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

Diff of /code/trunk/ChangeLog

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

revision 27 by nigel, Sat Feb 24 21:38:49 2007 UTC revision 886 by ph10, Tue Jan 17 14:25:07 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.30
5    ------------
6    
7    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
8        name is defined in ctype.h.
9    
10    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
11        only in quite long subpatterns.
12    
13    3.  Removed the function pcre_info(), which has been obsolete and deprecated
14        since it was replaced by pcre_fullinfo() in February 2000.
15    
16    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
17        match a (*MARK), and the match failed at the start of the subject, a
18        reference to memory before the start of the subject could occur. This bug
19        was introduced by fix 17 of release 8.21.
20    
21    5.  A reference to an unset group with zero minimum repetition was giving
22        totally wrong answers (in non-JavaScript-compatibility mode). For example,
23        /(another)?(\1?)test/ matched against "hello world test". This bug was
24        introduced in release 8.13.
25    
26    6.  Add support for 16-bit character strings (a large amount of work involving
27        many changes and refactorings).
28    
29    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31        from a file.
32    
33    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34        rounding is not applied in this particular case).
35    
36    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37        if they appear, or are escaped, in patterns.
38    
39    10. Get rid of a number of -Wunused-but-set-variable warnings.
40    
41    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
42        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
43        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
44        also returns the mark "x". This bug applied to capturing parentheses,
45        non-capturing parentheses, and atomic parentheses.
46    
47    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
48        information out of configure.ac instead of relying on pcre.h.generic, which
49        is not stored in the repository.
50    
51    
52    Version 8.21 12-Dec-2011
53    ------------------------
54    
55    1.  Updating the JIT compiler.
56    
57    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
58        are added as well.
59    
60    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
61        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
62        calling _pcre_jit_exec. Some extra comments are added.
63    
64    4.  (*MARK) settings inside atomic groups that do not contain any capturing
65        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
66        was introduced by change 18 for 8.20.
67    
68    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
69        ECMA-262 standard.
70    
71    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
72        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
73        This bug was probably introduced by change 9 of 8.13.
74    
75    7.  While fixing 6 above, I noticed that a number of other items were being
76        incorrectly rejected as "not fixed length". This arose partly because newer
77        opcodes had not been added to the fixed-length checking code. I have (a)
78        corrected the bug and added tests for these items, and (b) arranged for an
79        error to occur if an unknown opcode is encountered while checking for fixed
80        length instead of just assuming "not fixed length". The items that were
81        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
82        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
83        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
84    
85    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
86        being incorrectly compiled and would have given unpredicatble results.
87    
88    9.  A possessively repeated subpattern with minimum repeat count greater than
89        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
90        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
91        the first (A) could occur when it should not.
92    
93    10. Add a cast and remove a redundant test from the code.
94    
95    11. JIT should use pcre_malloc/pcre_free for allocation.
96    
97    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
98        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
99        is anything other than /usr, -L is still shown).
100    
101    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
102    
103    14. Perl does not support \N without a following name in a [] class; PCRE now
104        also gives an error.
105    
106    15. If a forward reference was repeated with an upper limit of around 2000,
107        it caused the error "internal error: overran compiling workspace". The
108        maximum number of forward references (including repeats) was limited by the
109        internal workspace, and dependent on the LINK_SIZE. The code has been
110        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
111        the default depends on LINK_SIZE. There is a new upper limit (for safety)
112        of around 200,000 forward references. While doing this, I also speeded up
113        the filling in of repeated forward references.
114    
115    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
116        incorrectly expecting the subject to contain another "a" after the start.
117    
118    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
119        in the match, the SKIP should be ignored. This was not happening; instead
120        the SKIP was being treated as NOMATCH. For patterns such as
121        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
122        tested.
123    
124    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
125        now much more compatible with Perl, in particular in cases where the result
126        is a non-match for a non-anchored pattern. For example, if
127        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
128        "m", where previously it did not return a name. A side effect of this
129        change is that for partial matches, the last encountered mark name is
130        returned, as for non matches. A number of tests that were previously not
131        Perl-compatible have been moved into the Perl-compatible test files. The
132        refactoring has had the pleasing side effect of removing one argument from
133        the match() function, thus reducing its stack requirements.
134    
135    19. If the /S+ option was used in pcretest to study a pattern using JIT,
136        subsequent uses of /S (without +) incorrectly behaved like /S+.
137    
138    21. Retrieve executable code size support for the JIT compiler and fixing
139        some warnings.
140    
141    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
142        not work when the shorter character appeared right at the end of the
143        subject string.
144    
145    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
146        systems.
147    
148    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
149        output it when the /M option is used in pcretest.
150    
151    25. The CheckMan script was not being included in the distribution. Also, added
152        an explicit "perl" to run Perl scripts from the PrepareRelease script
153        because this is reportedly needed in Windows.
154    
155    26. If study data was being save in a file and studying had not found a set of
156        "starts with" bytes for the pattern, the data written to the file (though
157        never used) was taken from uninitialized memory and so caused valgrind to
158        complain.
159    
160    27. Updated RunTest.bat as provided by Sheri Pierce.
161    
162    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
163    
164    29. Computation of memory usage for the table of capturing group names was
165        giving an unnecessarily large value.
166    
167    
168    Version 8.20 21-Oct-2011
169    ------------------------
170    
171    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
172        a POSIX class. After further experiments with Perl, which convinced me that
173        Perl has bugs and confusions, a closing square bracket is no longer allowed
174        in a POSIX name. This bug also affected patterns with classes that started
175        with full stops.
176    
177    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
178        captured substring, but while checking the failing first alternative,
179        substring 1 is temporarily captured. If the output vector supplied to
180        pcre_exec() was not big enough for this capture, the yield of the function
181        was still zero ("insufficient space for captured substrings"). This cannot
182        be totally fixed without adding another stack variable, which seems a lot
183        of expense for a edge case. However, I have improved the situation in cases
184        such as /(a)(b)x|abc/ matched against "abc", where the return code
185        indicates that fewer than the maximum number of slots in the ovector have
186        been set.
187    
188    3.  Related to (2) above: when there are more back references in a pattern than
189        slots in the output vector, pcre_exec() uses temporary memory during
190        matching, and copies in the captures as far as possible afterwards. It was
191        using the entire output vector, but this conflicts with the specification
192        that only 2/3 is used for passing back captured substrings. Now it uses
193        only the first 2/3, for compatibility. This is, of course, another edge
194        case.
195    
196    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
197        main code base, and can be used by building with --enable-jit. When this is
198        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
199        runtime --no-jit option is given.
200    
201    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
202        ovector, the return from the function was zero, implying that there were
203        other matches that did not fit. The correct "exactly full" value is now
204        returned.
205    
206    6.  If a subpattern that was called recursively or as a subroutine contained
207        (*PRUNE) or any other control that caused it to give a non-standard return,
208        invalid errors such as "Error -26 (nested recursion at the same subject
209        position)" or even infinite loops could occur.
210    
211    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
212        computing the minimum length on reaching *ACCEPT, and so ended up with the
213        wrong value of 1 rather than 0. Further investigation indicates that
214        computing a minimum subject length in the presence of *ACCEPT is difficult
215        (think back references, subroutine calls), and so I have changed the code
216        so that no minimum is registered for a pattern that contains *ACCEPT.
217    
218    8.  If (*THEN) was present in the first (true) branch of a conditional group,
219        it was not handled as intended. [But see 16 below.]
220    
221    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
222        Sheri Pierce.
223    
224    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
225        the first byte in a match must be "a".
226    
227    11. Change 17 for 8.13 increased the recursion depth for patterns like
228        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
229        pattern contains any instances of (*THEN). If it does not, the old
230        optimizations are restored. It would be nice to do this on a per-group
231        basis, but at the moment that is not feasible.
