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

Diff of /code/trunk/ChangeLog

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

revision 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 412 by ph10, Sat Apr 11 10:34:37 2009 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 7.9 11-Apr-09
5    ---------------------
6    
7    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
8        (pcretest), all targets were linked against these libraries. This included
9        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
10        libraries. This caused unwanted dependencies to be created. This problem
11        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
12        pcretest is linked with readline.
13    
14    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
15        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
16        moved outside it again, because FALSE and TRUE are already defined in AIX,
17        but BOOL is not.
18    
19    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
20        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
21    
22    4.  The pcregrep documentation said spaces were inserted as well as colons (or
23        hyphens) following file names and line numbers when outputting matching
24        lines. This is not true; no spaces are inserted. I have also clarified the
25        wording for the --colour (or --color) option.
26    
27    5.  In pcregrep, when --colour was used with -o, the list of matching strings
28        was not coloured; this is different to GNU grep, so I have changed it to be
29        the same.
30    
31    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
32        each matching line was coloured. Now it goes on to look for further matches
33        of any of the test patterns, which is the same behaviour as GNU grep.
34    
35    7.  A pattern that could match an empty string could cause pcregrep to loop; it
36        doesn't make sense to accept an empty string match in pcregrep, so I have
37        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
38        seems to be how GNU grep behaves.
39    
40    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
41        start or after a newline", because the conditional assertion was not being
42        correctly handled. The rule now is that both the assertion and what follows
43        in the first alternative must satisfy the test.
44    
45    9.  If auto-callout was enabled in a pattern with a conditional group whose
46        condition was an assertion, PCRE could crash during matching, both with
47        pcre_exec() and pcre_dfa_exec().
48    
49    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
50        used for matching.
51    
52    11. Unicode property support in character classes was not working for
53        characters (bytes) greater than 127 when not in UTF-8 mode.
54    
55    12. Added the -M command line option to pcretest.
56    
57    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
58    
59    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
60    
61    16. Added comments and documentation about mis-use of no_arg in the C++
62        wrapper.
63    
64    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
65        from Martin Jerabek that uses macro names for all relevant character and
66        string constants.
67    
68    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
69        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
70        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
71        these, but not everybody uses configure.
72    
73    19. A conditional group that had only one branch was not being correctly
74        recognized as an item that could match an empty string. This meant that an
75        enclosing group might also not be so recognized, causing infinite looping
76        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
77        with the subject "ab", where knowledge that the repeated group can match
78        nothing is needed in order to break the loop.
79    
80    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
81        exec(), but without supplying a callout function, matching went wrong.
82    
83    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
84        leak if the size of the offset vector was greater than 30. When the vector
85        is smaller, the saved offsets during recursion go onto a local stack
86        vector, but for larger vectors malloc() is used. It was failing to free
87        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
88        error, in fact).
89    
90    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
91        heapframe that is used only when UTF-8 support is enabled. This caused no
92        problem, but was untidy.
93    
94    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
95        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
96        included within another project.
97    
98    24. Steven Van Ingelgem's patches to add more options to the CMake support,
99        slightly modified by me:
100    
101          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
102              not building pcregrep.
103    
104          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
105              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
106    
107    25. Forward references, both numeric and by name, in patterns that made use of
108        duplicate group numbers, could behave incorrectly or give incorrect errors,
109        because when scanning forward to find the reference group, PCRE was not
110        taking into account the duplicate group numbers. A pattern such as
111        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
112    
113    26. Changed a few more instances of "const unsigned char *" to USPTR, making
114        the feature of a custom pointer more persuasive (as requested by a user).
115    
116    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
117        pcretest.c, inside #ifndefs, because it seems they are sometimes already
118        pre-defined.
119    
120    28. Added support for (*UTF8) at the start of a pattern.
121    
122    
123    
124    Version 7.8 05-Sep-08
125    ---------------------
126    
127    1.  Replaced UCP searching code with optimized version as implemented for Ad
128        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
129        stage table and inline lookup instead of a function, giving speed ups of 2
130        to 5 times on some simple patterns that I tested. Permission was given to
131        distribute the MultiStage2.py script that generates the tables (it's not in
132        the tarball, but is in the Subversion repository).
133    
134    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
135        scripts.
136    
137    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
138        a group with a zero qualifier. The result of the study could be incorrect,
139        or the function might crash, depending on the pattern.
140    
141    4.  Caseless matching was not working for non-ASCII characters in back
142        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
143        It now works when Unicode Property Support is available.
144    
145    5.  In pcretest, an escape such as \x{de} in the data was always generating
146        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
147        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
148        truncation.
149    
150    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
151    
152    7.  Added two (int) casts to pcregrep when printing the difference of two
153        pointers, in case they are 64-bit values.
154    
155    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
156        test 2 if it fails.
157    
158    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
159        and a #define of that name to empty if it is not externally set. This is to
160        allow users of MSVC to set it if necessary.
161    
162    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
163        the convenience functions in the pcre_get.c source file.
164    
165    11. An option change at the start of a pattern that had top-level alternatives
166        could cause overwriting and/or a crash. This command provoked a crash in
167        some environments:
168    
169          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
170    
171        This potential security problem was recorded as CVE-2008-2371.
172    
173    12. For a pattern where the match had to start at the beginning or immediately
174        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
175        pcre_dfa_exec() could read past the end of the passed subject if there was
176        no match. To help with detecting such bugs (e.g. with valgrind), I modified
177        pcretest so that it places the subject at the end of its malloc-ed buffer.
178    
179    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
180        exec() might read past the end of the data buffer in UTF-8 mode.
181    
182    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
183        the data contained the byte 0x85 as part of a UTF-8 character within its
184        first line. This applied both to normal and DFA matching.
185    
186    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
187        /^[^d]*?$/8 failed to match "abc".
188    
189    16. Added a missing copyright notice to pcrecpp_internal.h.
190    
191    17. Make it more clear in the documentation that values returned from
192        pcre_exec() in ovector are byte offsets, not character counts.
193    
194    18. Tidied a few places to stop certain compilers from issuing warnings.
195    
196    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
197        supplied by Stefan Weber. I made a further small update for 7.8 because
198        there is a change of source arrangements: the pcre_searchfuncs.c module is
199        replaced by pcre_ucd.c.
