155    
156    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
157        PCRE did not set those parentheses (unlike Perl). I have now found a way to
158        make it do so. The string so far is captured, making this feature
159        compatible with Perl.
160    
161    23. The tests have been re-organized, adding tests 11 and 12, to make it
162        possible to check the Perl 5.10 features against Perl 5.10.
163    
164    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
165        pattern matches a fixed length string. PCRE did not allow this; now it
166        does. Neither allows recursion.
167    
168    25. I finally figured out how to implement a request to provide the minimum
169        length of subject string that was needed in order to match a given pattern.
170        (It was back references and recursion that I had previously got hung up
171        on.) This code has now been added to pcre_study(); it finds a lower bound
172        to the length of subject needed. It is not necessarily the greatest lower
173        bound, but using it to avoid searching strings that are too short does give
174        some useful speed-ups. The value is available to calling programs via
175        pcre_fullinfo().
176    
177    26. While implementing 25, I discovered to my embarrassment that pcretest had
178        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
179        study optimizations had never been tested with that matching function.
180        Oops. What is worse, even when it was passed study data, there was a bug in
181        pcre_dfa_exec() that meant it never actually used it. Double oops. There
182        were also very few tests of studied patterns with pcre_dfa_exec().
183    
184    27. If (?| is used to create subpatterns with duplicate numbers, they are now
185        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
186        on the other side of the coin, they are no longer allowed to have different
187        names, because these cannot be distinguished in PCRE, and this has caused
188        confusion. (This is a difference from Perl.)
189    
190    28. When duplicate subpattern names are present (necessarily with different
191        numbers, as required by 27 above), and a test is made by name in a
192        conditional pattern, either for a subpattern having been matched, or for
193        recursion in such a pattern, all the associated numbered subpatterns are
194        tested, and the overall condition is true if the condition is true for any
195        one of them. This is the way Perl works, and is also more like the way
196        testing by number works.
197    
198    
199    Version 7.9 11-Apr-09
200    ---------------------
201    
202    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
203        (pcretest), all targets were linked against these libraries. This included
204        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
205        libraries. This caused unwanted dependencies to be created. This problem
206        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
207        pcretest is linked with readline.
208    
209    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
210        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
211        moved outside it again, because FALSE and TRUE are already defined in AIX,
212        but BOOL is not.
213    
214    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
215        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
216    
217    4.  The pcregrep documentation said spaces were inserted as well as colons (or
218        hyphens) following file names and line numbers when outputting matching
219        lines. This is not true; no spaces are inserted. I have also clarified the
220        wording for the --colour (or --color) option.
221    
222    5.  In pcregrep, when --colour was used with -o, the list of matching strings
223        was not coloured; this is different to GNU grep, so I have changed it to be
224        the same.
225    
226    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
227        each matching line was coloured. Now it goes on to look for further matches
228        of any of the test patterns, which is the same behaviour as GNU grep.
229    
230    7.  A pattern that could match an empty string could cause pcregrep to loop; it
231        doesn't make sense to accept an empty string match in pcregrep, so I have
232        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
233        seems to be how GNU grep behaves.
234    
235    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
236        start or after a newline", because the conditional assertion was not being
237        correctly handled. The rule now is that both the assertion and what follows
238        in the first alternative must satisfy the test.
239    
240    9.  If auto-callout was enabled in a pattern with a conditional group whose
241        condition was an assertion, PCRE could crash during matching, both with
242        pcre_exec() and pcre_dfa_exec().
243    
244    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
245        used for matching.
246    
247    11. Unicode property support in character classes was not working for
248        characters (bytes) greater than 127 when not in UTF-8 mode.
249    
250    12. Added the -M command line option to pcretest.
251    
252    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
253    
254    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
255    
256    16. Added comments and documentation about mis-use of no_arg in the C++
257        wrapper.
258    
259    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
260        from Martin Jerabek that uses macro names for all relevant character and
261        string constants.
262    
263    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
264        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
265        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
266        these, but not everybody uses configure.
267    
268    19. A conditional group that had only one branch was not being correctly
269        recognized as an item that could match an empty string. This meant that an
270        enclosing group might also not be so recognized, causing infinite looping
271        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
272        with the subject "ab", where knowledge that the repeated group can match
273        nothing is needed in order to break the loop.
274    
275    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
276        exec(), but without supplying a callout function, matching went wrong.
277    
278    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
279        leak if the size of the offset vector was greater than 30. When the vector
280        is smaller, the saved offsets during recursion go onto a local stack
281        vector, but for larger vectors malloc() is used. It was failing to free
282        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
283        error, in fact).
284    
285    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
286        heapframe that is used only when UTF-8 support is enabled. This caused no
287        problem, but was untidy.
288    
289    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
290        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
291        included within another project.
292    
293    24. Steven Van Ingelgem's patches to add more options to the CMake support,
294        slightly modified by me:
295    
296          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
297              not building pcregrep.
298    
299          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
300              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
301    
302    25. Forward references, both numeric and by name, in patterns that made use of
303        duplicate group numbers, could behave incorrectly or give incorrect errors,
304        because when scanning forward to find the reference group, PCRE was not
305        taking into account the duplicate group numbers. A pattern such as
306        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
307    
308    26. Changed a few more instances of "const unsigned char *" to USPTR, making
309        the feature of a custom pointer more persuasive (as requested by a user).
310    
311    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
312        pcretest.c, inside #ifndefs, because it seems they are sometimes already
313        pre-defined.
314    
315    28. Added support for (*UTF8) at the start of a pattern.
316    
317    29. Arrange for flags added by the "release type" setting in CMake to be shown
318        in the configuration summary.
319    
320    
321    Version 7.8 05-Sep-08
322    ---------------------
323    
324    1.  Replaced UCP searching code with optimized version as implemented for Ad
325        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
326        stage table and inline lookup instead of a function, giving speed ups of 2
327        to 5 times on some simple patterns that I tested. Permission was given to
328        distribute the MultiStage2.py script that generates the tables (it's not in
329        the tarball, but is in the Subversion repository).
330    
331    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
332        scripts.
333    
334    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
335        a group with a zero qualifier. The result of the study could be incorrect,
336        or the function might crash, depending on the pattern.
337    
338    4.  Caseless matching was not working for non-ASCII characters in back
339        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
340        It now works when Unicode Property Support is available.
341    
342    5.  In pcretest, an escape such as \x{de} in the data was always generating
343        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
344        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
345        truncation.
346    
347    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
348    
349    7.  Added two (int) casts to pcregrep when printing the difference of two
350        pointers, in case they are 64-bit values.
351    
352    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
353        test 2 if it fails.
354    
355    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
356        and a #define of that name to empty if it is not externally set. This is to
357        allow users of MSVC to set it if necessary.
358    
359    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
360        the convenience functions in the pcre_get.c source file.
361    
362    11. An option change at the start of a pattern that had top-level alternatives
363        could cause overwriting and/or a crash. This command provoked a crash in
364        some environments:
365    
366          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
367    
368        This potential security problem was recorded as CVE-2008-2371.
369    
370    12. For a pattern where the match had to start at the beginning or immediately
371        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
372        pcre_dfa_exec() could read past the end of the passed subject if there was
373        no match. To help with detecting such bugs (e.g. with valgrind), I modified
374        pcretest so that it places the subject at the end of its malloc-ed buffer.
375    
376    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
377        exec() might read past the end of the data buffer in UTF-8 mode.
378    
379    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
380        the data contained the byte 0x85 as part of a UTF-8 character within its
381        first line. This applied both to normal and DFA matching.
382    
383    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
384        /^[^d]*?$/8 failed to match "abc".
385    
386    16. Added a missing copyright notice to pcrecpp_internal.h.
387    
388    17. Make it more clear in the documentation that values returned from
389        pcre_exec() in ovector are byte offsets, not character counts.
390    
391    18. Tidied a few places to stop certain compilers from issuing warnings.
392    
393    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
394        supplied by Stefan Weber. I made a further small update for 7.8 because
395        there is a change of source arrangements: the pcre_searchfuncs.c module is
396        replaced by pcre_ucd.c.
397    
398    
399    Version 7.7 07-May-08
400    ---------------------
401    
402    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
403        a string to a long long, pretend we don't even have a long long." This is
404        done by checking for the strtoq, strtoll, and _strtoi64 functions.
405    
406    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
407        pre-7.6 versions, which defined a global no_arg variable instead of putting
408        it in the RE class. (See also #8 below.)
409    
410    3.  Remove a line of dead code, identified by coverity and reported by Nuno
411        Lopes.
412    
413    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
414    
415        (1) The include/exclude patterns were being applied to the whole pathnames
416            of files, instead of just to the final components.
417    
418        (2) If there was more than one level of directory, the subdirectories were
419            skipped unless they satisfied the include/exclude conditions. This is
420            inconsistent with GNU grep (and could even be seen as contrary to the
421            pcregrep specification - which I improved to make it absolutely clear).
422            The action now is always to scan all levels of directory, and just
423            apply the include/exclude patterns to regular files.
424    
425    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
426        --exclude_dir in the tests to avoid scanning .svn directories.
427    
428    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
429        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
430        doesn't support NULs in patterns.
431    
432    7.  Added some missing "const"s to declarations of static tables in
433        pcre_compile.c and pcre_dfa_exec.c.
434    
435    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
436        caused by fix #2  above. (Subsequently also a second patch to fix the
437        first patch. And a third patch - this was a messy problem.)
438    
439    9.  Applied Craig's patch to remove the use of push_back().
440    
441    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
442        matching function regexec().
443    
444    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
445        which, however, unlike Perl's \g{...}, are subroutine calls, not back
446        references. PCRE supports relative numbers with this syntax (I don't think
447        Oniguruma does).
448    
449    12. Previously, a group with a zero repeat such as (...){0} was completely
450        omitted from the compiled regex. However, this means that if the group
451        was called as a subroutine from elsewhere in the pattern, things went wrong
452        (an internal error was given). Such groups are now left in the compiled
453        pattern, with a new opcode that causes them to be skipped at execution
454        time.
455    
456    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
457        to the way PCRE behaves:
458    
459        (a) A lone ] character is dis-allowed (Perl treats it as data).
460    
461        (b) A back reference to an unmatched subpattern matches an empty string
462            (Perl fails the current match path).
463    
464        (c) A data ] in a character class must be notated as \] because if the
465            first data character in a class is ], it defines an empty class. (In
466            Perl it is not possible to have an empty class.) The empty class []
467            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
468            The negative empty class [^] matches any one character, independently
469            of the DOTALL setting.
470    
471    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
472        non-existent subpattern following a character class starting with ']' and
473        containing () gave an internal compiling error instead of "reference to
474        non-existent subpattern". Fortunately, when the pattern did exist, the
475        compiled code was correct. (When scanning forwards to check for the
476        existencd of the subpattern, it was treating the data ']' as terminating
477        the class, so got the count wrong. When actually compiling, the reference
478        was subsequently set up correctly.)
479    
480    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
481        it was being rejected as not supported by pcre_dfa_exec(), even though
482        other assertions are supported. I have made pcre_dfa_exec() support
483        (*FAIL).
484    
485    16. The implementation of 13c above involved the invention of a new opcode,
486        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
487        cannot be changed at match time, I realized I could make a small
488        improvement to matching performance by compiling OP_ALLANY instead of
489        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
490        on the OP_ANY path.
491    
492    17. Compiling pcretest on Windows with readline support failed without the
493        following two fixes: (1) Make the unistd.h include conditional on
494        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
495    
496    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
497        ncurses library to be included for pcretest when ReadLine support is
498        requested, but also to allow for it to be overridden. This patch came from
499        Daniel Bergström.
500    
501    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
502        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
503        any errors with the current Unicode tables. Thanks to Peter Kankowski for
504        spotting this.
505    
506    
507    Version 7.6 28-Jan-08
508    ---------------------
509    
510    1.  A character class containing a very large number of characters with
511        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
512        overflow.
513    
514    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
515        HAVE_LONG_LONG is not defined.
516    
517    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
518        bring them up to date and include new features. This patch includes:
519    
520        - Fixed PH's badly added libz and libbz2 support.
521        - Fixed a problem with static linking.
522        - Added pcredemo. [But later removed - see 7 below.]
523        - Fixed dftables problem and added an option.
524        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
525            HAVE_LONG_LONG.
526        - Added readline support for pcretest.
527        - Added an listing of the option settings after cmake has run.
528    
529    4.  A user submitted a patch to Makefile that makes it easy to create
530        "pcre.dll" under mingw when using Configure/Make. I added stuff to
531        Makefile.am that cause it to include this special target, without
532        affecting anything else. Note that the same mingw target plus all
533        the other distribution libraries and programs are now supported
534        when configuring with CMake (see 6 below) instead of with
535        Configure/Make.
536    
537    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
538        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
539        exported in the Windows port". It has not yet been confirmed that the patch
540        solves the problem, but it does no harm.