232    
233    12. In some environments, the output of pcretest -C is CRLF terminated. This
234        broke RunTest's code that checks for the link size. A single white space
235        character after the value is now allowed for.
236    
237    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
238        For "fr", it uses the Windows-specific input and output files.
239    
240    14. If (*THEN) appeared in a group that was called recursively or as a
241        subroutine, it did not work as intended. [But see next item.]
242    
243    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
244        pattern fragments (but not containing any | characters). If A and B are
245        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
246        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
247        D. In other words, Perl considers parentheses that do not contain any |
248        characters to be part of a surrounding alternative, whereas PCRE was
249        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
250        differently. PCRE now behaves in the same way as Perl, except in the case
251        of subroutine/recursion calls such as (?1) which have in any case always
252        been different (but PCRE had them first :-).
253    
254    16. Related to 15 above: Perl does not treat the | in a conditional group as
255        creating alternatives. Such a group is treated in the same way as an
256        ordinary group without any | characters when processing (*THEN). PCRE has
257        been changed to match Perl's behaviour.
258    
259    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
260        RunGrepTest script failed.
261    
262    18. Change 22 for version 13 caused atomic groups to use more stack. This is
263        inevitable for groups that contain captures, but it can lead to a lot of
264        stack use in large patterns. The old behaviour has been restored for atomic
265        groups that do not contain any capturing parentheses.
266    
267    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
268        suppress the check for a minimum subject length at run time. (If it was
269        given to pcre_exec() or pcre_dfa_exec() it did work.)
270    
271    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
272        fail to work when decoding hex characters in data strings in EBCDIC
273        environments.
274    
275    21. It appears that in at least one Mac OS environment, the isxdigit() function
276        is implemented as a macro that evaluates to its argument more than once,
277        contravening the C 90 Standard (I haven't checked a later standard). There
278        was an instance in pcretest which caused it to go wrong when processing
279        \x{...} escapes in subject strings. The has been rewritten to avoid using
280        things like p++ in the argument of isxdigit().
281    
282    
283    Version 8.13 16-Aug-2011
284    ------------------------
285    
286    1.  The Unicode data tables have been updated to Unicode 6.0.0.
287    
288    2.  Two minor typos in pcre_internal.h have been fixed.
289    
290    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
291        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
292        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
293    
294    4.  There were a number of related bugs in the code for matching backrefences
295        caselessly in UTF-8 mode when codes for the characters concerned were
296        different numbers of bytes. For example, U+023A and U+2C65 are an upper
297        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
298        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
299        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
300        2-byte code at the end of the subject (it thought there wasn't enough data
301        left).
302    
303    5.  Comprehensive information about what went wrong is now returned by
304        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
305        as the output vector has at least 2 elements. The offset of the start of
306        the failing character and a reason code are placed in the vector.
307    
308    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
309        now returned is for the first byte of the failing character, instead of the
310        last byte inspected. This is an incompatible change, but I hope it is small
311        enough not to be a problem. It makes the returned offset consistent with
312        pcre_exec() and pcre_dfa_exec().
313    
314    7.  pcretest now gives a text phrase as well as the error number when
315        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
316        failure, the offset and reason code are output.
317    
318    8.  When \R was used with a maximizing quantifier it failed to skip backwards
319        over a \r\n pair if the subsequent match failed. Instead, it just skipped
320        back over a single character (\n). This seems wrong (because it treated the
321        two characters as a single entity when going forwards), conflicts with the
322        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
323        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
324        has been changed.
325    
326    9.  Some internal refactoring has changed the processing so that the handling
327        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
328        time (the PCRE_DOTALL option was changed this way some time ago: version
329        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
330        which was always a bit of a fudge. It also means that there is one less
331        argument for the match() function, which reduces its stack requirements
332        slightly. This change also fixes an incompatibility with Perl: the pattern
333        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
334    
335    10. More internal refactoring has drastically reduced the number of recursive
336        calls to match() for possessively repeated groups such as (abc)++ when
337        using pcre_exec().
338    
339    11. While implementing 10, a number of bugs in the handling of groups were
340        discovered and fixed:
341    
342        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
343        (a|)*(?1) gave a compile-time internal error.
344        ((a|)+)+  did not notice that the outer group could match an empty string.
345        (^a|^)+   was not marked as anchored.
346        (.*a|.*)+ was not marked as matching at start or after a newline.
347    
348    12. Yet more internal refactoring has removed another argument from the match()
349        function. Special calls to this function are now indicated by setting a
350        value in a variable in the "match data" data block.
351    
352    13. Be more explicit in pcre_study() instead of relying on "default" for
353        opcodes that mean there is no starting character; this means that when new
354        ones are added and accidentally left out of pcre_study(), testing should
355        pick them up.
356    
357    14. The -s option of pcretest has been documented for ages as being an old
358        synonym of -m (show memory usage). I have changed it to mean "force study
359        for every regex", that is, assume /S for every regex. This is similar to -i
360        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
361        using it. It makes it easier to run collections of tests with and without
362        study enabled, and thereby test pcre_study() more easily. All the standard
363        tests are now run with and without -s (but some patterns can be marked as
364        "never study" - see 20 below).
365    
366    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
367        restoration of the capturing data to the outer values was not happening
368        correctly.
369    
370    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
371        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
372        pattern had matched an empty string, and so incorrectly returned a no
373        match.
374    
375    17. There was optimizing code for the last branch of non-capturing parentheses,
376        and also for the obeyed branch of a conditional subexpression, which used
377        tail recursion to cut down on stack usage. Unfortunately, now that there is
378        the possibility of (*THEN) occurring in these branches, tail recursion is
379        no longer possible because the return has to be checked for (*THEN). These
380        two optimizations have therefore been removed. [But see 8.20/11 above.]
381    
382    18. If a pattern containing \R was studied, it was assumed that \R always
383        matched two bytes, thus causing the minimum subject length to be
384        incorrectly computed because \R can also match just one byte.
385    
386    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
387        was incorrectly computed.
388    
389    20. If /S is present twice on a test pattern in pcretest input, it now
390        *disables* studying, thereby overriding the use of -s on the command line
391        (see 14 above). This is necessary for one or two tests to keep the output
392        identical in both cases.
393    
394    21. When (*ACCEPT) was used in an assertion that matched an empty string and
395        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
396    
397    22. When an atomic group that contained a capturing parenthesis was
398        successfully matched, but the branch in which it appeared failed, the
399        capturing was not being forgotten if a higher numbered group was later
400        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
401        group 1 to "a", when in fact it should be unset. This applied to multi-
402        branched capturing and non-capturing groups, repeated or not, and also to
403        positive assertions (capturing in negative assertions does not happen
404        in PCRE) and also to nested atomic groups.
405    
406    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
407        subject after a captured substring, to make it easier to tell which of a
408        number of identical substrings has been captured.
409    
410    24. The way atomic groups are processed by pcre_exec() has been changed so that
411        if they are repeated, backtracking one repetition now resets captured
412        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
413        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
414        "aaa". Previously, it would have been "a". As part of this code
415        refactoring, the way recursive calls are handled has also been changed.
416    
417    25. If an assertion condition captured any substrings, they were not passed
418        back unless some other capturing happened later. For example, if
419        (?(?=(a))a) was matched against "a", no capturing was returned.
420    
421    26. When studying a pattern that contained subroutine calls or assertions,
422        the code for finding the minimum length of a possible match was handling
423        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
424        group 1 called group 2 while simultaneously a separate group 2 called group
425        1). A stack overflow occurred in this case. I have fixed this by limiting
426        the recursion depth to 10.
427    
428    27. Updated RunTest.bat in the distribution to the version supplied by Tom
429        Fortmann. This supports explicit test numbers on the command line, and has
430        argument validation and error reporting.
431    
432    28. An instance of \X with an unlimited repeat could fail if at any point the
433        first character it looked at was a mark character.
434    
435    29. Some minor code refactoring concerning Unicode properties and scripts
436        should reduce the stack requirement of match() slightly.