200    
201    
202    Version 7.7 07-May-08
203    ---------------------
204    
205    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
206        a string to a long long, pretend we don't even have a long long." This is
207        done by checking for the strtoq, strtoll, and _strtoi64 functions.
208    
209    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
210        pre-7.6 versions, which defined a global no_arg variable instead of putting
211        it in the RE class. (See also #8 below.)
212    
213    3.  Remove a line of dead code, identified by coverity and reported by Nuno
214        Lopes.
215    
216    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
217    
218        (1) The include/exclude patterns were being applied to the whole pathnames
219            of files, instead of just to the final components.
220    
221        (2) If there was more than one level of directory, the subdirectories were
222            skipped unless they satisfied the include/exclude conditions. This is
223            inconsistent with GNU grep (and could even be seen as contrary to the
224            pcregrep specification - which I improved to make it absolutely clear).
225            The action now is always to scan all levels of directory, and just
226            apply the include/exclude patterns to regular files.
227    
228    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
229        --exclude_dir in the tests to avoid scanning .svn directories.
230    
231    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
232        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
233        doesn't support NULs in patterns.
234    
235    7.  Added some missing "const"s to declarations of static tables in
236        pcre_compile.c and pcre_dfa_exec.c.
237    
238    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
239        caused by fix #2  above. (Subsequently also a second patch to fix the
240        first patch. And a third patch - this was a messy problem.)
241    
242    9.  Applied Craig's patch to remove the use of push_back().
243    
244    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
245        matching function regexec().
246    
247    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
248        which, however, unlike Perl's \g{...}, are subroutine calls, not back
249        references. PCRE supports relative numbers with this syntax (I don't think
250        Oniguruma does).
251    
252    12. Previously, a group with a zero repeat such as (...){0} was completely
253        omitted from the compiled regex. However, this means that if the group
254        was called as a subroutine from elsewhere in the pattern, things went wrong
255        (an internal error was given). Such groups are now left in the compiled
256        pattern, with a new opcode that causes them to be skipped at execution
257        time.
258    
259    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
260        to the way PCRE behaves:
261    
262        (a) A lone ] character is dis-allowed (Perl treats it as data).
263    
264        (b) A back reference to an unmatched subpattern matches an empty string
265            (Perl fails the current match path).
266    
267        (c) A data ] in a character class must be notated as \] because if the
268            first data character in a class is ], it defines an empty class. (In
269            Perl it is not possible to have an empty class.) The empty class []
270            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
271            The negative empty class [^] matches any one character, independently
272            of the DOTALL setting.
273    
274    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
275        non-existent subpattern following a character class starting with ']' and
276        containing () gave an internal compiling error instead of "reference to
277        non-existent subpattern". Fortunately, when the pattern did exist, the
278        compiled code was correct. (When scanning forwards to check for the
279        existencd of the subpattern, it was treating the data ']' as terminating
280        the class, so got the count wrong. When actually compiling, the reference
281        was subsequently set up correctly.)
282    
283    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
284        it was being rejected as not supported by pcre_dfa_exec(), even though
285        other assertions are supported. I have made pcre_dfa_exec() support
286        (*FAIL).
287    
288    16. The implementation of 13c above involved the invention of a new opcode,
289        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
290        cannot be changed at match time, I realized I could make a small
291        improvement to matching performance by compiling OP_ALLANY instead of
292        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
293        on the OP_ANY path.
294    
295    17. Compiling pcretest on Windows with readline support failed without the
296        following two fixes: (1) Make the unistd.h include conditional on
297        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
298    
299    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
300        ncurses library to be included for pcretest when ReadLine support is
301        requested, but also to allow for it to be overridden. This patch came from
302        Daniel Bergström.
303    
304    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
305        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
306        any errors with the current Unicode tables. Thanks to Peter Kankowski for
307        spotting this.
308    
309    
310    Version 7.6 28-Jan-08
311    ---------------------
312    
313    1.  A character class containing a very large number of characters with
314        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
315        overflow.
316    
317    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
318        HAVE_LONG_LONG is not defined.
319    
320    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
321        bring them up to date and include new features. This patch includes:
322    
323        - Fixed PH's badly added libz and libbz2 support.
324        - Fixed a problem with static linking.
325        - Added pcredemo. [But later removed - see 7 below.]
326        - Fixed dftables problem and added an option.
327        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
328            HAVE_LONG_LONG.
329        - Added readline support for pcretest.
330        - Added an listing of the option settings after cmake has run.
331    
332    4.  A user submitted a patch to Makefile that makes it easy to create
333        "pcre.dll" under mingw when using Configure/Make. I added stuff to
334        Makefile.am that cause it to include this special target, without
335        affecting anything else. Note that the same mingw target plus all
336        the other distribution libraries and programs are now supported
337        when configuring with CMake (see 6 below) instead of with
338        Configure/Make.
339    
340    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
341        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
342        exported in the Windows port". It has not yet been confirmed that the patch
343        solves the problem, but it does no harm.
344    
345    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
346        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
347        with CMake, and also correct the comment about stack recursion.
348    
349    7.  Remove the automatic building of pcredemo from the ./configure system and
350        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
351        of a program that users should build themselves after PCRE is installed, so
352        building it automatically is not really right. What is more, it gave
353        trouble in some build environments.
354    
355    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
356    
357    
358    Version 7.5 10-Jan-08
359    ---------------------
360    
361    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
362        values in parens when parsing an RE using the C++ wrapper."
363    
364    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
365        Characters greater than 255 were excluded from the class instead of being
366        included.
367    
368    3.  The same bug as (2) above applied to negated POSIX classes such as
369        [:^space:].
370    
371    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
372        defined or documented. It seems to have been a typo for PCRE_STATIC, so
373        I have changed it.
374    
375    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
376        first named subpattern) and a construct such as (?&a) would reference the
377        first named subpattern whose name started with "a" (in other words, the
378        length check was missing). Both these problems are fixed. "Subpattern name
379        expected" is now given for (?&) (a zero-length name), and this patch also
380        makes it give the same error for \k'' (previously it complained that that
381        was a reference to a non-existent subpattern).
382    
383    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
384        this is right because (?- can be followed by option settings as well as by
385        digits. I have, however, made the messages clearer.
386    
387    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
388        than the number used in the conditional) now cause a compile-time error.
389        This is actually not compatible with Perl, which accepts such patterns, but
390        treats the conditional as always being FALSE (as PCRE used to), but it
391        seems to me that giving a diagnostic is better.