541    
542    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
543        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
544        with CMake, and also correct the comment about stack recursion.
545    
546    7.  Remove the automatic building of pcredemo from the ./configure system and
547        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
548        of a program that users should build themselves after PCRE is installed, so
549        building it automatically is not really right. What is more, it gave
550        trouble in some build environments.
551    
552    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
553    
554    
555    Version 7.5 10-Jan-08
556    ---------------------
557    
558    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
559        values in parens when parsing an RE using the C++ wrapper."
560    
561    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
562        Characters greater than 255 were excluded from the class instead of being
563        included.
564    
565    3.  The same bug as (2) above applied to negated POSIX classes such as
566        [:^space:].
567    
568    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
569        defined or documented. It seems to have been a typo for PCRE_STATIC, so
570        I have changed it.
571    
572    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
573        first named subpattern) and a construct such as (?&a) would reference the
574        first named subpattern whose name started with "a" (in other words, the
575        length check was missing). Both these problems are fixed. "Subpattern name
576        expected" is now given for (?&) (a zero-length name), and this patch also
577        makes it give the same error for \k'' (previously it complained that that
578        was a reference to a non-existent subpattern).
579    
580    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
581        this is right because (?- can be followed by option settings as well as by
582        digits. I have, however, made the messages clearer.
583    
584    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
585        than the number used in the conditional) now cause a compile-time error.
586        This is actually not compatible with Perl, which accepts such patterns, but
587        treats the conditional as always being FALSE (as PCRE used to), but it
588        seems to me that giving a diagnostic is better.
589    
590    8.  Change "alphameric" to the more common word "alphanumeric" in comments
591        and messages.
592    
593    9.  Fix two occurrences of "backslash" in comments that should have been
594        "backspace".
595    
596    10. Remove two redundant lines of code that can never be obeyed (their function
597        was moved elsewhere).
598    
599    11. The program that makes PCRE's Unicode character property table had a bug
600        which caused it to generate incorrect table entries for sequences of
601        characters that have the same character type, but are in different scripts.
602        It amalgamated them into a single range, with the script of the first of
603        them. In other words, some characters were in the wrong script. There were
604        thirteen such cases, affecting characters in the following ranges:
605    
606          U+002b0 - U+002c1
607          U+0060c - U+0060d
608          U+0061e - U+00612
609          U+0064b - U+0065e
610          U+0074d - U+0076d
611          U+01800 - U+01805
612          U+01d00 - U+01d77
613          U+01d9b - U+01dbf
614          U+0200b - U+0200f
615          U+030fc - U+030fe
616          U+03260 - U+0327f
617          U+0fb46 - U+0fbb1
618          U+10450 - U+1049d
619    
620    12. The -o option (show only the matching part of a line) for pcregrep was not
621        compatible with GNU grep in that, if there was more than one match in a
622        line, it showed only the first of them. It now behaves in the same way as
623        GNU grep.
624    
625    13. If the -o and -v options were combined for pcregrep, it printed a blank
626        line for every non-matching line. GNU grep prints nothing, and pcregrep now
627        does the same. The return code can be used to tell if there were any
628        non-matching lines.
629    
630    14. Added --file-offsets and --line-offsets to pcregrep.
631    
632    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
633        infinitely looping recursion. The bug was that positive lookaheads were not
634        being skipped when checking for a possible empty match (negative lookaheads
635        and both kinds of lookbehind were skipped).
636    
637    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
638        inclusion of <windows.h> to before rather than after the definition of
639        INVALID_FILE_ATTRIBUTES (patch from David Byron).
640    
641    17. Specifying a possessive quantifier with a specific limit for a Unicode
642        character property caused pcre_compile() to compile bad code, which led at
643        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
644        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
645        caused the error; without that there was no problem.
646    
647    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
648    
649    19. Added --enable-pcretest-libreadline.
650    
651    20. In pcrecpp.cc, the variable 'count' was incremented twice in
652        RE::GlobalReplace(). As a result, the number of replacements returned was
653        double what it should be. I removed one of the increments, but Craig sent a
654        later patch that removed the other one (the right fix) and added unit tests
655        that check the return values (which was not done before).
656    
657    21. Several CMake things:
658    
659        (1) Arranged that, when cmake is used on Unix, the libraries end up with
660            the names libpcre and libpcreposix, not just pcre and pcreposix.
661    
662        (2) The above change means that pcretest and pcregrep are now correctly
663            linked with the newly-built libraries, not previously installed ones.
664    
665        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
666    
667    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
668        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
669        UTF-8 newline character). The key issue is that the pattern starts .*;
670        this means that the match must be either at the beginning, or after a
671        newline. The bug was in the code for advancing after a failed match and
672        checking that the new position followed a newline. It was not taking
673        account of UTF-8 characters correctly.
674    
675    23. PCRE was behaving differently from Perl in the way it recognized POSIX
676        character classes. PCRE was not treating the sequence [:...:] as a
677        character class unless the ... were all letters. Perl, however, seems to
678        allow any characters between [: and :], though of course it rejects as
679        unknown any "names" that contain non-letters, because all the known class
680        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
681        for example, whereas PCRE did not - it did not recognize a POSIX character
682        class. This seemed a bit dangerous, so the code has been changed to be
683        closer to Perl. The behaviour is not identical to Perl, because PCRE will
684        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
685        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
686        Perl does, and where it didn't before.
687    
688    24. Rewrite so as to remove the single use of %n from pcregrep because in some
689        Windows environments %n is disabled by default.
690    
691    
692    Version 7.4 21-Sep-07
693    ---------------------
694    
695    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
696        means that a class such as [\s] counted as "explicit reference to CR or
697        LF". That isn't really right - the whole point of the change was to try to
698        help when there was an actual mention of one of the two characters. So now
699        the change happens only if \r or \n (or a literal CR or LF) character is
700        encountered.
701    
702    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
703        of both had grown to the point where there were only 3 bits left.
704        Fortunately, there was spare space in the data structure, and so I have
705        moved the internal flags into a new 16-bit field to free up more option
706        bits.
707    
708    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
709        but did not set the internal JCHANGED flag - either of these is enough to
710        control the way the "get" function works - but the PCRE_INFO_JCHANGED
711        facility is supposed to tell if (?J) was ever used, so now (?J) at the
712        start sets both bits.
713    
714    4.  Added options (at build time, compile time, exec time) to change \R from
715        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
716    
717    5.  doc/pcresyntax.html was missing from the distribution.
718    
719    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
720        compatibility, even though it is no longer used.
721    
722    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
723        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
724        windows.h file is present (where different names are used). [This was
725        reversed later after testing - see 16 below.]
726    
727    8.  Changed all #include <config.h> to #include "config.h". There were also
728        some further <pcre.h> cases that I changed to "pcre.h".
729    
730    9.  When pcregrep was used with the --colour option, it missed the line ending
731        sequence off the lines that it output.
732    
733    10. It was pointed out to me that arrays of string pointers cause lots of
734        relocations when a shared library is dynamically loaded. A technique of
735        using a single long string with a table of offsets can drastically reduce
736        these. I have refactored PCRE in four places to do this. The result is
737        dramatic:
738    
739          Originally:                          290
740          After changing UCP table:            187
741          After changing error message table:   43
742          After changing table of "verbs"       36
743          After changing table of Posix names   22
744    
745        Thanks to the folks working on Gregex for glib for this insight.
746    
747    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
748        unicode-properties was also set.
749    
750    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
751    
752    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
753        checked only for CRLF.
754    
755    14. Added casts to pcretest.c to avoid compiler warnings.
756    
757    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
758    
759    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
760        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
761        entirely. This removes changes made in 7 above.
762    
763    17. The CMake files have been updated, and there is now more information about
764        building with CMake in the NON-UNIX-USE document.
765    
766    
767    Version 7.3 28-Aug-07
768    ---------------------
769    
770     1. In the rejigging of the build system that eventually resulted in 7.1, the
771        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
772        brackets there is not right, since it causes compilers to look for an
773        installed pcre.h, not the version that is in the source that is being
774        compiled (which of course may be different). I have changed it back to:
775    
776          #include "pcre.h"
777    
778        I have a vague recollection that the change was concerned with compiling in
779        different directories, but in the new build system, that is taken care of
780        by the VPATH setting the Makefile.
781    
782     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
783        when the subject happened to end in the byte 0x85 (e.g. if the last
784        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
785        characters but of course it shouldn't be taken as a newline when it is part
786        of another character. The bug was that, for an unlimited repeat of . in
787        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
788        characters when looking for a newline.
789    
790     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
791    
792     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
793        in debug output.
794    
795     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
796        long printing in the pcrecpp unittest when running under MinGW.
797    
798     6. ESC_K was left out of the EBCDIC table.
799    
800     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
801        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
802        limit also applies to "virtual nesting" when a pattern is recursive, and in
803        this case 1000 isn't so big. I have been able to remove this limit at the
804        expense of backing off one optimization in certain circumstances. Normally,
805        when pcre_exec() would call its internal match() function recursively and
806        immediately return the result unconditionally, it uses a "tail recursion"
807        feature to save stack. However, when a subpattern that can match an empty
808        string has an unlimited repetition quantifier, it no longer makes this
809        optimization. That gives it a stack frame in which to save the data for
810        checking that an empty string has been matched. Previously this was taken
811        from the 1000-entry workspace that had been reserved. So now there is no
812        explicit limit, but more stack is used.
813    
814     8. Applied Daniel's patches to solve problems with the import/export magic
815        syntax that is required for Windows, and which was going wrong for the
816        pcreposix and pcrecpp parts of the library. These were overlooked when this
817        problem was solved for the main library.
818    
819     9. There were some crude static tests to avoid integer overflow when computing
820        the size of patterns that contain repeated groups with explicit upper
821        limits. As the maximum quantifier is 65535, the maximum group length was
822        set at 30,000 so that the product of these two numbers did not overflow a
823        32-bit integer. However, it turns out that people want to use groups that
824        are longer than 30,000 bytes (though not repeat them that many times).
825        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
826        made it possible to implement the integer overflow checks in a much more
827        dynamic way, which I have now done. The artificial limitation on group
828        length has been removed - we now have only the limit on the total length of
829        the compiled pattern, which depends on the LINK_SIZE setting.
830    
831    10. Fixed a bug in the documentation for get/copy named substring when
832        duplicate names are permitted. If none of the named substrings are set, the
833        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
834        empty string.
835    
836    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
837        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
838        because the ] is interpreted as the first data character and the
839        terminating ] is not found. PCRE has been made compatible with Perl in this
840        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
841        cause memory overwriting.
842    
843    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
844        string has been matched (to stop an infinite loop). It was not recognizing
845        a conditional subpattern that could match an empty string if that
846        subpattern was within another subpattern. For example, it looped when
847        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
848        condition was not nested. This bug has been fixed.
849    
850    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
851        past the start of the subject in the presence of bytes with the top bit
852        set, for example "\x8aBCD".
853    
854    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
855        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
856    
857    14. Optimized (?!) to (*FAIL).
858    
859    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
860        This restricts code points to be within the range 0 to 0x10FFFF, excluding
861        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
862        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
863        does: it's just the validity check that is more restrictive.
864    
865    16. Inserted checks for integer overflows during escape sequence (backslash)
866        processing, and also fixed erroneous offset values for syntax errors during
867        backslash processing.
868    
869    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
870        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
871    
872    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
873        caused an overrun.
874    
875    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
876        something other than just ASCII characters) inside a group that had an
877        unlimited repeat caused a loop at compile time (while checking to see
878        whether the group could match an empty string).
879    
880    20. Debugging a pattern containing \p or \P could cause a crash. For example,
881        [\P{Any}] did so. (Error in the code for printing property names.)
882    
883    21. An orphan \E inside a character class could cause a crash.
884    
885    22. A repeated capturing bracket such as (A)? could cause a wild memory
886        reference during compilation.
887    
888    23. There are several functions in pcre_compile() that scan along a compiled
889        expression for various reasons (e.g. to see if it's fixed length for look
890        behind). There were bugs in these functions when a repeated \p or \P was
891        present in the pattern. These operators have additional parameters compared
892        with \d, etc, and these were not being taken into account when moving along
893        the compiled data. Specifically:
894    
895        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
896            length.
897    
898        (b) An item such as \pL+ within a repeated group could cause crashes or
899            loops.
900    
901        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
902            "reference to non-existent subpattern" error.
903    
904        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
905    
906    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
907        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
908    
909    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
910    
911    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
912        character were causing crashes (broken optimization).
913    
914    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
915        \p or \P) caused a compile-time loop.