437    
438    30. Added the '=' option to pcretest to check the setting of unused capturing
439        slots at the end of the pattern, which are documented as being -1, but are
440        not included in the return count.
441    
442    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
443        compiled something random. Now it gives a compile-time error (as does
444        Perl).
445    
446    32. A *MARK encountered during the processing of a positive assertion is now
447        recorded and passed back (compatible with Perl).
448    
449    33. If --only-matching or --colour was set on a pcregrep call whose pattern
450        had alternative anchored branches, the search for a second match in a line
451        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
452        matched the line "0102" twice. The same bug affected patterns that started
453        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
454        twice.
455    
456    34. Previously, PCRE did not allow quantification of assertions. However, Perl
457        does, and because of capturing effects, quantifying parenthesized
458        assertions may at times be useful. Quantifiers are now allowed for
459        parenthesized assertions.
460    
461    35. A minor code tidy in pcre_compile() when checking options for \R usage.
462    
463    36. \g was being checked for fancy things in a character class, when it should
464        just be a literal "g".
465    
466    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
467        appearance of a nested POSIX class supersedes an apparent external class.
468        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
469        unescaped square brackets may also appear as part of class names. For
470        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
471        more like Perl. (But see 8.20/1 above.)
472    
473    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
474        was because it thought it was \N{name}, which is not supported).
475    
476    39. Add minix to OS list not supporting the -S option in pcretest.
477    
478    40. PCRE tries to detect cases of infinite recursion at compile time, but it
479        cannot analyze patterns in sufficient detail to catch mutual recursions
480        such as ((?1))((?2)). There is now a runtime test that gives an error if a
481        subgroup is called recursively as a subpattern for a second time at the
482        same position in the subject string. In previous releases this might have
483        been caught by the recursion limit, or it might have run out of stack.
484    
485    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
486        happen only once. PCRE was, however incorrectly giving a compile time error
487        "recursive call could loop indefinitely" because it cannot analyze the
488        pattern in sufficient detail. The compile time test no longer happens when
489        PCRE is compiling a conditional subpattern, but actual runaway loops are
490        now caught at runtime (see 40 above).
491    
492    42. It seems that Perl allows any characters other than a closing parenthesis
493        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
494        has been changed to be the same.
495    
496    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
497        as not to get warnings when autogen.sh is called. Also changed
498        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
499    
500    44. To help people who use pcregrep to scan files containing exceedingly long
501        lines, the following changes have been made:
502    
503        (a) The default value of the buffer size parameter has been increased from
504            8K to 20K. (The actual buffer used is three times this size.)
505    
506        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
507            PCRE is built.
508    
509        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
510            to be set at run time.
511    
512        (d) Numerical values in pcregrep options can be followed by K or M, for
513            example --buffer-size=50K.
514    
515        (e) If a line being scanned overflows pcregrep's buffer, an error is now
516            given and the return code is set to 2.
517    
518    45. Add a pointer to the latest mark to the callout data block.
519    
520    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
521        partial match of an empty string instead of no match. This was specific to
522        the use of ".".
523    
524    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
525        complete match instead of a partial match. This bug was dependent on both
526        the PCRE_UTF8 and PCRE_DOTALL options being set.
527    
528    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
529        starting byte set, because \b was not being ignored.
530    
531    
532    Version 8.12 15-Jan-2011
533    ------------------------
534    
535    1.  Fixed some typos in the markup of the man pages, and wrote a script that
536        checks for such things as part of the documentation building process.
537    
538    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
539        --match-limit and --recursion-limit options (added for 8.11). In
540        particular, this made one of the standard tests fail. (The integer value
541        went into the wrong half of a long int.)
542    
543    3.  If the --colour option was given to pcregrep with -v (invert match), it
544        did strange things, either producing crazy output, or crashing. It should,
545        of course, ignore a request for colour when reporting lines that do not
546        match.
547    
548    4.  Another pcregrep bug caused similar problems if --colour was specified with
549        -M (multiline) and the pattern match finished with a line ending.
550    
551    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
552        matched in multiline mode, the following line was shown as part of the
553        match. This seems wrong, so I have changed it.
554    
555    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
556        the check for further matches in the same line (so they could be coloured)
557        to overrun the end of the current line. If another match was found, it was
558        incorrectly shown (and then shown again when found in the next line).
559    
560    7.  If pcregrep was compiled under Windows, there was a reference to the
561        function pcregrep_exit() before it was defined. I am assuming this was
562        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
563        reported by a user. I've moved the definition above the reference.
564    
565    
566    Version 8.11 10-Dec-2010
567    ------------------------
568    
569    1.  (*THEN) was not working properly if there were untried alternatives prior
570        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
571        backtracked to try for "b" instead of moving to the next alternative branch
572        at the same level (in this case, to look for "c"). The Perl documentation
573        is clear that when (*THEN) is backtracked onto, it goes to the "next
574        alternative in the innermost enclosing group".
575    
576    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
577        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
578        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
579        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
580        (*THEN).
581    
582    3.  If \s appeared in a character class, it removed the VT character from
583        the class, even if it had been included by some previous item, for example
584        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
585        of \s, but is part of the POSIX "space" class.)
586    
587    4.  A partial match never returns an empty string (because you can always
588        match an empty string at the end of the subject); however the checking for
589        an empty string was starting at the "start of match" point. This has been
590        changed to the "earliest inspected character" point, because the returned
591        data for a partial match starts at this character. This means that, for
592        example, /(?<=abc)def/ gives a partial match for the subject "abc"
593        (previously it gave "no match").
594    
595    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
596        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
597        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
598        has an implication that the given string is incomplete (because a partial
599        match is preferred over a full match). For this reason, these items now
600        give a partial match in this situation. [Aside: previously, the one case
601        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
602        match rather than a full match, which was wrong by the old rules, but is
603        now correct.]
604    
605    6.  There was a bug in the handling of #-introduced comments, recognized when
606        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
607        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
608        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
609        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
610        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
611        places in pcre_compile().
612    
613    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
614        comments when looking ahead for named forward references to subpatterns,
615        the only newline sequence it recognized was NL. It now handles newlines
616        according to the set newline convention.
617    
618    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
619        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
620        cater for a lack of strerror(). These oversights have been fixed.
621    
622    9.  Added --match-limit and --recursion-limit to pcregrep.
623    
624    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
625    
626    11. When the -o option was used, pcregrep was setting a return code of 1, even
627        when matches were found, and --line-buffered was not being honoured.
628    
629    12. Added an optional parentheses number to the -o and --only-matching options
630        of pcregrep.
631    
632    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
633        can match an empty string. The code to do it in pcretest and pcredemo
634        needed fixing:
635    
636        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
637            only one byte after an empty string match just before CRLF (this case
638            just got forgotten; "any" and "anycrlf" were OK).
639    
640        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
641            mode when an empty string match preceded an ASCII character followed by
642            a non-ASCII character. (The code for advancing by one character rather
643            than one byte was nonsense.)
644    
645        (c) The pcredemo.c sample program did not have any code at all to handle
646            the cases when CRLF is a valid newline sequence.
647    
648    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
649        as a starting offset was within the subject string. There is now a new
650        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
651        negative or greater than the length of the string. In order to test this,
652        pcretest is extended to allow the setting of negative starting offsets.
653    
654    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
655        starting offset points to the beginning of a UTF-8 character was
656        unnecessarily clumsy. I tidied it up.
657    
658    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
659        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
660    
661    17. Nobody had reported that the --include_dir option, which was added in
662        release 7.7 should have been called --include-dir (hyphen, not underscore)
663        for compatibility with GNU grep. I have changed it to --include-dir, but
664        left --include_dir as an undocumented synonym, and the same for
665        --exclude-dir, though that is not available in GNU grep, at least as of
666        release 2.5.4.