392    
393    8.  Change "alphameric" to the more common word "alphanumeric" in comments
394        and messages.
395    
396    9.  Fix two occurrences of "backslash" in comments that should have been
397        "backspace".
398    
399    10. Remove two redundant lines of code that can never be obeyed (their function
400        was moved elsewhere).
401    
402    11. The program that makes PCRE's Unicode character property table had a bug
403        which caused it to generate incorrect table entries for sequences of
404        characters that have the same character type, but are in different scripts.
405        It amalgamated them into a single range, with the script of the first of
406        them. In other words, some characters were in the wrong script. There were
407        thirteen such cases, affecting characters in the following ranges:
408    
409          U+002b0 - U+002c1
410          U+0060c - U+0060d
411          U+0061e - U+00612
412          U+0064b - U+0065e
413          U+0074d - U+0076d
414          U+01800 - U+01805
415          U+01d00 - U+01d77
416          U+01d9b - U+01dbf
417          U+0200b - U+0200f
418          U+030fc - U+030fe
419          U+03260 - U+0327f
420          U+0fb46 - U+0fbb1
421          U+10450 - U+1049d
422    
423    12. The -o option (show only the matching part of a line) for pcregrep was not
424        compatible with GNU grep in that, if there was more than one match in a
425        line, it showed only the first of them. It now behaves in the same way as
426        GNU grep.
427    
428    13. If the -o and -v options were combined for pcregrep, it printed a blank
429        line for every non-matching line. GNU grep prints nothing, and pcregrep now
430        does the same. The return code can be used to tell if there were any
431        non-matching lines.
432    
433    14. Added --file-offsets and --line-offsets to pcregrep.
434    
435    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
436        infinitely looping recursion. The bug was that positive lookaheads were not
437        being skipped when checking for a possible empty match (negative lookaheads
438        and both kinds of lookbehind were skipped).
439    
440    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
441        inclusion of <windows.h> to before rather than after the definition of
442        INVALID_FILE_ATTRIBUTES (patch from David Byron).
443    
444    17. Specifying a possessive quantifier with a specific limit for a Unicode
445        character property caused pcre_compile() to compile bad code, which led at
446        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
447        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
448        caused the error; without that there was no problem.
449    
450    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
451    
452    19. Added --enable-pcretest-libreadline.
453    
454    20. In pcrecpp.cc, the variable 'count' was incremented twice in
455        RE::GlobalReplace(). As a result, the number of replacements returned was
456        double what it should be. I removed one of the increments, but Craig sent a
457        later patch that removed the other one (the right fix) and added unit tests
458        that check the return values (which was not done before).
459    
460    21. Several CMake things:
461    
462        (1) Arranged that, when cmake is used on Unix, the libraries end up with
463            the names libpcre and libpcreposix, not just pcre and pcreposix.
464    
465        (2) The above change means that pcretest and pcregrep are now correctly
466            linked with the newly-built libraries, not previously installed ones.
467    
468        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
469    
470    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
471        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
472        UTF-8 newline character). The key issue is that the pattern starts .*;
473        this means that the match must be either at the beginning, or after a
474        newline. The bug was in the code for advancing after a failed match and
475        checking that the new position followed a newline. It was not taking
476        account of UTF-8 characters correctly.
477    
478    23. PCRE was behaving differently from Perl in the way it recognized POSIX
479        character classes. PCRE was not treating the sequence [:...:] as a
480        character class unless the ... were all letters. Perl, however, seems to
481        allow any characters between [: and :], though of course it rejects as
482        unknown any "names" that contain non-letters, because all the known class
483        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
484        for example, whereas PCRE did not - it did not recognize a POSIX character
485        class. This seemed a bit dangerous, so the code has been changed to be
486        closer to Perl. The behaviour is not identical to Perl, because PCRE will
487        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
488        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
489        Perl does, and where it didn't before.
490    
491    24. Rewrite so as to remove the single use of %n from pcregrep because in some
492        Windows environments %n is disabled by default.
493    
494    
495    Version 7.4 21-Sep-07
496    ---------------------
497    
498    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
499        means that a class such as [\s] counted as "explicit reference to CR or
500        LF". That isn't really right - the whole point of the change was to try to
501        help when there was an actual mention of one of the two characters. So now
502        the change happens only if \r or \n (or a literal CR or LF) character is
503        encountered.
504    
505    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
506        of both had grown to the point where there were only 3 bits left.
507        Fortunately, there was spare space in the data structure, and so I have
508        moved the internal flags into a new 16-bit field to free up more option
509        bits.
510    
511    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
512        but did not set the internal JCHANGED flag - either of these is enough to
513        control the way the "get" function works - but the PCRE_INFO_JCHANGED
514        facility is supposed to tell if (?J) was ever used, so now (?J) at the
515        start sets both bits.
516    
517    4.  Added options (at build time, compile time, exec time) to change \R from
518        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
519    
520    5.  doc/pcresyntax.html was missing from the distribution.
521    
522    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
523        compatibility, even though it is no longer used.
524    
525    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
526        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
527        windows.h file is present (where different names are used). [This was
528        reversed later after testing - see 16 below.]
529    
530    8.  Changed all #include <config.h> to #include "config.h". There were also
531        some further <pcre.h> cases that I changed to "pcre.h".
532    
533    9.  When pcregrep was used with the --colour option, it missed the line ending
534        sequence off the lines that it output.
535    
536    10. It was pointed out to me that arrays of string pointers cause lots of
537        relocations when a shared library is dynamically loaded. A technique of
538        using a single long string with a table of offsets can drastically reduce
539        these. I have refactored PCRE in four places to do this. The result is
540        dramatic:
541    
542          Originally:                          290
543          After changing UCP table:            187
544          After changing error message table:   43
545          After changing table of "verbs"       36
546          After changing table of Posix names   22
547    
548        Thanks to the folks working on Gregex for glib for this insight.
549    
550    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
551        unicode-properties was also set.
552    
553    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
554    
555    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
556        checked only for CRLF.
557    
558    14. Added casts to pcretest.c to avoid compiler warnings.
559    
560    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
561    
562    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
563        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
564        entirely. This removes changes made in 7 above.
565    
566    17. The CMake files have been updated, and there is now more information about
567        building with CMake in the NON-UNIX-USE document.