916    
917    28. More problems have arisen in unanchored patterns when CRLF is a valid line
918        break. For example, the unstudied pattern [\r\n]A does not match the string
919        "\r\nA" because change 7.0/46 below moves the current point on by two
920        characters after failing to match at the start. However, the pattern \nA
921        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
922        the same is true. There doesn't seem any very clean way out of this, but
923        what I have chosen to do makes the common cases work: PCRE now takes note
924        of whether there can be an explicit match for \r or \n anywhere in the
925        pattern, and if so, 7.0/46 no longer applies. As part of this change,
926        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
927        pattern has explicit CR or LF references.
928    
929    29. Added (*CR) etc for changing newline setting at start of pattern.
930    
931    
932    Version 7.2 19-Jun-07
933    ---------------------
934    
935     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
936        which is apparently normally available under Windows.
937    
938     2. Re-jig the pcregrep tests with different newline settings in an attempt
939        to make them independent of the local environment's newline setting.
940    
941     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
942    
943     4. Some of the "internals" tests were previously cut out when the link size
944        was not 2, because the output contained actual offsets. The recent new
945        "Z" feature of pcretest means that these can be cut out, making the tests
946        usable with all link sizes.
947    
948     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
949        stack recursion. This gives a massive performance boost under BSD, but just
950        a small improvement under Linux. However, it saves one field in the frame
951        in all cases.
952    
953     6. Added more features from the forthcoming Perl 5.10:
954    
955        (a) (?-n) (where n is a string of digits) is a relative subroutine or
956            recursion call. It refers to the nth most recently opened parentheses.
957    
958        (b) (?+n) is also a relative subroutine call; it refers to the nth next
959            to be opened parentheses.
960    
961        (c) Conditions that refer to capturing parentheses can be specified
962            relatively, for example, (?(-2)... or (?(+3)...
963    
964        (d) \K resets the start of the current match so that everything before
965            is not part of it.
966    
967        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
968    
969        (f) \g{name} is another synonym - part of Perl 5.10's unification of
970            reference syntax.
971    
972        (g) (?| introduces a group in which the numbering of parentheses in each
973            alternative starts with the same number.
974    
975        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
976    
977     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
978        PCRE_INFO_JCHANGED.
979    
980     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
981        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
982        for detecting groups that can match an empty string.
983    
984     9. A pattern with a very large number of alternatives (more than several
985        hundred) was running out of internal workspace during the pre-compile
986        phase, where pcre_compile() figures out how much memory will be needed. A
987        bit of new cunning has reduced the workspace needed for groups with
988        alternatives. The 1000-alternative test pattern now uses 12 bytes of
989        workspace instead of running out of the 4096 that are available.
990    
991    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
992    
993    11. Applied patch from Google to remove an optimization that didn't quite work.
994        The report of the bug said:
995    
996          pcrecpp::RE("a*").FullMatch("aaa") matches, while
997          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
998          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
999    
1000    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1001        it matched the wrong number of bytes.
1002    
1003    
1004    Version 7.1 24-Apr-07
1005    ---------------------
1006    
1007     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1008        that is more "standard", making use of automake and other Autotools. There
1009        is some re-arrangement of the files and adjustment of comments consequent
1010        on this.
1011    
1012     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1013        for recursive directory scanning broke on some systems because the files
1014        are not scanned in any specific order and on different systems the order
1015        was different. A call to "sort" has been inserted into RunGrepTest for the
1016        approprate test as a short-term fix. In the longer term there may be an
1017        alternative.
1018    
1019     3. I had an email from Eric Raymond about problems translating some of PCRE's
1020        man pages to HTML (despite the fact that I distribute HTML pages, some
1021        people do their own conversions for various reasons). The problems
1022        concerned the use of low-level troff macros .br and .in. I have therefore
1023        removed all such uses from the man pages (some were redundant, some could
1024        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1025        HTML has been updated to handle .nf/.fi and to complain if it encounters
1026        .br or .in.
1027    
1028     4. Updated comments in configure.ac that get placed in config.h.in and also
1029        arranged for config.h to be included in the distribution, with the name
1030        config.h.generic, for the benefit of those who have to compile without
1031        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1032    
1033     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1034        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1035        makevp.bat for the current PCRE, using the additional files
1036        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1037    
1038     6. A Windows user reported a minor discrepancy with test 2, which turned out
1039        to be caused by a trailing space on an input line that had got lost in his
1040        copy. The trailing space was an accident, so I've just removed it.
1041    
1042     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1043        that is needed.
1044    
1045     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1046        as "const" (a) because they are and (b) because it helps the PHP
1047        maintainers who have recently made a script to detect big data structures
1048        in the php code that should be moved to the .rodata section. I remembered
1049        to update Builducptable as well, so it won't revert if ucptable.h is ever
1050        re-created.
1051    
1052     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1053        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1054        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1055        support is not required. This saves 1.5-2K of code, which is important in
1056        some applications.
1057    
1058        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1059        so as not to refer to the tables, even though these functions will never be
1060        called when UTF-8 support is disabled. Otherwise there are problems with a
1061        shared library.
1062    
1063    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1064    
1065        (a) It was defining its arguments as char * instead of void *.
1066    
1067        (b) It was assuming that all moves were upwards in memory; this was true
1068            a long time ago when I wrote it, but is no longer the case.
1069    
1070        The emulated memove() is provided for those environments that have neither
1071        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1072        is clearly not the case, as these two bugs were recently reported.
1073    
1074    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1075        and Detrail to create the HTML documentation, the .txt form of the man
1076        pages, and it removes trailing spaces from listed files. It also creates
1077        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1078        case, it wraps all the #defines with #ifndefs. This script should be run
1079        before "make dist".
1080    
1081    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1082        with Unicode property support.
1083    
1084        (a) For a maximizing quantifier, if the two different cases of the
1085            character were of different lengths in their UTF-8 codings (there are
1086            some cases like this - I found 11), and the matching function had to
1087            back up over a mixture of the two cases, it incorrectly assumed they
1088            were both the same length.
1089    
1090        (b) When PCRE was configured to use the heap rather than the stack for
1091            recursion during matching, it was not correctly preserving the data for
1092            the other case of a UTF-8 character when checking ahead for a match
1093            while processing a minimizing repeat. If the check also involved
1094            matching a wide character, but failed, corruption could cause an
1095            erroneous result when trying to check for a repeat of the original
1096            character.
1097    
1098    13. Some tidying changes to the testing mechanism:
1099    
1100        (a) The RunTest script now detects the internal link size and whether there
1101            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1102            values substituted by "configure". (The RunGrepTest script already did
1103            this for UTF-8.) The configure.ac script no longer substitutes the
1104            relevant variables.
1105    
1106        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1107            with length and offset values. This means that the output is different
1108            for different internal link sizes. Test 2 is skipped for link sizes
1109            other than 2 because of this, bypassing the problem. Unfortunately,
1110            there was also a test in test 3 (the locale tests) that used /B and
1111            failed for link sizes other than 2. Rather than cut the whole test out,
1112            I have added a new /Z option to pcretest that replaces the length and
1113            offset values with spaces. This is now used to make test 3 independent
1114            of link size. (Test 2 will be tidied up later.)
1115    
1116    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1117        segmentation fault instead of returning the appropriate error message.
1118    
1119    15. In multiline mode when the newline sequence was set to "any", the pattern
1120        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1121        This doesn't seem right; it now treats the CRLF combination as the line
1122        ending, and so does not match in that case. It's only a pattern such as ^$
1123        that would hit this one: something like ^ABC$ would have failed after \r
1124        and then tried again after \r\n.
1125    
1126    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1127        in an attempt to make files that differ only in their line terminators
1128        compare equal. This works on Linux.
1129    
1130    17. Under certain error circumstances pcregrep might try to free random memory
1131        as it exited. This is now fixed, thanks to valgrind.
1132    
1133    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1134        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1135        was because its rules for how to advance for /g after matching an empty
1136        string at the end of a line did not allow for this case. They now check for
1137        it specially.
1138    
1139    20. pcretest is supposed to handle patterns and data of any length, by
1140        extending its buffers when necessary. It was getting this wrong when the
1141        buffer for a data line had to be extended.
1142    
1143    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1144        CRLF as a newline sequence.
1145    
1146    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1147        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1148        I have nevertheless tidied it up.
1149    
1150    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1151    
1152    24. Added a man page for pcre-config.
1153    
1154    
1155    Version 7.0 19-Dec-06
1156    ---------------------
1157    
1158     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1159        moving to gcc 4.1.1.
1160    
1161     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1162        sys/time.h, which is documented as needed for this function. It doesn't
1163        seem to matter on Linux, but it showed up on some releases of OS X.
1164    
1165     3. It seems that there are systems where bytes whose values are greater than
1166        127 match isprint() in the "C" locale. The "C" locale should be the
1167        default when a C program starts up. In most systems, only ASCII printing
1168        characters match isprint(). This difference caused the output from pcretest
1169        to vary, making some of the tests fail. I have changed pcretest so that:
1170    
1171        (a) When it is outputting text in the compiled version of a pattern, bytes
1172            other than 32-126 are always shown as hex escapes.
1173    
1174        (b) When it is outputting text that is a matched part of a subject string,
1175            it does the same, unless a different locale has been set for the match
1176            (using the /L modifier). In this case, it uses isprint() to decide.
1177    
1178     4. Fixed a major bug that caused incorrect computation of the amount of memory
1179        required for a compiled pattern when options that changed within the
1180        pattern affected the logic of the preliminary scan that determines the
1181        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1182        that the computed length was too small. The symptoms of this bug were
1183        either the PCRE error "internal error: code overflow" from pcre_compile(),
1184        or a glibc crash with a message such as "pcretest: free(): invalid next
1185        size (fast)". Examples of patterns that provoked this bug (shown in
1186        pcretest format) are:
1187    
1188          /(?-x: )/x
1189          /(?x)(?-x: \s*#\s*)/
1190          /((?i)[\x{c0}])/8
1191          /(?i:[\x{c0}])/8
1192    
1193        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1194        is now done differently.
1195    
1196     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1197        wrapper classes; (b) implement a new function in the C++ scanner that is
1198        more efficient than the old way of doing things because it avoids levels of
1199        recursion in the regex matching; (c) add a paragraph to the documentation
1200        for the FullMatch() function.
1201    
1202     6. The escape sequence \n was being treated as whatever was defined as
1203        "newline". Not only was this contrary to the documentation, which states
1204        that \n is character 10 (hex 0A), but it also went horribly wrong when
1205        "newline" was defined as CRLF. This has been fixed.
1206    
1207     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1208        was being set to -1 for the "end of line" case (supposedly a value that no
1209        character can have). Though this value is never used (the check for end of
1210        line is "zero bytes in current character"), it caused compiler complaints.
1211        I've changed it to 0xffffffff.
1212    
1213     8. In pcre_version.c, the version string was being built by a sequence of
1214        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1215        string (as it is for production releases) called a macro with an empty
1216        argument. The C standard says the result of this is undefined. The gcc
1217        compiler treats it as an empty string (which was what was wanted) but it is
1218        reported that Visual C gives an error. The source has been hacked around to
1219        avoid this problem.
1220    
1221     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1222        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1223        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1224        of them did).
1225    
1226    10. Originally, pcretest opened its input and output without "b"; then I was
1227        told that "b" was needed in some environments, so it was added for release
1228        5.0 to both the input and output. (It makes no difference on Unix-like
1229        systems.) Later I was told that it is wrong for the input on Windows. I've
1230        now abstracted the modes into two macros, to make it easier to fiddle with
1231        them, and removed "b" from the input mode under Windows.
1232    
1233    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1234    
1235    12. Added -help and --help to pcretest as an official way of being reminded
1236        of the options.
1237    
1238    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1239        and pcrecpp.cc because they annoy compilers at high warning levels.
1240    
1241    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1242    
1243    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1244        scripts are not C programs :-) and which was not noticed because it works
1245        on Linux.
1246    
1247    16. pcretest is supposed to handle any length of pattern and data line (as one
1248        line or as a continued sequence of lines) by extending its input buffer if
1249        necessary. This feature was broken for very long pattern lines, leading to
1250        a string of junk being passed to pcre_compile() if the pattern was longer
1251        than about 50K.
1252    
1253    17. I have done a major re-factoring of the way pcre_compile() computes the
1254        amount of memory needed for a compiled pattern. Previously, there was code
1255        that made a preliminary scan of the pattern in order to do this. That was
1256        OK when PCRE was new, but as the facilities have expanded, it has become
1257        harder and harder to keep it in step with the real compile phase, and there
1258        have been a number of bugs (see for example, 4 above). I have now found a
1259        cunning way of running the real compile function in a "fake" mode that
1260        enables it to compute how much memory it would need, while actually only
1261        ever using a few hundred bytes of working memory and without too many
1262        tests of the mode. This should make future maintenance and development
1263        easier. A side effect of this work is that the limit of 200 on the nesting
1264        depth of parentheses has been removed (though this was never a serious
1265        limitation, I suspect). However, there is a downside: pcre_compile() now
1266        runs more slowly than before (30% or more, depending on the pattern). I
1267        hope this isn't a big issue. There is no effect on runtime performance.