667    
668    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
669        characters from a string of bytes) have been redefined so as not to use
670        loops, in order to improve performance in some environments. At the same
671        time, I abstracted some of the common code into auxiliary macros to save
672        repetition (this should not affect the compiled code).
673    
674    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
675        compile-time error is now given if \c is not followed by an ASCII
676        character, that is, a byte less than 128. (In EBCDIC mode, the code is
677        different, and any byte value is allowed.)
678    
679    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
680        START_OPTIMIZE option, which is now allowed at compile time - but just
681        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
682        to pcregrep and other applications that have no direct access to PCRE
683        options. The new /Y option in pcretest sets this option when calling
684        pcre_compile().
685    
686    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
687        back references. Groups containing recursive back references were forced to
688        be atomic by that change, but in the case of named groups, the amount of
689        memory required was incorrectly computed, leading to "Failed: internal
690        error: code overflow". This has been fixed.
691    
692    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
693        pcretest.c, to avoid build problems in some Borland environments.
694    
695    
696    Version 8.10 25-Jun-2010
697    ------------------------
698    
699    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
700        THEN.
701    
702    2.  (*ACCEPT) was not working when inside an atomic group.
703    
704    3.  Inside a character class, \B is treated as a literal by default, but
705        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
706        causes the error). The code is unchanged, but I tidied the documentation.
707    
708    4.  Inside a character class, PCRE always treated \R and \X as literals,
709        whereas Perl faults them if its -w option is set. I have changed PCRE so
710        that it faults them when PCRE_EXTRA is set.
711    
712    5.  Added support for \N, which always matches any character other than
713        newline. (It is the same as "." when PCRE_DOTALL is not set.)
714    
715    6.  When compiling pcregrep with newer versions of gcc which may have
716        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
717        declared with attribute warn_unused_result" were given. Just casting the
718        result to (void) does not stop the warnings; a more elaborate fudge is
719        needed. I've used a macro to implement this.
720    
721    7.  Minor change to pcretest.c to avoid a compiler warning.
722    
723    8.  Added four artifical Unicode properties to help with an option to make
724        \s etc use properties (see next item). The new properties are: Xan
725        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
726    
727    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
728        use Unicode properties. (*UCP) at the start of a pattern can be used to set
729        this option. Modified pcretest to add /W to test this facility. Added
730        REG_UCP to make it available via the POSIX interface.
731    
732    10. Added --line-buffered to pcregrep.
733    
734    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
735        studied, and the match started with a letter with a code point greater than
736        127 whose first byte was different to the first byte of the other case of
737        the letter, the other case of this starting letter was not recognized
738        (#976).
739    
740    12. If a pattern that was studied started with a repeated Unicode property
741        test, for example, \p{Nd}+, there was the theoretical possibility of
742        setting up an incorrect bitmap of starting bytes, but fortunately it could
743        not have actually happened in practice until change 8 above was made (it
744        added property types that matched character-matching opcodes).
745    
746    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
747        possible starting bytes for non-anchored patterns.
748    
749    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
750        \R, and also a number of cases that involve Unicode properties, both
751        explicit and implicit when PCRE_UCP is set.
752    
753    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
754        input, it could crash or give wrong results if characters with values
755        greater than 0xc0 were present in the subject string. (Detail: it assumed
756        UTF-8 input when processing these items.)
757    
758    16. Added a lot of (int) casts to avoid compiler warnings in systems where
759        size_t is 64-bit (#991).
760    
761    17. Added a check for running out of memory when PCRE is compiled with
762        --disable-stack-for-recursion (#990).
763    
764    18. If the last data line in a file for pcretest does not have a newline on
765        the end, a newline was missing in the output.
766    
767    19. The default pcre_chartables.c file recognizes only ASCII characters (values
768        less than 128) in its various bitmaps. However, there is a facility for
769        generating tables according to the current locale when PCRE is compiled. It
770        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
771        characters, are recognized by isspace() and therefore were getting set in
772        these tables, and indeed these tables seem to approximate to ISO 8859. This
773        caused a problem in UTF-8 mode when pcre_study() was used to create a list
774        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
775        which of course cannot start UTF-8 characters. I have changed the code so
776        that only real ASCII characters (less than 128) and the correct starting
777        bytes for UTF-8 encodings are set for characters greater than 127 when in
778        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
779        altogether.)
780    
781    20. Added the /T option to pcretest so as to be able to run tests with non-
782        standard character tables, thus making it possible to include the tests
783        used for 19 above in the standard set of tests.
784    
785    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
786        reference to a subpattern the other side of a comment that contains an
787        opening parenthesis caused either an internal compiling error, or a
788        reference to the wrong subpattern.
789    
790    
791    Version 8.02 19-Mar-2010
792    ------------------------
793    
794    1.  The Unicode data tables have been updated to Unicode 5.2.0.
795    
796    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
797        configured.
798    
799    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
800        original author of that file, following a query about its status.
801    
802    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
803        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
804    
805    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
806        quantifier applied to a forward-referencing subroutine call, could compile
807        incorrect code or give the error "internal error: previously-checked
808        referenced subpattern not found".
809    
810    6.  Both MS Visual Studio and Symbian OS have problems with initializing
811        variables to point to external functions. For these systems, therefore,
812        pcre_malloc etc. are now initialized to local functions that call the
813        relevant global functions.
814    
815    7.  There were two entries missing in the vectors called coptable and poptable
816        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
817        I've fixed the data, and added a kludgy way of testing at compile time that
818        the lengths are correct (equal to the number of opcodes).
819    
820    8.  Following on from 7, I added a similar kludge to check the length of the
821        eint vector in pcreposix.c.
822    
823    9.  Error texts for pcre_compile() are held as one long string to avoid too
824        much relocation at load time. To find a text, the string is searched,
825        counting zeros. There was no check for running off the end of the string,
826        which could happen if a new error number was added without updating the
827        string.
828    
829    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
830    
831    11. \K was not working if it appeared in an atomic group or in a group that
832        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
833        \K is "not well defined" if used in an assertion. PCRE now accepts it if
834        the assertion is positive, but not if it is negative.
835    
836    12. Change 11 fortuitously reduced the size of the stack frame used in the
837        "match()" function of pcre_exec.c by one pointer. Forthcoming
838        implementation of support for (*MARK) will need an extra pointer on the
839        stack; I have reserved it now, so that the stack frame size does not
840        decrease.
841    
842    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
843        item in branch that calls a recursion is a subroutine call - as in the
844        second branch in the above example - was incorrectly given the compile-
845        time error "recursive call could loop indefinitely" because pcre_compile()
846        was not correctly checking the subroutine for matching a non-empty string.
847    
848    14. The checks for overrunning compiling workspace could trigger after an
849        overrun had occurred. This is a "should never occur" error, but it can be
850        triggered by pathological patterns such as hundreds of nested parentheses.
851        The checks now trigger 100 bytes before the end of the workspace.
852    
853    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
854    
855    
856    Version 8.01 19-Jan-2010
857    ------------------------
858    
859    1.  If a pattern contained a conditional subpattern with only one branch (in
860        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
861        computed the wrong minimum data length (which is of course zero for such
862        subpatterns). This could cause incorrect "no match" results.
863    
864    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
865        the pattern is reset in the first branch, pcre_compile() failed with
866        "internal error: code overflow at offset...". This happened only when
867        the reset was to the original external option setting. (An optimization
868        abstracts leading options settings into an external setting, which was the
869        cause of this.)
870    
871    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
872        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
873        assertion pattern did not match (meaning that the assertion was true), it
874        was incorrectly treated as false if the SKIP had been reached during the
875        matching. This also applied to assertions used as conditions.
876    
877    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
878        assertion subpattern, including such a pattern used as a condition,
879        unpredictable results occurred, instead of the error return
880        PCRE_ERROR_DFA_UITEM.
881    
882    5.  The C++ GlobalReplace function was not working like Perl for the special
883        situation when an empty string is matched. It now does the fancy magic
884        stuff that is necessary.