568    
569    
570    Version 7.3 28-Aug-07
571    ---------------------
572    
573     1. In the rejigging of the build system that eventually resulted in 7.1, the
574        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
575        brackets there is not right, since it causes compilers to look for an
576        installed pcre.h, not the version that is in the source that is being
577        compiled (which of course may be different). I have changed it back to:
578    
579          #include "pcre.h"
580    
581        I have a vague recollection that the change was concerned with compiling in
582        different directories, but in the new build system, that is taken care of
583        by the VPATH setting the Makefile.
584    
585     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
586        when the subject happened to end in the byte 0x85 (e.g. if the last
587        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
588        characters but of course it shouldn't be taken as a newline when it is part
589        of another character. The bug was that, for an unlimited repeat of . in
590        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
591        characters when looking for a newline.
592    
593     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
594    
595     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
596        in debug output.
597    
598     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
599        long printing in the pcrecpp unittest when running under MinGW.
600    
601     6. ESC_K was left out of the EBCDIC table.
602    
603     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
604        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
605        limit also applies to "virtual nesting" when a pattern is recursive, and in
606        this case 1000 isn't so big. I have been able to remove this limit at the
607        expense of backing off one optimization in certain circumstances. Normally,
608        when pcre_exec() would call its internal match() function recursively and
609        immediately return the result unconditionally, it uses a "tail recursion"
610        feature to save stack. However, when a subpattern that can match an empty
611        string has an unlimited repetition quantifier, it no longer makes this
612        optimization. That gives it a stack frame in which to save the data for
613        checking that an empty string has been matched. Previously this was taken
614        from the 1000-entry workspace that had been reserved. So now there is no
615        explicit limit, but more stack is used.
616    
617     8. Applied Daniel's patches to solve problems with the import/export magic
618        syntax that is required for Windows, and which was going wrong for the
619        pcreposix and pcrecpp parts of the library. These were overlooked when this
620        problem was solved for the main library.
621    
622     9. There were some crude static tests to avoid integer overflow when computing
623        the size of patterns that contain repeated groups with explicit upper
624        limits. As the maximum quantifier is 65535, the maximum group length was
625        set at 30,000 so that the product of these two numbers did not overflow a
626        32-bit integer. However, it turns out that people want to use groups that
627        are longer than 30,000 bytes (though not repeat them that many times).
628        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
629        made it possible to implement the integer overflow checks in a much more
630        dynamic way, which I have now done. The artificial limitation on group
631        length has been removed - we now have only the limit on the total length of
632        the compiled pattern, which depends on the LINK_SIZE setting.
633    
634    10. Fixed a bug in the documentation for get/copy named substring when
635        duplicate names are permitted. If none of the named substrings are set, the
636        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
637        empty string.
638    
639    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
640        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
641        because the ] is interpreted as the first data character and the
642        terminating ] is not found. PCRE has been made compatible with Perl in this
643        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
644        cause memory overwriting.
645    
646    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
647        string has been matched (to stop an infinite loop). It was not recognizing
648        a conditional subpattern that could match an empty string if that
649        subpattern was within another subpattern. For example, it looped when
650        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
651        condition was not nested. This bug has been fixed.
652    
653    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
654        past the start of the subject in the presence of bytes with the top bit
655        set, for example "\x8aBCD".
656    
657    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
658        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
659    
660    14. Optimized (?!) to (*FAIL).
661    
662    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
663        This restricts code points to be within the range 0 to 0x10FFFF, excluding
664        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
665        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
666        does: it's just the validity check that is more restrictive.
667    
668    16. Inserted checks for integer overflows during escape sequence (backslash)
669        processing, and also fixed erroneous offset values for syntax errors during
670        backslash processing.
671    
672    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
673        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
674    
675    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
676        caused an overrun.
677    
678    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
679        something other than just ASCII characters) inside a group that had an
680        unlimited repeat caused a loop at compile time (while checking to see
681        whether the group could match an empty string).
682    
683    20. Debugging a pattern containing \p or \P could cause a crash. For example,
684        [\P{Any}] did so. (Error in the code for printing property names.)
685    
686    21. An orphan \E inside a character class could cause a crash.
687    
688    22. A repeated capturing bracket such as (A)? could cause a wild memory
689        reference during compilation.
690    
691    23. There are several functions in pcre_compile() that scan along a compiled
692        expression for various reasons (e.g. to see if it's fixed length for look
693        behind). There were bugs in these functions when a repeated \p or \P was
694        present in the pattern. These operators have additional parameters compared
695        with \d, etc, and these were not being taken into account when moving along
696        the compiled data. Specifically:
697    
698        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
699            length.
700    
701        (b) An item such as \pL+ within a repeated group could cause crashes or
702            loops.
703    
704        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
705            "reference to non-existent subpattern" error.
706    
707        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
708    
709    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
710        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
711    
712    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
713    
714    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
715        character were causing crashes (broken optimization).
716    
717    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
718        \p or \P) caused a compile-time loop.
719    
720    28. More problems have arisen in unanchored patterns when CRLF is a valid line
721        break. For example, the unstudied pattern [\r\n]A does not match the string
722        "\r\nA" because change 7.0/46 below moves the current point on by two
723        characters after failing to match at the start. However, the pattern \nA
724        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
725        the same is true. There doesn't seem any very clean way out of this, but
726        what I have chosen to do makes the common cases work: PCRE now takes note
727        of whether there can be an explicit match for \r or \n anywhere in the
728        pattern, and if so, 7.0/46 no longer applies. As part of this change,
729        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
730        pattern has explicit CR or LF references.
731    
732    29. Added (*CR) etc for changing newline setting at start of pattern.
733    
734    
735    Version 7.2 19-Jun-07
736    ---------------------
737    
738     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
739        which is apparently normally available under Windows.
740    
741     2. Re-jig the pcregrep tests with different newline settings in an attempt
742        to make them independent of the local environment's newline setting.
743    
744     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
745    
746     4. Some of the "internals" tests were previously cut out when the link size
747        was not 2, because the output contained actual offsets. The recent new
748        "Z" feature of pcretest means that these can be cut out, making the tests
749        usable with all link sizes.
750    
751     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
752        stack recursion. This gives a massive performance boost under BSD, but just
753        a small improvement under Linux. However, it saves one field in the frame
754        in all cases.