1268    
1269    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1270        newline (only possible for the last line of a file) and it was a
1271        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1272    
1273    19. Added additional timing features to pcretest. (1) The -tm option now times
1274        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1275        separate command line item, by a number that specifies the number of
1276        repeats to use when timing. The default is 50000; this gives better
1277        precision, but takes uncomfortably long for very large patterns.
1278    
1279    20. Extended pcre_study() to be more clever in cases where a branch of a
1280        subpattern has no definite first character. For example, (a*|b*)[cd] would
1281        previously give no result from pcre_study(). Now it recognizes that the
1282        first character must be a, b, c, or d.
1283    
1284    21. There was an incorrect error "recursive call could loop indefinitely" if
1285        a subpattern (or the entire pattern) that was being tested for matching an
1286        empty string contained only one non-empty item after a nested subpattern.
1287        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1288        incorrectly, because the \d was being skipped in the check.
1289    
1290    22. The pcretest program now has a new pattern option /B and a command line
1291        option -b, which is equivalent to adding /B to every pattern. This causes
1292        it to show the compiled bytecode, without the additional information that
1293        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1294        is the same as /B/I).
1295    
1296    23. A new optimization is now able automatically to treat some sequences such
1297        as a*b as a*+b. More specifically, if something simple (such as a character
1298        or a simple class like \d) has an unlimited quantifier, and is followed by
1299        something that cannot possibly match the quantified thing, the quantifier
1300        is automatically "possessified".
1301    
1302    24. A recursive reference to a subpattern whose number was greater than 39
1303        went wrong under certain circumstances in UTF-8 mode. This bug could also
1304        have affected the operation of pcre_study().
1305    
1306    25. Realized that a little bit of performance could be had by replacing
1307        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1308    
1309    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1310    
1311    27. Possessive quantifiers such as a++ were previously implemented by turning
1312        them into atomic groups such as ($>a+). Now they have their own opcodes,
1313        which improves performance. This includes the automatically created ones
1314        from 23 above.
1315    
1316    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1317        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1318        the first matched character to be a colon. This applied both to named and
1319        numbered groups.
1320    
1321    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1322    
1323    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1324        building PCRE on Windows easier, so I have included it in the distribution.
1325    
1326    31. There is now a check in pcretest against a ridiculously large number being
1327        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1328        loop, the loop is abandoned.
1329    
1330    32. Forward references to subpatterns in conditions such as (?(2)...) where
1331        subpattern 2 is defined later cause pcre_compile() to search forwards in
1332        the pattern for the relevant set of parentheses. This search went wrong
1333        when there were unescaped parentheses in a character class, parentheses
1334        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1335    
1336    33. "Subroutine" calls and backreferences were previously restricted to
1337        referencing subpatterns earlier in the regex. This restriction has now
1338        been removed.
1339    
1340    34. Added a number of extra features that are going to be in Perl 5.10. On the
1341        whole, these are just syntactic alternatives for features that PCRE had
1342        previously implemented using the Python syntax or my own invention. The
1343        other formats are all retained for compatibility.
1344    
1345        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1346            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1347            also .NET compatible.
1348    
1349        (b) A recursion or subroutine call to a named group can now be defined as
1350            (?&name) as well as (?P>name).
1351    
1352        (c) A backreference to a named group can now be defined as \k<name> or
1353            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1354            5.10, are also .NET compatible.
1355    
1356        (d) A conditional reference to a named group can now use the syntax
1357            (?(<name>) or (?('name') as well as (?(name).
1358    
1359        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1360            groups (named and numbered) that are never evaluated inline, but can be
1361            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1362            is always false. There may be only one alternative in such a group.
1363    
1364        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1365            as the simple (?(R). The condition is true only if the most recent
1366            recursion is that of the given number or name. It does not search out
1367            through the entire recursion stack.
1368    
1369        (g) The escape \gN or \g{N} has been added, where N is a positive or
1370            negative number, specifying an absolute or relative reference.
1371    
1372    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1373        some "unreachable code" warnings.
1374    
1375    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1376        things, this adds five new scripts.
1377    
1378    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1379        There were also incompatibilities regarding the handling of \Q..\E inside
1380        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1381        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1382    
1383    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1384        matches an empty string, and forcibly breaks the loop. There were bugs in
1385        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1386        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1387        separate and independent bugs (that affected different cases) have been
1388        fixed.
1389    
1390    39. Refactored the code to abolish the use of different opcodes for small
1391        capturing bracket numbers. This is a tidy that I avoided doing when I
1392        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1393        The new approach is not only tidier, it makes it possible to reduce the
1394        memory needed to fix the previous bug (38).
1395    
1396    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1397        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1398        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1399        mode.
1400    
1401    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1402        report.
1403    
1404    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1405        copy construction and assignment in the C++ wrapper.
1406    
1407    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1408        couple of bugs that could have given wrong results in the "--newline=crlf"
1409        case.
1410    
1411    44. Added a number of casts and did some reorganization of signed/unsigned int
1412        variables following suggestions from Dair Grant. Also renamed the variable
1413        "this" as "item" because it is a C++ keyword.
1414    
1415    45. Arranged for dftables to add
1416    
1417          #include "pcre_internal.h"
1418    
1419        to pcre_chartables.c because without it, gcc 4.x may remove the array
1420        definition from the final binary if PCRE is built into a static library and
1421        dead code stripping is activated.
1422    
1423    46. For an unanchored pattern, if a match attempt fails at the start of a
1424        newline sequence, and the newline setting is CRLF or ANY, and the next two
1425        characters are CRLF, advance by two characters instead of one.
1426    
1427    
1428    Version 6.7 04-Jul-06
1429    ---------------------
1430    
1431     1. In order to handle tests when input lines are enormously long, pcretest has
1432        been re-factored so that it automatically extends its buffers when
1433        necessary. The code is crude, but this _is_ just a test program. The
1434        default size has been increased from 32K to 50K.
1435    
1436     2. The code in pcre_study() was using the value of the re argument before
1437        testing it for NULL. (Of course, in any sensible call of the function, it
1438        won't be NULL.)
1439    
1440     3. The memmove() emulation function in pcre_internal.h, which is used on
1441        systems that lack both memmove() and bcopy() - that is, hardly ever -
1442        was missing a "static" storage class specifier.
1443    
1444     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1445        containing an extended class (one that cannot be represented by a bitmap
1446        because it contains high-valued characters or Unicode property items, e.g.
1447        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1448        pattern, but PCRE should not loop if you do not (it no longer does).
1449        [Detail: two cases were found: (a) a repeated subpattern containing an
1450        extended class; (b) a recursive reference to a subpattern that followed a
1451        previous extended class. It wasn't skipping over the extended class
1452        correctly when UTF-8 mode was not set.]
1453    
1454     5. A negated single-character class was not being recognized as fixed-length
1455        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1456        compile error "lookbehind assertion is not fixed length".
1457    
1458     6. The RunPerlTest auxiliary script was showing an unexpected difference
1459        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1460        write a Perl script that can interpret lines of an input file either as
1461        byte characters or as UTF-8, which is what "perltest" was being required to
1462        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1463        can't do is switch easily at run time between having the "use utf8;" pragma
1464        or not. In the end, I fudged it by using the RunPerlTest script to insert
1465        "use utf8;" explicitly for the UTF-8 tests.
1466    
1467     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1468        the end of the subject string, contrary to the documentation and to what
1469        Perl does. This was true of both matching functions. Now it matches only at
1470        the start of the subject and immediately after *internal* newlines.
1471    
1472     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1473        a pointer to an int instead of a pointer to an unsigned long int. This
1474        caused problems on 64-bit systems.
1475    
1476     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1477        instance of the 'standard' template library not being so standard".
1478    
1479    10. There was no check on the number of named subpatterns nor the maximum
1480        length of a subpattern name. The product of these values is used to compute
1481        the size of the memory block for a compiled pattern. By supplying a very
1482        long subpattern name and a large number of named subpatterns, the size
1483        computation could be caused to overflow. This is now prevented by limiting
1484        the length of names to 32 characters, and the number of named subpatterns
1485        to 10,000.
1486    
1487    11. Subpatterns that are repeated with specific counts have to be replicated in
1488        the compiled pattern. The size of memory for this was computed from the
1489        length of the subpattern and the repeat count. The latter is limited to
1490        65535, but there was no limit on the former, meaning that integer overflow
1491        could in principle occur. The compiled length of a repeated subpattern is
1492        now limited to 30,000 bytes in order to prevent this.
1493    
1494    12. Added the optional facility to have named substrings with the same name.
1495    
1496    13. Added the ability to use a named substring as a condition, using the
1497        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1498        are numbers (not recommended). Forward references are permitted.
1499    
1500    14. Added forward references in named backreferences (if you see what I mean).
1501    
1502    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1503        pattern could run off the end of the subject. For example, the pattern
1504        "(?s)(.{1,5})"8 did this with the subject "ab".
1505    
1506    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1507        PCRE_CASELESS was set when matching characters that were quantified with ?
1508        or *.
1509    
1510    17. A character class other than a single negated character that had a minimum
1511        but no maximum quantifier - for example [ab]{6,} - was not handled
1512        correctly by pce_dfa_exec(). It would match only one character.
1513    
1514    18. A valid (though odd) pattern that looked like a POSIX character
1515        class but used an invalid character after [ (for example [[,abc,]]) caused
1516        pcre_compile() to give the error "Failed: internal error: code overflow" or
1517        in some cases to crash with a glibc free() error. This could even happen if
1518        the pattern terminated after [[ but there just happened to be a sequence of
1519        letters, a binary zero, and a closing ] in the memory that followed.
1520    
1521    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1522        over the years. Originally (before any Unicode support), just the bottom 8
1523        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1524        output from "man perlunicode" includes this:
1525    
1526          The regular expression compiler produces polymorphic opcodes.  That
1527          is, the pattern adapts to the data and automatically switches to
1528          the Unicode character scheme when presented with Unicode data--or
1529          instead uses a traditional byte scheme when presented with byte
1530          data.
1531    
1532        Sadly, a wide octal escape does not cause a switch, and in a string with
1533        no other multibyte characters, these octal escapes are treated as before.
1534        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1535        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1536        Unicode string.
1537    
1538        I have not perpetrated such confusion in PCRE. Up till now, it took just
1539        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1540        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1541        translate to the appropriate multibyte character.
1542    
1543    29. Applied some refactoring to reduce the number of warnings from Microsoft
1544        and Borland compilers. This has included removing the fudge introduced
1545        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1546        a warning about an unused variable.
1547    
1548    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1549        characters since release 4.0, because Perl (from release 5.004) does not.
1550        [Or at least, is documented not to: some releases seem to be in conflict
1551        with the documentation.] However, when a pattern was studied with
1552        pcre_study() and all its branches started with \s, PCRE still included VT
1553        as a possible starting character. Of course, this did no harm; it just
1554        caused an unnecessary match attempt.
1555    
1556    22. Removed a now-redundant internal flag bit that recorded the fact that case
1557        dependency changed within the pattern. This was once needed for "required
1558        byte" processing, but is no longer used. This recovers a now-scarce options
1559        bit. Also moved the least significant internal flag bit to the most-
1560        significant bit of the word, which was not previously used (hangover from
1561        the days when it was an int rather than a uint) to free up another bit for
1562        the future.
1563    
1564    23. Added support for CRLF line endings as well as CR and LF. As well as the
1565        default being selectable at build time, it can now be changed at runtime
1566        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1567        specify that it is scanning data with non-default line endings.
1568    
1569    24. Changed the definition of CXXLINK to make it agree with the definition of
1570        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1571    
1572    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1573        recursions. This makes a big different to stack usage for some patterns.
1574    
1575    26. If a subpattern containing a named recursion or subroutine reference such
1576        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1577        the space required for the compiled pattern went wrong and gave too small a
1578        value. Depending on the environment, this could lead to "Failed: internal
1579        error: code overflow at offset 49" or "glibc detected double free or
1580        corruption" errors.
1581    
1582    27. Applied patches from Google (a) to support the new newline modes and (b) to
1583        advance over multibyte UTF-8 characters in GlobalReplace.
1584    
1585    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1586        difference for some implementation of PCRE in some Windows version.
1587    
1588    29. Added some extra testing facilities to pcretest:
1589    
1590        \q<number>   in a data line sets the "match limit" value
1591        \Q<number>   in a data line sets the "match recursion limt" value
1592        -S <number>  sets the stack size, where <number> is in megabytes
1593    
1594        The -S option isn't available for Windows.
1595    
1596    
1597    Version 6.6 06-Feb-06
1598    ---------------------
1599    
1600     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1601        in pcreposix.h. I have copied the definition from pcre.h.