885    
886    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
887        removed. (These were left over from very, very early versions of PCRE.)
888    
889    7.  Some cosmetic changes to the code to make life easier when compiling it
890        as part of something else:
891    
892        (a) Change DEBUG to PCRE_DEBUG.
893    
894        (b) In pcre_compile(), rename the member of the "branch_chain" structure
895            called "current" as "current_branch", to prevent a collision with the
896            Linux macro when compiled as a kernel module.
897    
898        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
899            prevent a collision with the Linux macro when compiled as a kernel
900            module.
901    
902    8.  In pcre_compile() there are some checks for integer overflows that used to
903        cast potentially large values to (double). This has been changed to that
904        when building, a check for int64_t is made, and if it is found, it is used
905        instead, thus avoiding the use of floating point arithmetic. (There is no
906        other use of FP in PCRE.) If int64_t is not found, the fallback is to
907        double.
908    
909    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
910        2005 (difference between two addresses compared to an unsigned value).
911    
912    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
913        custom one, because of the following reported problem in Windows:
914    
915          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
916              under Win32.
917          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
918              therefore missing the function definition.
919          - The compiler thus generates a "C" signature for the test function.
920          - The linker fails to find the "C" function.
921          - PCRE fails to configure if asked to do so against libbz2.
922    
923    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
924        messages were output:
925    
926          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
927          rerunning libtoolize, to keep the correct libtool macros in-tree.
928          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
929    
930        I have done both of these things.
931    
932    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
933        most of the time, it *can* run out if it is given a pattern that contains a
934        runaway infinite recursion. I updated the discussion in the pcrestack man
935        page.
936    
937    13. Now that we have gone to the x.xx style of version numbers, the minor
938        version may start with zero. Using 08 or 09 is a bad idea because users
939        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
940        interpreted as invalid octal numbers. I've updated the previous comment in
941        configure.ac, and also added a check that gives an error if 08 or 09 are
942        used.
943    
944    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
945        causing partial matching to fail when the end of the subject matched \W
946        in a UTF-8 pattern where \W was quantified with a minimum of 3.
947    
948    15. There were some discrepancies between the declarations in pcre_internal.h
949        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
950        their definitions. The declarations used "const uschar *" and the
951        definitions used USPTR. Even though USPTR is normally defined as "const
952        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
953        reported that: "This difference in casting confuses some C++ compilers, for
954        example, SunCC recognizes above declarations as different functions and
955        generates broken code for hbpcre." I have changed the declarations to use
956        USPTR.
957    
958    16. GNU libtool is named differently on some systems. The autogen.sh script now
959        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
960        (FreeBSD).
961    
962    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
963        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
964        comment: "Figure out how to create a longlong from a string: strtoll and
965        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
966        instance, but it only takes 2 args instead of 3!"
967    
968    18. A subtle bug concerned with back references has been fixed by a change of
969        specification, with a corresponding code fix. A pattern such as
970        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
971        refers, was giving matches when it shouldn't. For example, xa=xaaa would
972        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
973        same bug. Such groups have to be quantified to be useful, or contained
974        inside another quantified group. (If there's no repetition, the reference
975        can never match.) The problem arises because, having left the group and
976        moved on to the rest of the pattern, a later failure that backtracks into
977        the group uses the captured value from the final iteration of the group
978        rather than the correct earlier one. I have fixed this in PCRE by forcing
979        any group that contains a reference to itself to be an atomic group; that
980        is, there cannot be any backtracking into it once it has completed. This is
981        similar to recursive and subroutine calls.
982    
983    
984    Version 8.00 19-Oct-09
985    ----------------------
986    
987    1.  The table for translating pcre_compile() error codes into POSIX error codes
988        was out-of-date, and there was no check on the pcre_compile() error code
989        being within the table. This could lead to an OK return being given in
990        error.
991    
992    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
993        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
994        in a Windows environment.
995    
996    3.  The pcregrep --count option prints the count for each file even when it is
997        zero, as does GNU grep. However, pcregrep was also printing all files when
998        --files-with-matches was added. Now, when both options are given, it prints
999        counts only for those files that have at least one match. (GNU grep just
1000        prints the file name in this circumstance, but including the count seems
1001        more useful - otherwise, why use --count?) Also ensured that the
1002        combination -clh just lists non-zero counts, with no names.
1003    
1004    4.  The long form of the pcregrep -F option was incorrectly implemented as
1005        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1006        but it seems right to fix it, and I didn't think it was worth preserving
1007        the old behaviour.
1008    
1009    5.  The command line items --regex=pattern and --regexp=pattern were not
1010        recognized by pcregrep, which required --regex pattern or --regexp pattern
1011        (with a space rather than an '='). The man page documented the '=' forms,
1012        which are compatible with GNU grep; these now work.
1013    
1014    6.  No libpcreposix.pc file was created for pkg-config; there was just
1015        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1016    
1017    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1018        when UCP support is not needed, by modifying the Python script that
1019        generates it from Unicode data files. This should not matter if the module
1020        is correctly used as a library, but I received one complaint about 50K of
1021        unwanted data. My guess is that the person linked everything into his
1022        program rather than using a library. Anyway, it does no harm.
1023    
1024    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1025        was a minimum greater than 1 for a wide character in a possessive
1026        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1027        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1028        character. Chaos in the form of incorrect output or a compiling loop could
1029        result.
1030    
1031    9.  The restrictions on what a pattern can contain when partial matching is
1032        requested for pcre_exec() have been removed. All patterns can now be
1033        partially matched by this function. In addition, if there are at least two
1034        slots in the offset vector, the offset of the earliest inspected character
1035        for the match and the offset of the end of the subject are set in them when
1036        PCRE_ERROR_PARTIAL is returned.
1037    
1038    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1039        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1040        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1041        and may be more useful for multi-segment matching.
1042    
1043    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1044        used to be given if ever the end of the subject was reached; now it is
1045        given only if matching could not proceed because another character was
1046        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1047        string "Z", which now yields "no match" instead of "partial match". In the
1048        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1049        final character ended with (*FAIL).
1050    
1051    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1052        if the pattern had a "must contain" character that was already found in the
1053        earlier partial match, unless partial matching was again requested. For
1054        example, with the pattern /dog.(body)?/, the "must contain" character is
1055        "g". If the first part-match was for the string "dog", restarting with
1056        "sbody" failed. This bug has been fixed.
1057    
1058    13. The string returned by pcre_dfa_exec() after a partial match has been
1059        changed so that it starts at the first inspected character rather than the
1060        first character of the match. This makes a difference only if the pattern
1061        starts with a lookbehind assertion or \b or \B (\K is not supported by
1062        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1063        matching functions compatible, and I think it's the right thing to do.
1064    
1065    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1066        so that the demonstration program is easily available in environments where
1067        PCRE has not been installed from source.
1068    
1069    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1070        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1071        library.
1072    
1073    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1074        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1075        is not the first non-POSIX option to be added. Clearly some people find
1076        these options useful.
1077    
1078    17. If a caller to the POSIX matching function regexec() passes a non-zero
1079        value for nmatch with a NULL value for pmatch, the value of
1080        nmatch is forced to zero.
1081    
1082    18. RunGrepTest did not have a test for the availability of the -u option of
1083        the diff command, as RunTest does. It now checks in the same way as
1084        RunTest, and also checks for the -b option.
1085    
1086    19. If an odd number of negated classes containing just a single character
1087        interposed, within parentheses, between a forward reference to a named
1088        subpattern and the definition of the subpattern, compilation crashed with
1089        an internal error, complaining that it could not find the referenced
1090        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1091        [The bug was that it was starting one character too far in when skipping
1092        over the character class, thus treating the ] as data rather than
1093        terminating the class. This meant it could skip too much.]