755    
756     6. Added more features from the forthcoming Perl 5.10:
757    
758        (a) (?-n) (where n is a string of digits) is a relative subroutine or
759            recursion call. It refers to the nth most recently opened parentheses.
760    
761        (b) (?+n) is also a relative subroutine call; it refers to the nth next
762            to be opened parentheses.
763    
764        (c) Conditions that refer to capturing parentheses can be specified
765            relatively, for example, (?(-2)... or (?(+3)...
766    
767        (d) \K resets the start of the current match so that everything before
768            is not part of it.
769    
770        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
771    
772        (f) \g{name} is another synonym - part of Perl 5.10's unification of
773            reference syntax.
774    
775        (g) (?| introduces a group in which the numbering of parentheses in each
776            alternative starts with the same number.
777    
778        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
779    
780     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
781        PCRE_INFO_JCHANGED.
782    
783     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
784        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
785        for detecting groups that can match an empty string.
786    
787     9. A pattern with a very large number of alternatives (more than several
788        hundred) was running out of internal workspace during the pre-compile
789        phase, where pcre_compile() figures out how much memory will be needed. A
790        bit of new cunning has reduced the workspace needed for groups with
791        alternatives. The 1000-alternative test pattern now uses 12 bytes of
792        workspace instead of running out of the 4096 that are available.
793    
794    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
795    
796    11. Applied patch from Google to remove an optimization that didn't quite work.
797        The report of the bug said:
798    
799          pcrecpp::RE("a*").FullMatch("aaa") matches, while
800          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
801          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
802    
803    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
804        it matched the wrong number of bytes.
805    
806    
807    Version 7.1 24-Apr-07
808    ---------------------
809    
810     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
811        that is more "standard", making use of automake and other Autotools. There
812        is some re-arrangement of the files and adjustment of comments consequent
813        on this.
814    
815     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
816        for recursive directory scanning broke on some systems because the files
817        are not scanned in any specific order and on different systems the order
818        was different. A call to "sort" has been inserted into RunGrepTest for the
819        approprate test as a short-term fix. In the longer term there may be an
820        alternative.
821    
822     3. I had an email from Eric Raymond about problems translating some of PCRE's
823        man pages to HTML (despite the fact that I distribute HTML pages, some
824        people do their own conversions for various reasons). The problems
825        concerned the use of low-level troff macros .br and .in. I have therefore
826        removed all such uses from the man pages (some were redundant, some could
827        be replaced by .nf/.fi pairs). The 132html script that I use to generate
828        HTML has been updated to handle .nf/.fi and to complain if it encounters
829        .br or .in.
830    
831     4. Updated comments in configure.ac that get placed in config.h.in and also
832        arranged for config.h to be included in the distribution, with the name
833        config.h.generic, for the benefit of those who have to compile without
834        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
835    
836     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
837        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
838        makevp.bat for the current PCRE, using the additional files
839        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
840    
841     6. A Windows user reported a minor discrepancy with test 2, which turned out
842        to be caused by a trailing space on an input line that had got lost in his
843        copy. The trailing space was an accident, so I've just removed it.
844    
845     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
846        that is needed.
847    
848     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
849        as "const" (a) because they are and (b) because it helps the PHP
850        maintainers who have recently made a script to detect big data structures
851        in the php code that should be moved to the .rodata section. I remembered
852        to update Builducptable as well, so it won't revert if ucptable.h is ever
853        re-created.
854    
855     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
856        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
857        order to be able to cut out the UTF-8 tables in the latter when UTF-8
858        support is not required. This saves 1.5-2K of code, which is important in
859        some applications.
860    
861        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
862        so as not to refer to the tables, even though these functions will never be
863        called when UTF-8 support is disabled. Otherwise there are problems with a
864        shared library.
865    
866    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
867    
868        (a) It was defining its arguments as char * instead of void *.
869    
870        (b) It was assuming that all moves were upwards in memory; this was true
871            a long time ago when I wrote it, but is no longer the case.
872    
873        The emulated memove() is provided for those environments that have neither
874        memmove() nor bcopy(). I didn't think anyone used it these days, but that
875        is clearly not the case, as these two bugs were recently reported.
876    
877    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
878        and Detrail to create the HTML documentation, the .txt form of the man
879        pages, and it removes trailing spaces from listed files. It also creates
880        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
881        case, it wraps all the #defines with #ifndefs. This script should be run
882        before "make dist".
883    
884    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
885        with Unicode property support.
886    
887        (a) For a maximizing quantifier, if the two different cases of the
888            character were of different lengths in their UTF-8 codings (there are
889            some cases like this - I found 11), and the matching function had to
890            back up over a mixture of the two cases, it incorrectly assumed they
891            were both the same length.
892    
893        (b) When PCRE was configured to use the heap rather than the stack for
894            recursion during matching, it was not correctly preserving the data for
895            the other case of a UTF-8 character when checking ahead for a match
896            while processing a minimizing repeat. If the check also involved
897            matching a wide character, but failed, corruption could cause an
898            erroneous result when trying to check for a repeat of the original
899            character.
900    
901    13. Some tidying changes to the testing mechanism:
902    
903        (a) The RunTest script now detects the internal link size and whether there
904            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
905            values substituted by "configure". (The RunGrepTest script already did
906            this for UTF-8.) The configure.ac script no longer substitutes the
907            relevant variables.
908    
909        (b) The debugging options /B and /D in pcretest show the compiled bytecode
910            with length and offset values. This means that the output is different
911            for different internal link sizes. Test 2 is skipped for link sizes
912            other than 2 because of this, bypassing the problem. Unfortunately,
913            there was also a test in test 3 (the locale tests) that used /B and
914            failed for link sizes other than 2. Rather than cut the whole test out,
915            I have added a new /Z option to pcretest that replaces the length and
916            offset values with spaces. This is now used to make test 3 independent
917            of link size. (Test 2 will be tidied up later.)
918    
919    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
920        segmentation fault instead of returning the appropriate error message.
921    
922    15. In multiline mode when the newline sequence was set to "any", the pattern
923        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
924        This doesn't seem right; it now treats the CRLF combination as the line
925        ending, and so does not match in that case. It's only a pattern such as ^$
926        that would hit this one: something like ^ABC$ would have failed after \r
927        and then tried again after \r\n.
928    
929    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
930        in an attempt to make files that differ only in their line terminators
931        compare equal. This works on Linux.