1602    
1603     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1604        because pcre.h is no longer a built file.
1605    
1606     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1607        not normally included in the compiled code.
1608    
1609    
1610    Version 6.5 01-Feb-06
1611    ---------------------
1612    
1613     1. When using the partial match feature with pcre_dfa_exec(), it was not
1614        anchoring the second and subsequent partial matches at the new starting
1615        point. This could lead to incorrect results. For example, with the pattern
1616        /1234/, partially matching against "123" and then "a4" gave a match.
1617    
1618     2. Changes to pcregrep:
1619    
1620        (a) All non-match returns from pcre_exec() were being treated as failures
1621            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1622            error message is output. Some extra information is given for the
1623            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1624            probably the only errors that are likely to be caused by users (by
1625            specifying a regex that has nested indefinite repeats, for instance).
1626            If there are more than 20 of these errors, pcregrep is abandoned.
1627    
1628        (b) A binary zero was treated as data while matching, but terminated the
1629            output line if it was written out. This has been fixed: binary zeroes
1630            are now no different to any other data bytes.
1631    
1632        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1633            used to set a locale for matching. The --locale=xxxx long option has
1634            been added (no short equivalent) to specify a locale explicitly on the
1635            pcregrep command, overriding the environment variables.
1636    
1637        (d) When -B was used with -n, some line numbers in the output were one less
1638            than they should have been.
1639    
1640        (e) Added the -o (--only-matching) option.
1641    
1642        (f) If -A or -C was used with -c (count only), some lines of context were
1643            accidentally printed for the final match.
1644    
1645        (g) Added the -H (--with-filename) option.
1646    
1647        (h) The combination of options -rh failed to suppress file names for files
1648            that were found from directory arguments.
1649    
1650        (i) Added the -D (--devices) and -d (--directories) options.
1651    
1652        (j) Added the -F (--fixed-strings) option.
1653    
1654        (k) Allow "-" to be used as a file name for -f as well as for a data file.
1655    
1656        (l) Added the --colo(u)r option.
1657    
1658        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1659            is not present by default.
1660    
1661     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1662        items such as (?R) or (?1), when the recursion could match a number of
1663        alternatives. If it matched one of the alternatives, but subsequently,
1664        outside the recursion, there was a failure, the code tried to back up into
1665        the recursion. However, because of the way PCRE is implemented, this is not
1666        possible, and the result was an incorrect result from the match.
1667    
1668        In order to prevent this happening, the specification of recursion has
1669        been changed so that all such subpatterns are automatically treated as
1670        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1671    
1672     4. I had overlooked the fact that, in some locales, there are characters for
1673        which isalpha() is true but neither isupper() nor islower() are true. In
1674        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1675        and ordfeminine) are like this. This affected the treatment of \w and \W
1676        when they appeared in character classes, but not when they appeared outside
1677        a character class. The bit map for "word" characters is now created
1678        separately from the results of isalnum() instead of just taking it from the
1679        upper, lower, and digit maps. (Plus the underscore character, of course.)
1680    
1681     5. The above bug also affected the handling of POSIX character classes such as
1682        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1683        permanent tables. Instead, the bit maps for such a class were previously
1684        created as the appropriate unions of the upper, lower, and digit bitmaps.
1685        Now they are created by subtraction from the [[:word:]] class, which has
1686        its own bitmap.
1687    
1688     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1689        It is created by subtracting the vertical space characters (\x09, \x0a,
1690        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1691        subtraction was done in the overall bitmap for a character class, meaning
1692        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1693        be recognized. This bug has been fixed.
1694    
1695     7. Patches from the folks at Google:
1696    
1697          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1698          real life, but is still worth protecting against".
1699    
1700          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1701          regular expressions".
1702    
1703          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1704          have it.
1705    
1706          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1707          "configure" and the latter not, in order to fix a problem somebody had
1708          with compiling the Arg class on HP-UX.
1709    
1710          (e) Improve the error-handling of the C++ wrapper a little bit.
1711    
1712          (f) New tests for checking recursion limiting.
1713    
1714     8. The pcre_memmove() function, which is used only if the environment does not
1715        have a standard memmove() function (and is therefore rarely compiled),
1716        contained two bugs: (a) use of int instead of size_t, and (b) it was not
1717        returning a result (though PCRE never actually uses the result).
1718    
1719     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1720        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1721        returned instead of calling malloc() with an overflowing number that would
1722        most likely cause subsequent chaos.
1723    
1724    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1725    
1726    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1727        with this option is matched, the nmatch and pmatch options of regexec() are
1728        ignored.
1729    
1730    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1731        provided in case anyone wants to the the POSIX interface with UTF-8
1732        strings.
1733    
1734    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1735        C++ linking (needed for some HP-UX environments).
1736    
1737    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1738        (unused parameter) and in the pcre_printint() function (omitted "default"
1739        switch label when the default is to do nothing).
1740    
1741    15. Added some code to make it possible, when PCRE is compiled as a C++
1742        library, to replace subject pointers for pcre_exec() with a smart pointer
1743        class, thus making it possible to process discontinuous strings.
1744    
1745    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1746        much the same function. They were added by different people who were trying
1747        to make PCRE easy to compile on non-Unix systems. It has been suggested
1748        that PCRE_EXPORT be abolished now that there is more automatic apparatus
1749        for compiling on Windows systems. I have therefore replaced it with
1750        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1751        defaults to "extern" for C or "extern C" for C++, which works fine on
1752        Unix-like systems. It is now possible to override the value of PCRE_DATA_
1753        SCOPE with something explicit in config.h. In addition:
1754    
1755        (a) pcreposix.h still had just "extern" instead of either of these macros;
1756            I have replaced it with PCRE_DATA_SCOPE.
1757    
1758        (b) Functions such as _pcre_xclass(), which are internal to the library,
1759            but external in the C sense, all had PCRE_EXPORT in their definitions.
1760            This is apparently wrong for the Windows case, so I have removed it.
1761            (It makes no difference on Unix-like systems.)
1762    
1763    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1764        of recursive calls to match(). This is different to MATCH_LIMIT because
1765        that limits the total number of calls to match(), not all of which increase
1766        the depth of recursion. Limiting the recursion depth limits the amount of
1767        stack (or heap if NO_RECURSE is set) that is used. The default can be set
1768        when PCRE is compiled, and changed at run time. A patch from Google adds
1769        this functionality to the C++ interface.
1770    
1771    18. Changes to the handling of Unicode character properties:
1772    
1773        (a) Updated the table to Unicode 4.1.0.
1774    
1775        (b) Recognize characters that are not in the table as "Cn" (undefined).
1776    
1777        (c) I revised the way the table is implemented to a much improved format
1778            which includes recognition of ranges. It now supports the ranges that
1779            are defined in UnicodeData.txt, and it also amalgamates other
1780            characters into ranges. This has reduced the number of entries in the
1781            table from around 16,000 to around 3,000, thus reducing its size
1782            considerably. I realized I did not need to use a tree structure after
1783            all - a binary chop search is just as efficient. Having reduced the
1784            number of entries, I extended their size from 6 bytes to 8 bytes to
1785            allow for more data.
1786    
1787        (d) Added support for Unicode script names via properties such as \p{Han}.
1788    
1789    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1790        matching that character.
1791    
1792    20. When matching a repeated Unicode property with a minimum greater than zero,
1793        (for example \pL{2,}), PCRE could look past the end of the subject if it
1794        reached it while seeking the minimum number of characters. This could
1795        happen only if some of the characters were more than one byte long, because
1796        there is a check for at least the minimum number of bytes.
1797    
1798    21. Refactored the implementation of \p and \P so as to be more general, to
1799        allow for more different types of property in future. This has changed the
1800        compiled form incompatibly. Anybody with saved compiled patterns that use
1801        \p or \P will have to recompile them.
1802    
1803    22. Added "Any" and "L&" to the supported property types.
1804    
1805    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1806        but give a compile time error if the value is greater than 0xff.
1807    
1808    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1809        accidentally not being installed or uninstalled.
1810    
1811    25. The pcre.h file was built from pcre.h.in, but the only changes that were
1812        made were to insert the current release number. This seemed silly, because
1813        it made things harder for people building PCRE on systems that don't run
1814        "configure". I have turned pcre.h into a distributed file, no longer built
1815        by "configure", with the version identification directly included. There is
1816        no longer a pcre.h.in file.
1817    
1818        However, this change necessitated a change to the pcre-config script as
1819        well. It is built from pcre-config.in, and one of the substitutions was the
1820        release number. I have updated configure.ac so that ./configure now finds
1821        the release number by grepping pcre.h.
1822    
1823    26. Added the ability to run the tests under valgrind.
1824    
1825    
1826    Version 6.4 05-Sep-05
1827    ---------------------
1828    
1829     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1830        "--" to be printed when multiple files were scanned, even when none of the
1831        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1832        consider it to be a bug, and have restored the previous behaviour.
1833    
1834     2. A couple of code tidies to get rid of compiler warnings.
1835    
1836     3. The pcretest program used to cheat by referring to symbols in the library
1837        whose names begin with _pcre_. These are internal symbols that are not
1838        really supposed to be visible externally, and in some environments it is
1839        possible to suppress them. The cheating is now confined to including
1840        certain files from the library's source, which is a bit cleaner.
1841    
1842     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1843        file's purpose clearer.
1844    
1845     5. Reorganized pcre_ucp_findchar().
1846    
1847    
1848    Version 6.3 15-Aug-05
1849    ---------------------
1850    
1851     1. The file libpcre.pc.in did not have general read permission in the tarball.
1852    
1853     2. There were some problems when building without C++ support:
1854    
1855        (a) If C++ support was not built, "make install" and "make test" still
1856            tried to test it.
1857    
1858        (b) There were problems when the value of CXX was explicitly set. Some
1859            changes have been made to try to fix these, and ...
1860    
1861        (c) --disable-cpp can now be used to explicitly disable C++ support.
1862    
1863        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
1864            backslash in a target when C++ was disabled. This confuses some
1865            versions of "make", apparently. Using an intermediate variable solves
1866            this. (Same for CPP_LOBJ.)
1867    
1868     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
1869        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
1870        necessary on certain architectures.
1871    
1872     4. Added a setting of -export-symbols-regex to the link command to remove
1873        those symbols that are exported in the C sense, but actually are local
1874        within the library, and not documented. Their names all begin with
1875        "_pcre_". This is not a perfect job, because (a) we have to except some
1876        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
1877        available (and never for static libraries). I have made a note to try to
1878        find a way round (a) in the future.
1879    
1880    
1881    Version 6.2 01-Aug-05
1882    ---------------------
1883    
1884     1. There was no test for integer overflow of quantifier values. A construction
1885        such as {1111111111111111} would give undefined results. What is worse, if
1886        a minimum quantifier for a parenthesized subpattern overflowed and became
1887        negative, the calculation of the memory size went wrong. This could have
1888        led to memory overwriting.
1889    
1890     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
1891    
1892     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
1893        operating environments where this matters.
1894    
1895     4. Applied Giuseppe Maxia's patch to add additional features for controlling
1896        PCRE options from within the C++ wrapper.
1897    
1898     5. Named capturing subpatterns were not being correctly counted when a pattern
1899        was compiled. This caused two problems: (a) If there were more than 100
1900        such subpatterns, the calculation of the memory needed for the whole
1901        compiled pattern went wrong, leading to an overflow error. (b) Numerical
1902        back references of the form \12, where the number was greater than 9, were
1903        not recognized as back references, even though there were sufficient
1904        previous subpatterns.
1905    
1906     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
1907        versions of gcc, e.g. 2.95.4.
1908    
1909    
1910    Version 6.1 21-Jun-05
1911    ---------------------
1912    
1913     1. There was one reference to the variable "posix" in pcretest.c that was not
1914        surrounded by "#if !defined NOPOSIX".
1915    
1916     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
1917        the cross-check on the old pcre_info() function, for the benefit of the
1918        cut-down version of PCRE that is currently imported into Exim.
1919    
1920     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
1921        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
1922        patterns that start (?i) but I don't think that matters. The use of (?i) is
1923        just an example; this all applies to the other options as well.
1924    
1925     4. Since libtool seems to echo the compile commands it is issuing, the output
1926        from "make" can be reduced a bit by putting "@" in front of each libtool
1927        compile command.
1928    
1929     5. Patch from the folks at Google for configure.in to be a bit more thorough
1930        in checking for a suitable C++ installation before trying to compile the
1931        C++ stuff. This should fix a reported problem when a compiler was present,
1932        but no suitable headers.
1933    
1934     6. The man pages all had just "PCRE" as their title. I have changed them to
1935        be the relevant file name. I have also arranged that these names are
1936        retained in the file doc/pcre.txt, which is a concatenation in text format
1937        of all the man pages except the little individual ones for each function.