1094    
1095    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1096        /g option in pcretest when the pattern contains \K, which makes it possible
1097        to have an empty string match not at the start, even when the pattern is
1098        anchored. Updated pcretest and pcredemo to use this option.
1099    
1100    21. If the maximum number of capturing subpatterns in a recursion was greater
1101        than the maximum at the outer level, the higher number was returned, but
1102        with unset values at the outer level. The correct (outer level) value is
1103        now given.
1104    
1105    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1106        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1107        make it do so. The string so far is captured, making this feature
1108        compatible with Perl.
1109    
1110    23. The tests have been re-organized, adding tests 11 and 12, to make it
1111        possible to check the Perl 5.10 features against Perl 5.10.
1112    
1113    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1114        pattern matches a fixed length string. PCRE did not allow this; now it
1115        does. Neither allows recursion.
1116    
1117    25. I finally figured out how to implement a request to provide the minimum
1118        length of subject string that was needed in order to match a given pattern.
1119        (It was back references and recursion that I had previously got hung up
1120        on.) This code has now been added to pcre_study(); it finds a lower bound
1121        to the length of subject needed. It is not necessarily the greatest lower
1122        bound, but using it to avoid searching strings that are too short does give
1123        some useful speed-ups. The value is available to calling programs via
1124        pcre_fullinfo().
1125    
1126    26. While implementing 25, I discovered to my embarrassment that pcretest had
1127        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1128        study optimizations had never been tested with that matching function.
1129        Oops. What is worse, even when it was passed study data, there was a bug in
1130        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1131        were also very few tests of studied patterns with pcre_dfa_exec().
1132    
1133    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1134        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1135        on the other side of the coin, they are no longer allowed to have different
1136        names, because these cannot be distinguished in PCRE, and this has caused
1137        confusion. (This is a difference from Perl.)
1138    
1139    28. When duplicate subpattern names are present (necessarily with different
1140        numbers, as required by 27 above), and a test is made by name in a
1141        conditional pattern, either for a subpattern having been matched, or for
1142        recursion in such a pattern, all the associated numbered subpatterns are
1143        tested, and the overall condition is true if the condition is true for any
1144        one of them. This is the way Perl works, and is also more like the way
1145        testing by number works.
1146    
1147    
1148    Version 7.9 11-Apr-09
1149    ---------------------
1150    
1151    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1152        (pcretest), all targets were linked against these libraries. This included
1153        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1154        libraries. This caused unwanted dependencies to be created. This problem
1155        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1156        pcretest is linked with readline.
1157    
1158    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1159        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1160        moved outside it again, because FALSE and TRUE are already defined in AIX,
1161        but BOOL is not.
1162    
1163    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1164        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1165    
1166    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1167        hyphens) following file names and line numbers when outputting matching
1168        lines. This is not true; no spaces are inserted. I have also clarified the
1169        wording for the --colour (or --color) option.
1170    
1171    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1172        was not coloured; this is different to GNU grep, so I have changed it to be
1173        the same.
1174    
1175    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1176        each matching line was coloured. Now it goes on to look for further matches
1177        of any of the test patterns, which is the same behaviour as GNU grep.
1178    
1179    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1180        doesn't make sense to accept an empty string match in pcregrep, so I have
1181        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1182        seems to be how GNU grep behaves.
1183    
1184    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1185        start or after a newline", because the conditional assertion was not being
1186        correctly handled. The rule now is that both the assertion and what follows
1187        in the first alternative must satisfy the test.
1188    
1189    9.  If auto-callout was enabled in a pattern with a conditional group whose
1190        condition was an assertion, PCRE could crash during matching, both with
1191        pcre_exec() and pcre_dfa_exec().
1192    
1193    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1194        used for matching.
1195    
1196    11. Unicode property support in character classes was not working for
1197        characters (bytes) greater than 127 when not in UTF-8 mode.
1198    
1199    12. Added the -M command line option to pcretest.
1200    
1201    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1202    
1203    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1204    
1205    16. Added comments and documentation about mis-use of no_arg in the C++
1206        wrapper.
1207    
1208    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1209        from Martin Jerabek that uses macro names for all relevant character and
1210        string constants.
1211    
1212    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1213        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1214        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1215        these, but not everybody uses configure.
1216    
1217    19. A conditional group that had only one branch was not being correctly
1218        recognized as an item that could match an empty string. This meant that an
1219        enclosing group might also not be so recognized, causing infinite looping
1220        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1221        with the subject "ab", where knowledge that the repeated group can match
1222        nothing is needed in order to break the loop.
1223    
1224    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1225        exec(), but without supplying a callout function, matching went wrong.
1226    
1227    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1228        leak if the size of the offset vector was greater than 30. When the vector
1229        is smaller, the saved offsets during recursion go onto a local stack
1230        vector, but for larger vectors malloc() is used. It was failing to free
1231        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1232        error, in fact).
1233    
1234    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1235        heapframe that is used only when UTF-8 support is enabled. This caused no
1236        problem, but was untidy.
1237    
1238    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1239        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1240        included within another project.
1241    
1242    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1243        slightly modified by me:
1244    
1245          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1246              not building pcregrep.
1247    
1248          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1249              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1250    
1251    25. Forward references, both numeric and by name, in patterns that made use of
1252        duplicate group numbers, could behave incorrectly or give incorrect errors,
1253        because when scanning forward to find the reference group, PCRE was not
1254        taking into account the duplicate group numbers. A pattern such as
1255        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1256    
1257    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1258        the feature of a custom pointer more persuasive (as requested by a user).
1259    
1260    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1261        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1262        pre-defined.
1263    
1264    28. Added support for (*UTF8) at the start of a pattern.
1265    
1266    29. Arrange for flags added by the "release type" setting in CMake to be shown
1267        in the configuration summary.
1268    
1269    
1270    Version 7.8 05-Sep-08
1271    ---------------------
1272    
1273    1.  Replaced UCP searching code with optimized version as implemented for Ad
1274        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1275        stage table and inline lookup instead of a function, giving speed ups of 2
1276        to 5 times on some simple patterns that I tested. Permission was given to
1277        distribute the MultiStage2.py script that generates the tables (it's not in
1278        the tarball, but is in the Subversion repository).
1279    
1280    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1281        scripts.
1282    
1283    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1284        a group with a zero qualifier. The result of the study could be incorrect,
1285        or the function might crash, depending on the pattern.
1286    
1287    4.  Caseless matching was not working for non-ASCII characters in back
1288        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1289        It now works when Unicode Property Support is available.
1290    
1291    5.  In pcretest, an escape such as \x{de} in the data was always generating
1292        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1293        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1294        truncation.
1295    
1296    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1297    
1298    7.  Added two (int) casts to pcregrep when printing the difference of two
1299        pointers, in case they are 64-bit values.
1300    
1301    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1302        test 2 if it fails.
1303    
1304    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1305        and a #define of that name to empty if it is not externally set. This is to
1306        allow users of MSVC to set it if necessary.
1307    
1308    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1309        the convenience functions in the pcre_get.c source file.
1310    
1311    11. An option change at the start of a pattern that had top-level alternatives
1312        could cause overwriting and/or a crash. This command provoked a crash in
1313        some environments:
1314    
1315          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1316    
1317        This potential security problem was recorded as CVE-2008-2371.
1318    
1319    12. For a pattern where the match had to start at the beginning or immediately
1320        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1321        pcre_dfa_exec() could read past the end of the passed subject if there was
1322        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1323        pcretest so that it places the subject at the end of its malloc-ed buffer.
1324    
1325    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1326        exec() might read past the end of the data buffer in UTF-8 mode.
1327    
1328    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1329        the data contained the byte 0x85 as part of a UTF-8 character within its
1330        first line. This applied both to normal and DFA matching.
1331    
1332    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1333        /^[^d]*?$/8 failed to match "abc".
1334    
1335    16. Added a missing copyright notice to pcrecpp_internal.h.