932    
933    17. Under certain error circumstances pcregrep might try to free random memory
934        as it exited. This is now fixed, thanks to valgrind.
935    
936    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
937        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
938        was because its rules for how to advance for /g after matching an empty
939        string at the end of a line did not allow for this case. They now check for
940        it specially.
941    
942    20. pcretest is supposed to handle patterns and data of any length, by
943        extending its buffers when necessary. It was getting this wrong when the
944        buffer for a data line had to be extended.
945    
946    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
947        CRLF as a newline sequence.
948    
949    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
950        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
951        I have nevertheless tidied it up.
952    
953    23. Added some casts to kill warnings from HP-UX ia64 compiler.
954    
955    24. Added a man page for pcre-config.
956    
957    
958    Version 7.0 19-Dec-06
959    ---------------------
960    
961     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
962        moving to gcc 4.1.1.
963    
964     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
965        sys/time.h, which is documented as needed for this function. It doesn't
966        seem to matter on Linux, but it showed up on some releases of OS X.
967    
968     3. It seems that there are systems where bytes whose values are greater than
969        127 match isprint() in the "C" locale. The "C" locale should be the
970        default when a C program starts up. In most systems, only ASCII printing
971        characters match isprint(). This difference caused the output from pcretest
972        to vary, making some of the tests fail. I have changed pcretest so that:
973    
974        (a) When it is outputting text in the compiled version of a pattern, bytes
975            other than 32-126 are always shown as hex escapes.
976    
977        (b) When it is outputting text that is a matched part of a subject string,
978            it does the same, unless a different locale has been set for the match
979            (using the /L modifier). In this case, it uses isprint() to decide.
980    
981     4. Fixed a major bug that caused incorrect computation of the amount of memory
982        required for a compiled pattern when options that changed within the
983        pattern affected the logic of the preliminary scan that determines the
984        length. The relevant options are -x, and -i in UTF-8 mode. The result was
985        that the computed length was too small. The symptoms of this bug were
986        either the PCRE error "internal error: code overflow" from pcre_compile(),
987        or a glibc crash with a message such as "pcretest: free(): invalid next
988        size (fast)". Examples of patterns that provoked this bug (shown in
989        pcretest format) are:
990    
991          /(?-x: )/x
992          /(?x)(?-x: \s*#\s*)/
993          /((?i)[\x{c0}])/8
994          /(?i:[\x{c0}])/8
995    
996        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
997        is now done differently.
998    
999     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1000        wrapper classes; (b) implement a new function in the C++ scanner that is
1001        more efficient than the old way of doing things because it avoids levels of
1002        recursion in the regex matching; (c) add a paragraph to the documentation
1003        for the FullMatch() function.
1004    
1005     6. The escape sequence \n was being treated as whatever was defined as
1006        "newline". Not only was this contrary to the documentation, which states
1007        that \n is character 10 (hex 0A), but it also went horribly wrong when
1008        "newline" was defined as CRLF. This has been fixed.
1009    
1010     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1011        was being set to -1 for the "end of line" case (supposedly a value that no
1012        character can have). Though this value is never used (the check for end of
1013        line is "zero bytes in current character"), it caused compiler complaints.
1014        I've changed it to 0xffffffff.
1015    
1016     8. In pcre_version.c, the version string was being built by a sequence of
1017        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1018        string (as it is for production releases) called a macro with an empty
1019        argument. The C standard says the result of this is undefined. The gcc
1020        compiler treats it as an empty string (which was what was wanted) but it is
1021        reported that Visual C gives an error. The source has been hacked around to
1022        avoid this problem.
1023    
1024     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1025        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1026        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1027        of them did).
1028    
1029    10. Originally, pcretest opened its input and output without "b"; then I was
1030        told that "b" was needed in some environments, so it was added for release
1031        5.0 to both the input and output. (It makes no difference on Unix-like
1032        systems.) Later I was told that it is wrong for the input on Windows. I've
1033        now abstracted the modes into two macros, to make it easier to fiddle with
1034        them, and removed "b" from the input mode under Windows.
1035    
1036    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1037    
1038    12. Added -help and --help to pcretest as an official way of being reminded
1039        of the options.
1040    
1041    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1042        and pcrecpp.cc because they annoy compilers at high warning levels.
1043    
1044    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1045    
1046    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1047        scripts are not C programs :-) and which was not noticed because it works
1048        on Linux.
1049    
1050    16. pcretest is supposed to handle any length of pattern and data line (as one
1051        line or as a continued sequence of lines) by extending its input buffer if
1052        necessary. This feature was broken for very long pattern lines, leading to
1053        a string of junk being passed to pcre_compile() if the pattern was longer
1054        than about 50K.
1055    
1056    17. I have done a major re-factoring of the way pcre_compile() computes the
1057        amount of memory needed for a compiled pattern. Previously, there was code
1058        that made a preliminary scan of the pattern in order to do this. That was
1059        OK when PCRE was new, but as the facilities have expanded, it has become
1060        harder and harder to keep it in step with the real compile phase, and there
1061        have been a number of bugs (see for example, 4 above). I have now found a
1062        cunning way of running the real compile function in a "fake" mode that
1063        enables it to compute how much memory it would need, while actually only
1064        ever using a few hundred bytes of working memory and without too many
1065        tests of the mode. This should make future maintenance and development
1066        easier. A side effect of this work is that the limit of 200 on the nesting
1067        depth of parentheses has been removed (though this was never a serious
1068        limitation, I suspect). However, there is a downside: pcre_compile() now
1069        runs more slowly than before (30% or more, depending on the pattern). I
1070        hope this isn't a big issue. There is no effect on runtime performance.
1071    
1072    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1073        newline (only possible for the last line of a file) and it was a
1074        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1075    
1076    19. Added additional timing features to pcretest. (1) The -tm option now times
1077        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1078        separate command line item, by a number that specifies the number of
1079        repeats to use when timing. The default is 50000; this gives better
1080        precision, but takes uncomfortably long for very large patterns.
1081    
1082    20. Extended pcre_study() to be more clever in cases where a branch of a
1083        subpattern has no definite first character. For example, (a*|b*)[cd] would
1084        previously give no result from pcre_study(). Now it recognizes that the
1085        first character must be a, b, c, or d.
1086    
1087    21. There was an incorrect error "recursive call could loop indefinitely" if
1088        a subpattern (or the entire pattern) that was being tested for matching an
1089        empty string contained only one non-empty item after a nested subpattern.