1938    
1939     7. The NON-UNIX-USE file had not been updated for the different set of source
1940        files that come with release 6. I also added a few comments about the C++
1941        wrapper.
1942    
1943    
1944    Version 6.0 07-Jun-05
1945    ---------------------
1946    
1947     1. Some minor internal re-organization to help with my DFA experiments.
1948    
1949     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
1950        didn't matter for the library itself when fully configured, but did matter
1951        when compiling without UCP support, or within Exim, where the ucp files are
1952        not imported.
1953    
1954     3. Refactoring of the library code to split up the various functions into
1955        different source modules. The addition of the new DFA matching code (see
1956        below) to a single monolithic source would have made it really too
1957        unwieldy, quite apart from causing all the code to be include in a
1958        statically linked application, when only some functions are used. This is
1959        relevant even without the DFA addition now that patterns can be compiled in
1960        one application and matched in another.
1961    
1962        The downside of splitting up is that there have to be some external
1963        functions and data tables that are used internally in different modules of
1964        the library but which are not part of the API. These have all had their
1965        names changed to start with "_pcre_" so that they are unlikely to clash
1966        with other external names.
1967    
1968     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
1969        a different (DFA) algorithm. Although it is slower than the original
1970        function, it does have some advantages for certain types of matching
1971        problem.
1972    
1973     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
1974        including restarting after a partial match.
1975    
1976     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
1977        defined when compiling for Windows was sent to me. I have put it into the
1978        code, though I have no means of testing or verifying it.
1979    
1980     7. Added the pcre_refcount() auxiliary function.
1981    
1982     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
1983        match before or at the first newline in the subject string. In pcretest,
1984        the /f option on a pattern can be used to set this.
1985    
1986     9. A repeated \w when used in UTF-8 mode with characters greater than 256
1987        would behave wrongly. This has been present in PCRE since release 4.0.
1988    
1989    10. A number of changes to the pcregrep command:
1990    
1991        (a) Refactored how -x works; insert ^(...)$ instead of setting
1992            PCRE_ANCHORED and checking the length, in preparation for adding
1993            something similar for -w.
1994    
1995        (b) Added the -w (match as a word) option.
1996    
1997        (c) Refactored the way lines are read and buffered so as to have more
1998            than one at a time available.
1999    
2000        (d) Implemented a pcregrep test script.
2001    
2002        (e) Added the -M (multiline match) option. This allows patterns to match
2003            over several lines of the subject. The buffering ensures that at least
2004            8K, or the rest of the document (whichever is the shorter) is available
2005            for matching (and similarly the previous 8K for lookbehind assertions).
2006    
2007        (f) Changed the --help output so that it now says
2008    
2009              -w, --word-regex(p)
2010    
2011            instead of two lines, one with "regex" and the other with "regexp"
2012            because that confused at least one person since the short forms are the
2013            same. (This required a bit of code, as the output is generated
2014            automatically from a table. It wasn't just a text change.)
2015    
2016        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2017            option but starts with a hyphen. Could be a pattern or a path name
2018            starting with a hyphen, for instance.
2019    
2020        (h) "-" can be given as a file name to represent stdin.
2021    
2022        (i) When file names are being printed, "(standard input)" is used for
2023            the standard input, for compatibility with GNU grep. Previously
2024            "<stdin>" was used.
2025    
2026        (j) The option --label=xxx can be used to supply a name to be used for
2027            stdin when file names are being printed. There is no short form.
2028    
2029        (k) Re-factored the options decoding logic because we are going to add
2030            two more options that take data. Such options can now be given in four
2031            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2032    
2033        (l) Added the -A, -B, and -C options for requesting that lines of context
2034            around matches be printed.
2035    
2036        (m) Added the -L option to print the names of files that do not contain
2037            any matching lines, that is, the complement of -l.
2038    
2039        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2040            continue to scan other files.
2041    
2042        (o) The -s option was incorrectly implemented. For compatibility with other
2043            greps, it now suppresses the error message for a non-existent or non-
2044            accessible file (but not the return code). There is a new option called
2045            -q that suppresses the output of matching lines, which was what -s was
2046            previously doing.
2047    
2048        (p) Added --include and --exclude options to specify files for inclusion
2049            and exclusion when recursing.
2050    
2051    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2052        Hopefully, it now does.
2053    
2054    12. Missing cast in pcre_study().
2055    
2056    13. Added an "uninstall" target to the makefile.
2057    
2058    14. Replaced "extern" in the function prototypes in Makefile.in with
2059        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2060        world, but is set differently for Windows.
2061    
2062    15. Added a second compiling function called pcre_compile2(). The only
2063        difference is that it has an extra argument, which is a pointer to an
2064        integer error code. When there is a compile-time failure, this is set
2065        non-zero, in addition to the error test pointer being set to point to an
2066        error message. The new argument may be NULL if no error number is required
2067        (but then you may as well call pcre_compile(), which is now just a
2068        wrapper). This facility is provided because some applications need a
2069        numeric error indication, but it has also enabled me to tidy up the way
2070        compile-time errors are handled in the POSIX wrapper.
2071    
2072    16. Added VPATH=.libs to the makefile; this should help when building with one
2073        prefix path and installing with another. (Or so I'm told by someone who
2074        knows more about this stuff than I do.)
2075    
2076    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2077        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2078        match everything, including newlines. This is not POSIX-compatible, but
2079        somebody wanted the feature. From pcretest it can be activated by using
2080        both the P and the s flags.
2081    
2082    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2083    
2084    19. libpcre.pc was being incorrectly installed as executable.
2085    
2086    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2087        it now also looks for '\r' so that it will work unmodified on Windows.
2088    
2089    21. Added Google's contributed C++ wrapper to the distribution.
2090    
2091    22. Added some untidy missing memory free() calls in pcretest, to keep
2092        Electric Fence happy when testing.
2093    
2094    
2095    
2096    Version 5.0 13-Sep-04
2097    ---------------------
2098    
2099     1. Internal change: literal characters are no longer packed up into items
2100        containing multiple characters in a single byte-string. Each character
2101        is now matched using a separate opcode. However, there may be more than one
2102        byte in the character in UTF-8 mode.
2103    
2104     2. The pcre_callout_block structure has two new fields: pattern_position and
2105        next_item_length. These contain the offset in the pattern to the next match
2106        item, and its length, respectively.
2107    
2108     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2109        insertion of callouts before each pattern item. Added the /C option to
2110        pcretest to make use of this.
2111    
2112     4. On the advice of a Windows user, the lines
2113    
2114          #if defined(_WIN32) || defined(WIN32)
2115          _setmode( _fileno( stdout ), 0x8000 );
2116          #endif  /* defined(_WIN32) || defined(WIN32) */
2117    
2118        have been added to the source of pcretest. This apparently does useful
2119        magic in relation to line terminators.
2120    
2121     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2122        for the benefit of those environments where the "b" makes a difference.
2123    
2124     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2125        to know about it. I have put a hack into configure.in that adds in code
2126        to set GCC=yes if CC=icc. This seems to end up at a point in the
2127        generated configure script that is early enough to affect the setting of
2128        compiler options, which is what is needed, but I have no means of testing
2129        whether it really works. (The user who reported this had patched the
2130        generated configure script, which of course I cannot do.)
2131    
2132        LATER: After change 22 below (new libtool files), the configure script
2133        seems to know about icc (and also ecc). Therefore, I have commented out
2134        this hack in configure.in.
2135    
2136     7. Added support for pkg-config (2 patches were sent in).
2137    
2138     8. Negated POSIX character classes that used a combination of internal tables
2139        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2140        [[:^ascii]]. Typically, they would match almost any characters. The other
2141        POSIX classes were not broken in this way.
2142    
2143     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2144        to find the match, as PCRE was deluded into thinking that the match had to
2145        start at the start point or following a newline. The same bug applied to
2146        patterns with negative forward assertions or any backward assertions
2147        preceding ".*" at the start, unless the pattern required a fixed first
2148        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2149    
2150    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2151        starting at the last subject character, bytes beyond the end of the subject
2152        string were read.
2153    
2154    11. Renamed the variable "class" as "classbits" to make life easier for C++
2155        users. (Previously there was a macro definition, but it apparently wasn't
2156        enough.)
2157    
2158    12. Added the new field "tables" to the extra data so that tables can be passed
2159        in at exec time, or the internal tables can be re-selected. This allows
2160        a compiled regex to be saved and re-used at a later time by a different
2161        program that might have everything at different addresses.
2162    
2163    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2164        -R library as well as a -L library.
2165    
2166    14. The debugging options of pcretest (-d on the command line or D on a
2167        pattern) showed incorrect output for anything following an extended class
2168        that contained multibyte characters and which was followed by a quantifier.
2169    
2170    15. Added optional support for general category Unicode character properties
2171        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2172        support. It adds about 90K to the size of the library. The meanings of the
2173        inbuilt class escapes such as \d and \s have NOT been changed.
2174    
2175    16. Updated pcredemo.c to include calls to free() to release the memory for the
2176        compiled pattern.
2177    
2178    17. The generated file chartables.c was being created in the source directory
2179        instead of in the building directory. This caused the build to fail if the
2180        source directory was different from the building directory, and was
2181        read-only.
2182    
2183    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2184        file. No doubt somebody will tell me if they don't make sense... Also added
2185        Dan Mooney's comments about building on OpenVMS.
2186    
2187    19. Added support for partial matching via the PCRE_PARTIAL option for
2188        pcre_exec() and the \P data escape in pcretest.
2189    
2190    20. Extended pcretest with 3 new pattern features:
2191    
2192        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2193              write the compiled pattern to the file whose name is "rest-of-line".
2194              This is a straight binary dump of the data, with the saved pointer to
2195              the character tables forced to be NULL. The study data, if any, is
2196              written too. After writing, pcretest reads a new pattern.
2197    
2198        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2199              compiled pattern from the given file. There must not be any
2200              occurrences of "<" in the file name (pretty unlikely); if there are,
2201              pcretest will instead treat the initial "<" as a pattern delimiter.
2202              After reading in the pattern, pcretest goes on to read data lines as
2203              usual.
2204    
2205        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2206              and 16-bit fields in a compiled pattern, to simulate a pattern that
2207              was compiled on a host of opposite endianness.
2208    
2209    21. The pcre-exec() function can now cope with patterns that were compiled on
2210        hosts of opposite endianness, with this restriction:
2211    
2212          As for any compiled expression that is saved and used later, the tables
2213          pointer field cannot be preserved; the extra_data field in the arguments
2214          to pcre_exec() should be used to pass in a tables address if a value
2215          other than the default internal tables were used at compile time.
2216    
2217    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2218        now diagnosed as an error. Previously, most of the time, a negative number
2219        would have been treated as zero, but if in addition "ovector" was passed as
2220        NULL, a crash could occur.
2221    
2222    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2223        new versions from the libtool 1.5 distribution (the last one is a copy of
2224        a file called libtool.m4). This seems to have fixed the need to patch
2225        "configure" to support Darwin 1.3 (which I used to do). However, I still
2226        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2227        workstation).
2228    
2229    24. Changed the PCRE licence to be the more standard "BSD" licence.
2230    
2231    
2232    Version 4.5 01-Dec-03
2233    ---------------------
2234    
2235     1. There has been some re-arrangement of the code for the match() function so
2236        that it can be compiled in a version that does not call itself recursively.
2237        Instead, it keeps those local variables that need separate instances for
2238        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2239        needs to "recurse". Keeping track of where control must go is done by means
2240        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2241        hide most of the details from the main code, and operates only if
2242        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2243        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2244        operating.
2245    
2246        To make it easier for callers to provide specially tailored get/free
2247        functions for this usage, two new functions, pcre_stack_malloc, and
2248        pcre_stack_free, are used. They are always called in strict stacking order,
2249        and the size of block requested is always the same.
2250    
2251        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2252        PCRE has been compiled to use the stack or the heap for recursion. The
2253        -C option of pcretest uses this to show which version is compiled.
2254    
2255        A new data escape \S, is added to pcretest; it causes the amounts of store
2256        obtained and freed by both kinds of malloc/free at match time to be added
2257        to the output.
2258    
2259     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2260        what's available on my current Linux desktop machine.
2261    
2262     3. When matching a UTF-8 string, the test for a valid string at the start has
2263        been extended. If start_offset is not zero, PCRE now checks that it points
2264        to a byte that is the start of a UTF-8 character. If not, it returns
2265        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2266        this is necessary because there may be backward assertions in the pattern.
2267        When matching the same subject several times, it may save resources to use
2268        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2269    
2270     4. The code for checking the validity of UTF-8 strings has been tightened so
2271        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2272        containing "overlong sequences".
2273    
2274     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2275        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2276        should have been "&", but it just so happened that all the cases this let
2277        through by mistake were picked up later in the function.