1336    
1337    17. Make it more clear in the documentation that values returned from
1338        pcre_exec() in ovector are byte offsets, not character counts.
1339    
1340    18. Tidied a few places to stop certain compilers from issuing warnings.
1341    
1342    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1343        supplied by Stefan Weber. I made a further small update for 7.8 because
1344        there is a change of source arrangements: the pcre_searchfuncs.c module is
1345        replaced by pcre_ucd.c.
1346    
1347    
1348    Version 7.7 07-May-08
1349    ---------------------
1350    
1351    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1352        a string to a long long, pretend we don't even have a long long." This is
1353        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1354    
1355    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1356        pre-7.6 versions, which defined a global no_arg variable instead of putting
1357        it in the RE class. (See also #8 below.)
1358    
1359    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1360        Lopes.
1361    
1362    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1363    
1364        (1) The include/exclude patterns were being applied to the whole pathnames
1365            of files, instead of just to the final components.
1366    
1367        (2) If there was more than one level of directory, the subdirectories were
1368            skipped unless they satisfied the include/exclude conditions. This is
1369            inconsistent with GNU grep (and could even be seen as contrary to the
1370            pcregrep specification - which I improved to make it absolutely clear).
1371            The action now is always to scan all levels of directory, and just
1372            apply the include/exclude patterns to regular files.
1373    
1374    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1375        --exclude_dir in the tests to avoid scanning .svn directories.
1376    
1377    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1378        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1379        doesn't support NULs in patterns.
1380    
1381    7.  Added some missing "const"s to declarations of static tables in
1382        pcre_compile.c and pcre_dfa_exec.c.
1383    
1384    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1385        caused by fix #2  above. (Subsequently also a second patch to fix the
1386        first patch. And a third patch - this was a messy problem.)
1387    
1388    9.  Applied Craig's patch to remove the use of push_back().
1389    
1390    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1391        matching function regexec().
1392    
1393    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1394        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1395        references. PCRE supports relative numbers with this syntax (I don't think
1396        Oniguruma does).
1397    
1398    12. Previously, a group with a zero repeat such as (...){0} was completely
1399        omitted from the compiled regex. However, this means that if the group
1400        was called as a subroutine from elsewhere in the pattern, things went wrong
1401        (an internal error was given). Such groups are now left in the compiled
1402        pattern, with a new opcode that causes them to be skipped at execution
1403        time.
1404    
1405    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1406        to the way PCRE behaves:
1407    
1408        (a) A lone ] character is dis-allowed (Perl treats it as data).
1409    
1410        (b) A back reference to an unmatched subpattern matches an empty string
1411            (Perl fails the current match path).
1412    
1413        (c) A data ] in a character class must be notated as \] because if the
1414            first data character in a class is ], it defines an empty class. (In
1415            Perl it is not possible to have an empty class.) The empty class []
1416            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1417            The negative empty class [^] matches any one character, independently
1418            of the DOTALL setting.
1419    
1420    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1421        non-existent subpattern following a character class starting with ']' and
1422        containing () gave an internal compiling error instead of "reference to
1423        non-existent subpattern". Fortunately, when the pattern did exist, the
1424        compiled code was correct. (When scanning forwards to check for the
1425        existencd of the subpattern, it was treating the data ']' as terminating
1426        the class, so got the count wrong. When actually compiling, the reference
1427        was subsequently set up correctly.)
1428    
1429    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1430        it was being rejected as not supported by pcre_dfa_exec(), even though
1431        other assertions are supported. I have made pcre_dfa_exec() support
1432        (*FAIL).
1433    
1434    16. The implementation of 13c above involved the invention of a new opcode,
1435        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1436        cannot be changed at match time, I realized I could make a small
1437        improvement to matching performance by compiling OP_ALLANY instead of
1438        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1439        on the OP_ANY path.
1440    
1441    17. Compiling pcretest on Windows with readline support failed without the
1442        following two fixes: (1) Make the unistd.h include conditional on
1443        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1444    
1445    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1446        ncurses library to be included for pcretest when ReadLine support is
1447        requested, but also to allow for it to be overridden. This patch came from
1448        Daniel Bergström.
1449    
1450    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1451        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1452        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1453        spotting this.
1454    
1455    
1456    Version 7.6 28-Jan-08
1457    ---------------------
1458    
1459    1.  A character class containing a very large number of characters with
1460        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1461        overflow.
1462    
1463    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1464        HAVE_LONG_LONG is not defined.
1465    
1466    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1467        bring them up to date and include new features. This patch includes:
1468    
1469        - Fixed PH's badly added libz and libbz2 support.
1470        - Fixed a problem with static linking.
1471        - Added pcredemo. [But later removed - see 7 below.]
1472        - Fixed dftables problem and added an option.
1473        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1474            HAVE_LONG_LONG.
1475        - Added readline support for pcretest.
1476        - Added an listing of the option settings after cmake has run.
1477    
1478    4.  A user submitted a patch to Makefile that makes it easy to create
1479        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1480        Makefile.am that cause it to include this special target, without
1481        affecting anything else. Note that the same mingw target plus all
1482        the other distribution libraries and programs are now supported
1483        when configuring with CMake (see 6 below) instead of with
1484        Configure/Make.
1485    
1486    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1487        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1488        exported in the Windows port". It has not yet been confirmed that the patch
1489        solves the problem, but it does no harm.
1490    
1491    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1492        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1493        with CMake, and also correct the comment about stack recursion.
1494    
1495    7.  Remove the automatic building of pcredemo from the ./configure system and
1496        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1497        of a program that users should build themselves after PCRE is installed, so
1498        building it automatically is not really right. What is more, it gave
1499        trouble in some build environments.
1500    
1501    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1502    
1503    
1504    Version 7.5 10-Jan-08
1505    ---------------------
1506    
1507    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1508        values in parens when parsing an RE using the C++ wrapper."
1509    
1510    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1511        Characters greater than 255 were excluded from the class instead of being
1512        included.
1513    
1514    3.  The same bug as (2) above applied to negated POSIX classes such as
1515        [:^space:].
1516    
1517    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1518        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1519        I have changed it.
1520    
1521    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1522        first named subpattern) and a construct such as (?&a) would reference the
1523        first named subpattern whose name started with "a" (in other words, the
1524        length check was missing). Both these problems are fixed. "Subpattern name
1525        expected" is now given for (?&) (a zero-length name), and this patch also
1526        makes it give the same error for \k'' (previously it complained that that
1527        was a reference to a non-existent subpattern).
1528    
1529    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1530        this is right because (?- can be followed by option settings as well as by
1531        digits. I have, however, made the messages clearer.
1532    
1533    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1534        than the number used in the conditional) now cause a compile-time error.
1535        This is actually not compatible with Perl, which accepts such patterns, but
1536        treats the conditional as always being FALSE (as PCRE used to), but it
1537        seems to me that giving a diagnostic is better.
1538    
1539    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1540        and messages.
1541    
1542    9.  Fix two occurrences of "backslash" in comments that should have been
1543        "backspace".
1544    
1545    10. Remove two redundant lines of code that can never be obeyed (their function
1546        was moved elsewhere).
1547    
1548    11. The program that makes PCRE's Unicode character property table had a bug
1549        which caused it to generate incorrect table entries for sequences of
1550        characters that have the same character type, but are in different scripts.
1551        It amalgamated them into a single range, with the script of the first of
1552        them. In other words, some characters were in the wrong script. There were
1553        thirteen such cases, affecting characters in the following ranges:
1554    
1555          U+002b0 - U+002c1
1556          U+0060c - U+0060d
1557          U+0061e - U+00612
1558          U+0064b - U+0065e
1559          U+0074d - U+0076d
1560          U+01800 - U+01805
1561          U+01d00 - U+01d77
1562          U+01d9b - U+01dbf
1563          U+0200b - U+0200f
1564          U+030fc - U+030fe
1565          U+03260 - U+0327f
1566          U+0fb46 - U+0fbb1
1567          U+10450 - U+1049d
1568    
1569    12. The -o option (show only the matching part of a line) for pcregrep was not
1570        compatible with GNU grep in that, if there was more than one match in a
1571        line, it showed only the first of them. It now behaves in the same way as
1572        GNU grep.
1573    
1574    13. If the -o and -v options were combined for pcregrep, it printed a blank
1575        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1576        does the same. The return code can be used to tell if there were any
1577        non-matching lines.