1090        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1091        incorrectly, because the \d was being skipped in the check.
1092    
1093    22. The pcretest program now has a new pattern option /B and a command line
1094        option -b, which is equivalent to adding /B to every pattern. This causes
1095        it to show the compiled bytecode, without the additional information that
1096        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1097        is the same as /B/I).
1098    
1099    23. A new optimization is now able automatically to treat some sequences such
1100        as a*b as a*+b. More specifically, if something simple (such as a character
1101        or a simple class like \d) has an unlimited quantifier, and is followed by
1102        something that cannot possibly match the quantified thing, the quantifier
1103        is automatically "possessified".
1104    
1105    24. A recursive reference to a subpattern whose number was greater than 39
1106        went wrong under certain circumstances in UTF-8 mode. This bug could also
1107        have affected the operation of pcre_study().
1108    
1109    25. Realized that a little bit of performance could be had by replacing
1110        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1111    
1112    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1113    
1114    27. Possessive quantifiers such as a++ were previously implemented by turning
1115        them into atomic groups such as ($>a+). Now they have their own opcodes,
1116        which improves performance. This includes the automatically created ones
1117        from 23 above.
1118    
1119    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1120        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1121        the first matched character to be a colon. This applied both to named and
1122        numbered groups.
1123    
1124    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1125    
1126    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1127        building PCRE on Windows easier, so I have included it in the distribution.
1128    
1129    31. There is now a check in pcretest against a ridiculously large number being
1130        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1131        loop, the loop is abandoned.
1132    
1133    32. Forward references to subpatterns in conditions such as (?(2)...) where
1134        subpattern 2 is defined later cause pcre_compile() to search forwards in
1135        the pattern for the relevant set of parentheses. This search went wrong
1136        when there were unescaped parentheses in a character class, parentheses
1137        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1138    
1139    33. "Subroutine" calls and backreferences were previously restricted to
1140        referencing subpatterns earlier in the regex. This restriction has now
1141        been removed.
1142    
1143    34. Added a number of extra features that are going to be in Perl 5.10. On the
1144        whole, these are just syntactic alternatives for features that PCRE had
1145        previously implemented using the Python syntax or my own invention. The
1146        other formats are all retained for compatibility.
1147    
1148        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1149            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1150            also .NET compatible.
1151    
1152        (b) A recursion or subroutine call to a named group can now be defined as
1153            (?&name) as well as (?P>name).
1154    
1155        (c) A backreference to a named group can now be defined as \k<name> or
1156            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1157            5.10, are also .NET compatible.
1158    
1159        (d) A conditional reference to a named group can now use the syntax
1160            (?(<name>) or (?('name') as well as (?(name).
1161    
1162        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1163            groups (named and numbered) that are never evaluated inline, but can be
1164            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1165            is always false. There may be only one alternative in such a group.
1166    
1167        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1168            as the simple (?(R). The condition is true only if the most recent
1169            recursion is that of the given number or name. It does not search out
1170            through the entire recursion stack.
1171    
1172        (g) The escape \gN or \g{N} has been added, where N is a positive or
1173            negative number, specifying an absolute or relative reference.
1174    
1175    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1176        some "unreachable code" warnings.
1177    
1178    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1179        things, this adds five new scripts.
1180    
1181    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1182        There were also incompatibilities regarding the handling of \Q..\E inside
1183        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1184        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1185    
1186    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1187        matches an empty string, and forcibly breaks the loop. There were bugs in
1188        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1189        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1190        separate and independent bugs (that affected different cases) have been
1191        fixed.
1192    
1193    39. Refactored the code to abolish the use of different opcodes for small
1194        capturing bracket numbers. This is a tidy that I avoided doing when I
1195        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1196        The new approach is not only tidier, it makes it possible to reduce the
1197        memory needed to fix the previous bug (38).
1198    
1199    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1200        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1201        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1202        mode.
1203    
1204    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1205        report.
1206    
1207    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1208        copy construction and assignment in the C++ wrapper.
1209    
1210    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1211        couple of bugs that could have given wrong results in the "--newline=crlf"
1212        case.
1213    
1214    44. Added a number of casts and did some reorganization of signed/unsigned int
1215        variables following suggestions from Dair Grant. Also renamed the variable
1216        "this" as "item" because it is a C++ keyword.
1217    
1218    45. Arranged for dftables to add
1219    
1220          #include "pcre_internal.h"
1221    
1222        to pcre_chartables.c because without it, gcc 4.x may remove the array
1223        definition from the final binary if PCRE is built into a static library and
1224        dead code stripping is activated.
1225    
1226    46. For an unanchored pattern, if a match attempt fails at the start of a
1227        newline sequence, and the newline setting is CRLF or ANY, and the next two
1228        characters are CRLF, advance by two characters instead of one.
1229    
1230    
1231    Version 6.7 04-Jul-06
1232    ---------------------
1233    
1234     1. In order to handle tests when input lines are enormously long, pcretest has
1235        been re-factored so that it automatically extends its buffers when
1236        necessary. The code is crude, but this _is_ just a test program. The
1237        default size has been increased from 32K to 50K.
1238    
1239     2. The code in pcre_study() was using the value of the re argument before
1240        testing it for NULL. (Of course, in any sensible call of the function, it
1241        won't be NULL.)
1242    
1243     3. The memmove() emulation function in pcre_internal.h, which is used on
1244        systems that lack both memmove() and bcopy() - that is, hardly ever -
1245        was missing a "static" storage class specifier.
1246    
1247     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1248        containing an extended class (one that cannot be represented by a bitmap
1249        because it contains high-valued characters or Unicode property items, e.g.
1250        [\pZ]). Almost always one would set UTF-8 mode when processing such a
1251        pattern, but PCRE should not loop if you do not (it no longer does).
1252        [Detail: two cases were found: (a) a repeated subpattern containing an
1253        extended class; (b) a recursive reference to a subpattern that followed a
1254        previous extended class. It wasn't skipping over the extended class
1255        correctly when UTF-8 mode was not set.]
1256    
1257     5. A negated single-character class was not being recognized as fixed-length
1258        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1259        compile error "lookbehind assertion is not fixed length".