2278    
2279     6. I had used a variable called "isblank" - this is a C99 function, causing
2280        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2281    
2282     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2283        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2284        so the version is shown; (c) stop "make test" echoing "./RunTest".
2285    
2286     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2287    
2288     9. The prototype for memmove() for systems that don't have it was using
2289        size_t, but the inclusion of the header that defines size_t was later. I've
2290        moved the #includes for the C headers earlier to avoid this.
2291    
2292    10. Added some adjustments to the code to make it easier to compiler on certain
2293        special systems:
2294    
2295          (a) Some "const" qualifiers were missing.
2296          (b) Added the macro EXPORT before all exported functions; by default this
2297              is defined to be empty.
2298          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2299              that it reads its output file name as an argument instead of writing
2300              to the standard output and assuming this can be redirected.
2301    
2302    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2303        class containing characters with values greater than 255, PCRE compilation
2304        went into a loop.
2305    
2306    12. A recursive reference to a subpattern that was within another subpattern
2307        that had a minimum quantifier of zero caused PCRE to crash. For example,
2308        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2309        recursion. If the recursively-called subpattern itself had a zero repeat,
2310        that was OK.
2311    
2312    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2313        buffer into which it was copied (for escape processing) was still set at
2314        1024, so long lines caused crashes.
2315    
2316    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2317        "internal error: code overflow...". This applied to any character class
2318        that was followed by a possessive quantifier.
2319    
2320    15. Modified the Makefile to add libpcre.la as a prerequisite for
2321        libpcreposix.la because I was told this is needed for a parallel build to
2322        work.
2323    
2324    16. If a pattern that contained .* following optional items at the start was
2325        studied, the wrong optimizing data was generated, leading to matching
2326        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2327        matching string must start with a or b or c. The correct conclusion for
2328        this pattern is that a match can start with any character.
2329    
2330    
2331    Version 4.4 13-Aug-03
2332    ---------------------
2333    
2334     1. In UTF-8 mode, a character class containing characters with values between
2335        127 and 255 was not handled correctly if the compiled pattern was studied.
2336        In fixing this, I have also improved the studying algorithm for such
2337        classes (slightly).
2338    
2339     2. Three internal functions had redundant arguments passed to them. Removal
2340        might give a very teeny performance improvement.
2341    
2342     3. Documentation bug: the value of the capture_top field in a callout is *one
2343        more than* the number of the hightest numbered captured substring.
2344    
2345     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2346        in incorrectly linking with a previously installed version. They now link
2347        explicitly with libpcre.la.
2348    
2349     5. configure.in no longer needs to recognize Cygwin specially.
2350    
2351     6. A problem in pcre.in for Windows platforms is fixed.
2352    
2353     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2354        pcretest, it used to include the size of the study block as part of its
2355        output. Unfortunately, the structure contains a field that has a different
2356        size on different hardware architectures. This meant that the tests that
2357        showed this size failed. As the block is currently always of a fixed size,
2358        this information isn't actually particularly useful in pcretest output, so
2359        I have just removed it.
2360    
2361     8. Three pre-processor statements accidentally did not start in column 1.
2362        Sadly, there are *still* compilers around that complain, even though
2363        standard C has not required this for well over a decade. Sigh.
2364    
2365     9. In pcretest, the code for checking callouts passed small integers in the
2366        callout_data field, which is a void * field. However, some picky compilers
2367        complained about the casts involved for this on 64-bit systems. Now
2368        pcretest passes the address of the small integer instead, which should get
2369        rid of the warnings.
2370    
2371    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2372        both compile and run time, and gives an error if an invalid UTF-8 sequence
2373        is found. There is a option for disabling this check in cases where the
2374        string is known to be correct and/or the maximum performance is wanted.
2375    
2376    11. In response to a bug report, I changed one line in Makefile.in from
2377    
2378            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2379        to
2380            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2381    
2382        to look similar to other lines, but I have no way of telling whether this
2383        is the right thing to do, as I do not use Windows. No doubt I'll get told
2384        if it's wrong...
2385    
2386    
2387    Version 4.3 21-May-03
2388    ---------------------
2389    
2390    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2391       Makefile.
2392    
2393    2. Some refactoring to improve the quality of the code:
2394    
2395       (i)   The utf8_table... variables are now declared "const".
2396    
2397       (ii)  The code for \cx, which used the "case flipping" table to upper case
2398             lower case letters, now just substracts 32. This is ASCII-specific,
2399             but the whole concept of \cx is ASCII-specific, so it seems
2400             reasonable.
2401    
2402       (iii) PCRE was using its character types table to recognize decimal and
2403             hexadecimal digits in the pattern. This is silly, because it handles
2404             only 0-9, a-f, and A-F, but the character types table is locale-
2405             specific, which means strange things might happen. A private
2406             table is now used for this - though it costs 256 bytes, a table is
2407             much faster than multiple explicit tests. Of course, the standard
2408             character types table is still used for matching digits in subject
2409             strings against \d.
2410    
2411       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2412             ending in _t are). So I've renamed it as ESC_tee.
2413    
2414    3. The first argument for regexec() in the POSIX wrapper should have been
2415       defined as "const".
2416    
2417    4. Changed pcretest to use malloc() for its buffers so that they can be
2418       Electric Fenced for debugging.
2419    
2420    5. There were several places in the code where, in UTF-8 mode, PCRE would try
2421       to read one or more bytes before the start of the subject string. Often this
2422       had no effect on PCRE's behaviour, but in some circumstances it could
2423       provoke a segmentation fault.
2424    
2425    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2426       to try to read one or more bytes before the start of the subject string.
2427    
2428    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2429       UTF-8 support could misbehave in various ways if the subject string
2430       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2431       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2432       back over UTF-8 characters.)
2433    
2434    
2435    Version 4.2 14-Apr-03
2436    ---------------------
2437    
2438    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2439    
2440    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2441         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2442         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2443         [WIN_PREFIX]: new variable, "cyg" for Cygwin
2444         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2445           and BUILD_EXEEXT
2446         Note: automatic setting of the BUILD variables is not yet working
2447         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2448           compile-time but not at link-time
2449         [LINK]: use for linking executables only
2450         make different versions for Windows and non-Windows
2451         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2452           libraries
2453         [LINK_FOR_BUILD]: new variable
2454         [OBJEXT]: use throughout
2455         [EXEEXT]: use throughout
2456         <winshared>: new target
2457         <wininstall>: new target
2458         <dftables.o>: use native compiler
2459         <dftables>: use native linker
2460         <install>: handle Windows platform correctly
2461         <clean>: ditto
2462         <check>: ditto
2463         copy DLL to top builddir before testing
2464    
2465       As part of these changes, -no-undefined was removed again. This was reported
2466       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2467       in any case.
2468    
2469    3. Some tidies to get rid of compiler warnings:
2470    
2471       . In the match_data structure, match_limit was an unsigned long int, whereas
2472         match_call_count was an int. I've made them both unsigned long ints.
2473    
2474       . In pcretest the fact that a const uschar * doesn't automatically cast to
2475         a void * provoked a warning.
2476    
2477       . Turning on some more compiler warnings threw up some "shadow" variables
2478         and a few more missing casts.
2479    
2480    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2481       option, a class that contained a single character with a value between 128
2482       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2483    
2484    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2485       option, a class that contained several characters, but with at least one
2486       whose value was between 128 and 255 caused PCRE to crash.
2487    
2488    
2489    Version 4.1 12-Mar-03
2490    ---------------------
2491    
2492    1. Compiling with gcc -pedantic found a couple of places where casts were
2493    needed, and a string in dftables.c that was longer than standard compilers are
2494    required to support.
2495    
2496    2. Compiling with Sun's compiler found a few more places where the code could
2497    be tidied up in order to avoid warnings.
2498    
2499    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2500    first of these names is deprecated in the latest Autoconf in favour of the name
2501    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2502    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2503    analogy I have changed it to CFLAGS_FOR_BUILD.
2504    
2505    4. Added -no-undefined to the linking command in the Makefile, because this is
2506    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2507    linking step for the pcreposix library.
2508    
2509    5. PCRE was failing to diagnose the case of two named groups with the same
2510    name.
2511    
2512    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2513    literal character that is needed in the subject for a match, and scans along to
2514    ensure that it is present before embarking on the full matching process. This
2515    saves time in cases of nested unlimited repeats that are never going to match.
2516    Problem: the scan can take a lot of time if the subject is very long (e.g.
2517    megabytes), thus penalizing straightforward matches. It is now done only if the
2518    amount of subject to be scanned is less than 1000 bytes.
2519    
2520    7. A lesser problem with the same optimization is that it was recording the
2521    first character of an anchored pattern as "needed", thus provoking a search
2522    right along the subject, even when the first match of the pattern was going to
2523    fail. The "needed" character is now not set for anchored patterns, unless it
2524    follows something in the pattern that is of non-fixed length. Thus, it still
2525    fulfils its original purpose of finding quick non-matches in cases of nested
2526    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2527    
2528    
2529    Version 4.0 17-Feb-03
2530    ---------------------
2531    
2532    1. If a comment in an extended regex that started immediately after a meta-item
2533    extended to the end of string, PCRE compiled incorrect data. This could lead to
2534    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2535    
2536    2. Moved to autoconf 2.53 and libtool 1.4.2.
2537    
2538    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2539    the special perltest8 script is no longer needed - all the tests can be run
2540    from a single perltest script.
2541    
2542    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2543    by \s. It has now been removed in PCRE. This means it isn't recognized as
2544    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2545    class [:space:] *does* include VT, thereby creating a mess.
2546    
2547    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2548    space and tab.
2549    
2550    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2551    its new features into the main test script, reducing the number of scripts.
2552    
2553    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2554    were backward compatible, and made the (?i) apply to the whole pattern, as if
2555    /i were given. Now it behaves more logically, and applies the option setting
2556    only to what follows. PCRE has been changed to follow suit. However, if it
2557    finds options settings right at the start of the pattern, it extracts them into
2558    the global options, as before. Thus, they show up in the info data.
2559    
2560    8. Added support for the \Q...\E escape sequence. Characters in between are
2561    treated as literals. This is slightly different from Perl in that $ and @ are
2562    also handled as literals inside the quotes. In Perl, they will cause variable
2563    interpolation. Note the following examples:
2564    
2565        Pattern            PCRE matches      Perl matches
2566    
2567        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
2568        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
2569        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
2570    
2571    For compatibility with Perl, \Q...\E sequences are recognized inside character
2572    classes as well as outside them.
2573    
2574    9. Re-organized 3 code statements in pcretest to avoid "overflow in
2575    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2576    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2577    signed/unsigned warnings.
2578    
2579    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2580    option for pcretest, so I've replaced it by a simple function that does just
2581    that job.
2582    
2583    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2584    "pcregrep -".
2585    
2586    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2587    Java package. This provides some syntactic sugar for simple cases of what my
2588    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2589    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2590    item, you can use this simplified notation. Note that only makes sense with
2591    greedy quantifiers. Consequently, the use of the possessive quantifier forces
2592    greediness, whatever the setting of the PCRE_UNGREEDY option.
2593    
2594    13. A change of greediness default within a pattern was not taking effect at
2595    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2596    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2597    was abstracted outside.
2598    
2599    14. PCRE now supports the \G assertion. It is true when the current matching
2600    position is at the start point of the match. This differs from \A when the
2601    starting offset is non-zero. Used with the /g option of pcretest (or similar
2602    code), it works in the same way as it does for Perl's /g option. If all
2603    alternatives of a regex begin with \G, the expression is anchored to the start
2604    match position, and the "anchored" flag is set in the compiled expression.
2605    
2606    15. Some bugs concerning the handling of certain option changes within patterns
2607    have been fixed. These applied to options other than (?ims). For example,
2608    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2609    been the other way round. Some of this was related to change 7 above.
2610    
2611    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2612    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2613    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2614    POSIX classes only within a class (e.g. /[[:alpha:]]/).
2615    
2616    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2617    mode. Unlike ".", it always matches newline, whatever the setting of
2618    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2619    assertions. Perl allows it, but it doesn't (in general) work because it can't
2620    calculate the length of the lookbehind. At least, that's the case for Perl
2621    5.8.0 - I've been told they are going to document that it doesn't work in
2622    future.
2623    
2624    18. Added an error diagnosis for escapes that PCRE does not support: these are
2625    \L, \l, \N, \P, \p, \U, \u, and \X.
2626    
2627    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2628    reading past the end of the pattern in cases such as /[abcd/.
2629    
2630    20. PCRE was getting more memory than necessary for patterns with classes that
2631    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2632    
2633    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2634    compiling PCRE for use with Virtual Pascal.
2635    
2636    22. Small fix to the Makefile to make it work properly if the build is done
2637    outside the source tree.
2638    
2639    23. Added a new extension: a condition to go with recursion. If a conditional
2640    subpattern starts with (?(R) the "true" branch is used if recursion has
2641    happened, whereas the "false" branch is used only at the top level.