1260    
1261     6. The RunPerlTest auxiliary script was showing an unexpected difference
1262        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1263        write a Perl script that can interpret lines of an input file either as
1264        byte characters or as UTF-8, which is what "perltest" was being required to
1265        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1266        can't do is switch easily at run time between having the "use utf8;" pragma
1267        or not. In the end, I fudged it by using the RunPerlTest script to insert
1268        "use utf8;" explicitly for the UTF-8 tests.
1269    
1270     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1271        the end of the subject string, contrary to the documentation and to what
1272        Perl does. This was true of both matching functions. Now it matches only at
1273        the start of the subject and immediately after *internal* newlines.
1274    
1275     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1276        a pointer to an int instead of a pointer to an unsigned long int. This
1277        caused problems on 64-bit systems.
1278    
1279     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1280        instance of the 'standard' template library not being so standard".
1281    
1282    10. There was no check on the number of named subpatterns nor the maximum
1283        length of a subpattern name. The product of these values is used to compute
1284        the size of the memory block for a compiled pattern. By supplying a very
1285        long subpattern name and a large number of named subpatterns, the size
1286        computation could be caused to overflow. This is now prevented by limiting
1287        the length of names to 32 characters, and the number of named subpatterns
1288        to 10,000.
1289    
1290    11. Subpatterns that are repeated with specific counts have to be replicated in
1291        the compiled pattern. The size of memory for this was computed from the
1292        length of the subpattern and the repeat count. The latter is limited to
1293        65535, but there was no limit on the former, meaning that integer overflow
1294        could in principle occur. The compiled length of a repeated subpattern is
1295        now limited to 30,000 bytes in order to prevent this.
1296    
1297    12. Added the optional facility to have named substrings with the same name.
1298    
1299    13. Added the ability to use a named substring as a condition, using the
1300        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1301        are numbers (not recommended). Forward references are permitted.
1302    
1303    14. Added forward references in named backreferences (if you see what I mean).
1304    
1305    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1306        pattern could run off the end of the subject. For example, the pattern
1307        "(?s)(.{1,5})"8 did this with the subject "ab".
1308    
1309    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1310        PCRE_CASELESS was set when matching characters that were quantified with ?
1311        or *.
1312    
1313    17. A character class other than a single negated character that had a minimum
1314        but no maximum quantifier - for example [ab]{6,} - was not handled
1315        correctly by pce_dfa_exec(). It would match only one character.
1316    
1317    18. A valid (though odd) pattern that looked like a POSIX character
1318        class but used an invalid character after [ (for example [[,abc,]]) caused
1319        pcre_compile() to give the error "Failed: internal error: code overflow" or
1320        in some cases to crash with a glibc free() error. This could even happen if
1321        the pattern terminated after [[ but there just happened to be a sequence of
1322        letters, a binary zero, and a closing ] in the memory that followed.
1323    
1324    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1325        over the years. Originally (before any Unicode support), just the bottom 8
1326        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1327        output from "man perlunicode" includes this:
1328    
1329          The regular expression compiler produces polymorphic opcodes.  That
1330          is, the pattern adapts to the data and automatically switches to
1331          the Unicode character scheme when presented with Unicode data--or
1332          instead uses a traditional byte scheme when presented with byte
1333          data.
1334    
1335        Sadly, a wide octal escape does not cause a switch, and in a string with
1336        no other multibyte characters, these octal escapes are treated as before.
1337        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
1338        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1339        Unicode string.
1340    
1341        I have not perpetrated such confusion in PCRE. Up till now, it took just
1342        the bottom 8 bits, as in old Perl. I have now made octal escapes with
1343        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1344        translate to the appropriate multibyte character.
1345    
1346    29. Applied some refactoring to reduce the number of warnings from Microsoft
1347        and Borland compilers. This has included removing the fudge introduced
1348        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1349        a warning about an unused variable.
1350    
1351    21. PCRE has not included VT (character 0x0b) in the set of whitespace
1352        characters since release 4.0, because Perl (from release 5.004) does not.
1353        [Or at least, is documented not to: some releases seem to be in conflict
1354        with the documentation.] However, when a pattern was studied with
1355        pcre_study() and all its branches started with \s, PCRE still included VT
1356        as a possible starting character. Of course, this did no harm; it just
1357        caused an unnecessary match attempt.
1358    
1359    22. Removed a now-redundant internal flag bit that recorded the fact that case
1360        dependency changed within the pattern. This was once needed for "required
1361        byte" processing, but is no longer used. This recovers a now-scarce options
1362        bit. Also moved the least significant internal flag bit to the most-
1363        significant bit of the word, which was not previously used (hangover from
1364        the days when it was an int rather than a uint) to free up another bit for
1365        the future.
1366    
1367    23. Added support for CRLF line endings as well as CR and LF. As well as the
1368        default being selectable at build time, it can now be changed at runtime
1369        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1370        specify that it is scanning data with non-default line endings.
1371    
1372    24. Changed the definition of CXXLINK to make it agree with the definition of
1373        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1374    
1375    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1376        recursions. This makes a big different to stack usage for some patterns.
1377    
1378    26. If a subpattern containing a named recursion or subroutine reference such
1379        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1380        the space required for the compiled pattern went wrong and gave too small a
1381        value. Depending on the environment, this could lead to "Failed: internal
1382        error: code overflow at offset 49" or "glibc detected double free or
1383        corruption" errors.
1384    
1385    27. Applied patches from Google (a) to support the new newline modes and (b) to
1386        advance over multibyte UTF-8 characters in GlobalReplace.
1387    
1388    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1389        difference for some implementation of PCRE in some Windows version.
1390    
1391    29. Added some extra testing facilities to pcretest:
1392    
1393        \q<number>   in a data line sets the "match limit" value
1394        \Q<number>   in a data line sets the "match recursion limt" value
1395        -S <number>  sets the stack size, where <number> is in megabytes
1396    
1397        The -S option isn't available for Windows.
1398    
1399    
1400    Version 6.6 06-Feb-06
1401    ---------------------
1402    
1403     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1404        in pcreposix.h. I have copied the definition from pcre.h.
1405    
1406     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1407        because pcre.h is no longer a built file.
1408    
1409     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1410        not normally included in the compiled code.
1411    
1412    
1413  Version 6.5 01-Feb-06  Version 6.5 01-Feb-06
1414  ---------------------  ---------------------
1415    

Legend:
Removed from v.87  
changed lines
  Added in v.412

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12