2642    
2643    24. When there was a very long string of literal characters (over 255 bytes
2644    without UTF support, over 250 bytes with UTF support), the computation of how
2645    much memory was required could be incorrect, leading to segfaults or other
2646    strange effects.
2647    
2648    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2649    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2650    there was a subsequent back reference to those brackets. This meant that, for
2651    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2652    possible to check for precisely this case. All we can do is abandon the
2653    optimization if .* occurs inside capturing brackets when there are any back
2654    references whatsoever. (See below for a better fix that came later.)
2655    
2656    26. The handling of the optimization for finding the first character of a
2657    non-anchored pattern, and for finding a character that is required later in the
2658    match were failing in some cases. This didn't break the matching; it just
2659    failed to optimize when it could. The way this is done has been re-implemented.
2660    
2661    27. Fixed typo in error message for invalid (?R item (it said "(?p").
2662    
2663    28. Added a new feature that provides some of the functionality that Perl
2664    provides with (?{...}). The facility is termed a "callout". The way it is done
2665    in PCRE is for the caller to provide an optional function, by setting
2666    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2667    global variable. By default it is unset, which disables all calling out. To get
2668    the function called, the regex must include (?C) at appropriate points. This
2669    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2670    This provides a means of identifying different callout points. When PCRE
2671    reaches such a point in the regex, if pcre_callout has been set, the external
2672    function is called. It is provided with data in a structure called
2673    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2674    matching continues; if it returns a non-zero value, the match at the current
2675    point fails. However, backtracking will occur if possible. [This was changed
2676    later and other features added - see item 49 below.]
2677    
2678    29. pcretest is upgraded to test the callout functionality. It provides a
2679    callout function that displays information. By default, it shows the start of
2680    the match and the current position in the text. There are some new data escapes
2681    to vary what happens:
2682    
2683        \C+         in addition, show current contents of captured substrings
2684        \C-         do not supply a callout function
2685        \C!n        return 1 when callout number n is reached
2686        \C!n!m      return 1 when callout number n is reached for the mth time
2687    
2688    30. If pcregrep was called with the -l option and just a single file name, it
2689    output "<stdin>" if a match was found, instead of the file name.
2690    
2691    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2692    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2693    pcre_exec(). This saves a malloc/free per call. The default value of
2694    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2695    when configuring.
2696    
2697    32. The default maximum size of a compiled pattern is 64K. There have been a
2698    few cases of people hitting this limit. The code now uses macros to handle the
2699    storing of links as offsets within the compiled pattern. It defaults to 2-byte
2700    links, but this can be changed to 3 or 4 bytes by --with-link-size when
2701    configuring. Tests 2 and 5 work only with 2-byte links because they output
2702    debugging information about compiled patterns.
2703    
2704    33. Internal code re-arrangements:
2705    
2706    (a) Moved the debugging function for printing out a compiled regex into
2707        its own source file (printint.c) and used #include to pull it into
2708        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2709        separate copies.
2710    
2711    (b) Defined the list of op-code names for debugging as a macro in
2712        internal.h so that it is next to the definition of the opcodes.
2713    
2714    (c) Defined a table of op-code lengths for simpler skipping along compiled
2715        code. This is again a macro in internal.h so that it is next to the
2716        definition of the opcodes.
2717    
2718    34. Added support for recursive calls to individual subpatterns, along the
2719    lines of Robin Houston's patch (but implemented somewhat differently).
2720    
2721    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2722    allow it to read and process whole directories in Win32. This code was
2723    contributed by Lionel Fourquaux; it has not been tested by me.
2724    
2725    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2726    used to name a group. Names consist of alphanumerics and underscores, and must
2727    be unique. Back references use the syntax (?P=name) and recursive calls use
2728    (?P>name) which is a PCRE extension to the Python extension. Groups still have
2729    numbers. The function pcre_fullinfo() can be used after compilation to extract
2730    a name/number map. There are three relevant calls:
2731    
2732      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
2733      PCRE_INFO_NAMECOUNT            yields the number of entries
2734      PCRE_INFO_NAMETABLE            yields a pointer to the map.
2735    
2736    The map is a vector of fixed-size entries. The size of each entry depends on
2737    the length of the longest name used. The first two bytes of each entry are the
2738    group number, most significant byte first. There follows the corresponding
2739    name, zero terminated. The names are in alphabetical order.
2740    
2741    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2742    case instead of 255. Making it the same both with and without UTF-8 support
2743    means that the same test output works with both.
2744    
2745    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2746    calling malloc() with a zero argument.
2747    
2748    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2749    optimization. I've improved things by keeping a bitmap of backreferences with
2750    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2751    fact referenced, the optimization can be applied. It is unlikely that a
2752    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2753    the match to follow \n) will appear inside brackets with a number greater than
2754    31, but if it does, any back reference > 31 suppresses the optimization.
2755    
2756    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2757    of disabling numbered capturing parentheses. Any opening parenthesis that is
2758    not followed by ? behaves as if it were followed by ?: but named parentheses
2759    can still be used for capturing (and they will acquire numbers in the usual
2760    way).
2761    
2762    41. Redesigned the return codes from the match() function into yes/no/error so
2763    that errors can be passed back from deep inside the nested calls. A malloc
2764    failure while inside a recursive subpattern call now causes the
2765    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2766    
2767    42. It is now possible to set a limit on the number of times the match()
2768    function is called in a call to pcre_exec(). This facility makes it possible to
2769    limit the amount of recursion and backtracking, though not in a directly
2770    obvious way, because the match() function is used in a number of different
2771    circumstances. The count starts from zero for each position in the subject
2772    string (for non-anchored patterns). The default limit is, for compatibility, a
2773    large number, namely 10 000 000. You can change this in two ways:
2774    
2775    (a) When configuring PCRE before making, you can use --with-match-limit=n
2776        to set a default value for the compiled library.
2777    
2778    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2779        a different value is set. See 45 below.
2780    
2781    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2782    
2783    43. Added a new function pcre_config(int, void *) to enable run-time extraction
2784    of things that can be changed at compile time. The first argument specifies
2785    what is wanted and the second points to where the information is to be placed.
2786    The current list of available information is:
2787    
2788      PCRE_CONFIG_UTF8
2789    
2790    The output is an integer that is set to one if UTF-8 support is available;
2791    otherwise it is set to zero.
2792    
2793      PCRE_CONFIG_NEWLINE
2794    
2795    The output is an integer that it set to the value of the code that is used for
2796    newline. It is either LF (10) or CR (13).
2797    
2798      PCRE_CONFIG_LINK_SIZE
2799    
2800    The output is an integer that contains the number of bytes used for internal
2801    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2802    
2803      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2804    
2805    The output is an integer that contains the threshold above which the POSIX
2806    interface uses malloc() for output vectors. See item 31 above.
2807    
2808      PCRE_CONFIG_MATCH_LIMIT
2809    
2810    The output is an unsigned integer that contains the default limit of the number
2811    of match() calls in a pcre_exec() execution. See 42 above.
2812    
2813    44. pcretest has been upgraded by the addition of the -C option. This causes it
2814    to extract all the available output from the new pcre_config() function, and to
2815    output it. The program then exits immediately.
2816    
2817    45. A need has arisen to pass over additional data with calls to pcre_exec() in
2818    order to support additional features. One way would have been to define
2819    pcre_exec2() (for example) with extra arguments, but this would not have been
2820    extensible, and would also have required all calls to the original function to
2821    be mapped to the new one. Instead, I have chosen to extend the mechanism that
2822    is used for passing in "extra" data from pcre_study().
2823    
2824    The pcre_extra structure is now exposed and defined in pcre.h. It currently
2825    contains the following fields:
2826    
2827      flags         a bitmap indicating which of the following fields are set
2828      study_data    opaque data from pcre_study()
2829      match_limit   a way of specifying a limit on match() calls for a specific
2830                      call to pcre_exec()
2831      callout_data  data for callouts (see 49 below)
2832    
2833    The flag bits are also defined in pcre.h, and are
2834    
2835      PCRE_EXTRA_STUDY_DATA
2836      PCRE_EXTRA_MATCH_LIMIT
2837      PCRE_EXTRA_CALLOUT_DATA
2838    
2839    The pcre_study() function now returns one of these new pcre_extra blocks, with
2840    the actual study data pointed to by the study_data field, and the
2841    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2842    before. That is, this change is entirely upwards-compatible and requires no
2843    change to existing code.
2844    
2845    If you want to pass in additional data to pcre_exec(), you can either place it
2846    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2847    block.
2848    
2849    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2850    data string contains the escape sequence \M, pcretest calls pcre_exec() several
2851    times with different match limits, until it finds the minimum value needed for
2852    pcre_exec() to complete. The value is then output. This can be instructive; for
2853    most simple matches the number is quite small, but for pathological cases it
2854    gets very large very quickly.
2855    
2856    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2857    returns the size of the data block pointed to by the study_data field in a
2858    pcre_extra block, that is, the value that was passed as the argument to
2859    pcre_malloc() when PCRE was getting memory in which to place the information
2860    created by pcre_study(). The fourth argument should point to a size_t variable.
2861    pcretest has been extended so that this information is shown after a successful
2862    pcre_study() call when information about the compiled regex is being displayed.
2863    
2864    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
2865    because what follows is always an absolute path. (Later: it turns out that this
2866    is more than cosmetic for MinGW, because it doesn't like empty path
2867    components.)
2868    
2869    49. Some changes have been made to the callout feature (see 28 above):
2870    
2871    (i)  A callout function now has three choices for what it returns:
2872    
2873           0  =>  success, carry on matching
2874         > 0  =>  failure at this point, but backtrack if possible
2875         < 0  =>  serious error, return this value from pcre_exec()
2876    
2877         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
2878         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
2879         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
2880         use by callout functions. It will never be used by PCRE itself.
2881    
2882    (ii) The pcre_extra structure (see 45 above) has a void * field called
2883         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
2884         pcre_callout_block structure has a field of the same name. The contents of
2885         the field passed in the pcre_extra structure are passed to the callout
2886         function in the corresponding field in the callout block. This makes it
2887         easier to use the same callout-containing regex from multiple threads. For
2888         testing, the pcretest program has a new data escape
2889    
2890           \C*n        pass the number n (may be negative) as callout_data
2891    
2892         If the callout function in pcretest receives a non-zero value as
2893         callout_data, it returns that value.
2894    
2895    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
2896    there were some redundant $(CFLAGS) in commands that are now specified as
2897    $(LINK), which already includes $(CFLAGS).
2898    
2899    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
2900    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
2901    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
2902    one-byte characters throughout. Note that case-insensitive matching applies
2903    only to characters whose values are less than 256. PCRE doesn't support the
2904    notion of cases for higher-valued characters.
2905    
2906    (i)   A character class whose characters are all within 0-255 is handled as
2907          a bit map, and the map is inverted for negative classes. Previously, a
2908          character > 255 always failed to match such a class; however it should
2909          match if the class was a negative one (e.g. [^ab]). This has been fixed.
2910    
2911    (ii)  A negated character class with a single character < 255 is coded as
2912          "not this character" (OP_NOT). This wasn't working properly when the test
2913          character was multibyte, either singly or repeated.
2914    
2915    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
2916          mode, for example: \x{100}{2,3}.
2917    
2918    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
2919          singly or repeated) now correctly test multibyte characters. However,
2920          PCRE doesn't recognize any characters with values greater than 255 as
2921          digits, spaces, or word characters. Such characters always match \D, \S,
2922          and \W, and never match \d, \s, or \w.
2923    
2924    (v)   Classes may now contain characters and character ranges with values
2925          greater than 255. For example: [ab\x{100}-\x{400}].
2926    
2927    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
2928          PCRE in UTF-8 mode.
2929    
2930    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
2931    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
2932    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
2933    value.)
2934    
2935    53. The single man page has become too large. I have therefore split it up into
2936    a number of separate man pages. These also give rise to individual HTML pages;
2937    these are now put in a separate directory, and there is an index.html page that
2938    lists them all. Some hyperlinking between the pages has been installed.
2939    
2940    54. Added convenience functions for handling named capturing parentheses.
2941    
2942    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
2943    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
2944    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
2945    are faulted.
2946    
2947    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
2948    calling configure. These values are used when compiling the dftables.c program
2949    which is run to generate the source of the default character tables. They
2950    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
2951    you will need to set these values.
2952    
2953    57. Updated the building process for Windows DLL, as provided by Fred Cox.
2954    
2955    
2956    Version 3.9 02-Jan-02
2957  ---------------------  ---------------------
2958    
2959  1. A bit of extraneous text had somehow crept into the pcregrep documentation.  1. A bit of extraneous text had somehow crept into the pcregrep documentation.

Legend:
Removed from v.61  
changed lines
  Added in v.474

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12