1578    
1579    14. Added --file-offsets and --line-offsets to pcregrep.
1580    
1581    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1582        infinitely looping recursion. The bug was that positive lookaheads were not
1583        being skipped when checking for a possible empty match (negative lookaheads
1584        and both kinds of lookbehind were skipped).
1585    
1586    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1587        inclusion of <windows.h> to before rather than after the definition of
1588        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1589    
1590    17. Specifying a possessive quantifier with a specific limit for a Unicode
1591        character property caused pcre_compile() to compile bad code, which led at
1592        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1593        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1594        caused the error; without that there was no problem.
1595    
1596    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1597    
1598    19. Added --enable-pcretest-libreadline.
1599    
1600    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1601        RE::GlobalReplace(). As a result, the number of replacements returned was
1602        double what it should be. I removed one of the increments, but Craig sent a
1603        later patch that removed the other one (the right fix) and added unit tests
1604        that check the return values (which was not done before).
1605    
1606    21. Several CMake things:
1607    
1608        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1609            the names libpcre and libpcreposix, not just pcre and pcreposix.
1610    
1611        (2) The above change means that pcretest and pcregrep are now correctly
1612            linked with the newly-built libraries, not previously installed ones.
1613    
1614        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1615    
1616    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1617        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1618        UTF-8 newline character). The key issue is that the pattern starts .*;
1619        this means that the match must be either at the beginning, or after a
1620        newline. The bug was in the code for advancing after a failed match and
1621        checking that the new position followed a newline. It was not taking
1622        account of UTF-8 characters correctly.
1623    
1624    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1625        character classes. PCRE was not treating the sequence [:...:] as a
1626        character class unless the ... were all letters. Perl, however, seems to
1627        allow any characters between [: and :], though of course it rejects as
1628        unknown any "names" that contain non-letters, because all the known class
1629        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1630        for example, whereas PCRE did not - it did not recognize a POSIX character
1631        class. This seemed a bit dangerous, so the code has been changed to be
1632        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1633        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1634        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1635        Perl does, and where it didn't before.
1636    
1637    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1638        Windows environments %n is disabled by default.
1639    
1640    
1641    Version 7.4 21-Sep-07
1642    ---------------------
1643    
1644    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1645        means that a class such as [\s] counted as "explicit reference to CR or
1646        LF". That isn't really right - the whole point of the change was to try to
1647        help when there was an actual mention of one of the two characters. So now
1648        the change happens only if \r or \n (or a literal CR or LF) character is
1649        encountered.
1650    
1651    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1652        of both had grown to the point where there were only 3 bits left.
1653        Fortunately, there was spare space in the data structure, and so I have
1654        moved the internal flags into a new 16-bit field to free up more option
1655        bits.
1656    
1657    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1658        but did not set the internal JCHANGED flag - either of these is enough to
1659        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1660        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1661        start sets both bits.
1662    
1663    4.  Added options (at build time, compile time, exec time) to change \R from
1664        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1665    
1666    5.  doc/pcresyntax.html was missing from the distribution.
1667    
1668    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1669        compatibility, even though it is no longer used.
1670    
1671    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1672        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1673        windows.h file is present (where different names are used). [This was
1674        reversed later after testing - see 16 below.]
1675    
1676    8.  Changed all #include <config.h> to #include "config.h". There were also
1677        some further <pcre.h> cases that I changed to "pcre.h".
1678    
1679    9.  When pcregrep was used with the --colour option, it missed the line ending
1680        sequence off the lines that it output.
1681    
1682    10. It was pointed out to me that arrays of string pointers cause lots of
1683        relocations when a shared library is dynamically loaded. A technique of
1684        using a single long string with a table of offsets can drastically reduce
1685        these. I have refactored PCRE in four places to do this. The result is
1686        dramatic:
1687    
1688          Originally:                          290
1689          After changing UCP table:            187
1690          After changing error message table:   43
1691          After changing table of "verbs"       36
1692          After changing table of Posix names   22
1693    
1694        Thanks to the folks working on Gregex for glib for this insight.
1695    
1696    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1697        unicode-properties was also set.
1698    
1699    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1700    
1701    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1702        checked only for CRLF.
1703    
1704    14. Added casts to pcretest.c to avoid compiler warnings.
1705    
1706    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1707    
1708    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1709        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1710        entirely. This removes changes made in 7 above.
1711    
1712    17. The CMake files have been updated, and there is now more information about
1713        building with CMake in the NON-UNIX-USE document.
1714    
1715    
1716    Version 7.3 28-Aug-07
1717    ---------------------
1718    
1719     1. In the rejigging of the build system that eventually resulted in 7.1, the
1720        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1721        brackets there is not right, since it causes compilers to look for an
1722        installed pcre.h, not the version that is in the source that is being
1723        compiled (which of course may be different). I have changed it back to:
1724    
1725          #include "pcre.h"
1726    
1727        I have a vague recollection that the change was concerned with compiling in
1728        different directories, but in the new build system, that is taken care of
1729        by the VPATH setting the Makefile.
1730    
1731     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1732        when the subject happened to end in the byte 0x85 (e.g. if the last
1733        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1734        characters but of course it shouldn't be taken as a newline when it is part
1735        of another character. The bug was that, for an unlimited repeat of . in
1736        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1737        characters when looking for a newline.
1738    
1739     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1740    
1741     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1742        in debug output.
1743    
1744     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1745        long printing in the pcrecpp unittest when running under MinGW.
1746    
1747     6. ESC_K was left out of the EBCDIC table.
1748    
1749     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1750        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1751        limit also applies to "virtual nesting" when a pattern is recursive, and in
1752        this case 1000 isn't so big. I have been able to remove this limit at the
1753        expense of backing off one optimization in certain circumstances. Normally,
1754        when pcre_exec() would call its internal match() function recursively and
1755        immediately return the result unconditionally, it uses a "tail recursion"
1756        feature to save stack. However, when a subpattern that can match an empty
1757        string has an unlimited repetition quantifier, it no longer makes this
1758        optimization. That gives it a stack frame in which to save the data for
1759        checking that an empty string has been matched. Previously this was taken
1760        from the 1000-entry workspace that had been reserved. So now there is no
1761        explicit limit, but more stack is used.
1762    
1763     8. Applied Daniel's patches to solve problems with the import/export magic
1764        syntax that is required for Windows, and which was going wrong for the
1765        pcreposix and pcrecpp parts of the library. These were overlooked when this
1766        problem was solved for the main library.
1767    
1768     9. There were some crude static tests to avoid integer overflow when computing
1769        the size of patterns that contain repeated groups with explicit upper
1770        limits. As the maximum quantifier is 65535, the maximum group length was
1771        set at 30,000 so that the product of these two numbers did not overflow a
1772        32-bit integer. However, it turns out that people want to use groups that
1773        are longer than 30,000 bytes (though not repeat them that many times).
1774        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1775        made it possible to implement the integer overflow checks in a much more
1776        dynamic way, which I have now done. The artificial limitation on group
1777        length has been removed - we now have only the limit on the total length of
1778        the compiled pattern, which depends on the LINK_SIZE setting.
1779    
1780    10. Fixed a bug in the documentation for get/copy named substring when
1781        duplicate names are permitted. If none of the named substrings are set, the
1782        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1783        empty string.
1784    
1785    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1786        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1787        because the ] is interpreted as the first data character and the
1788        terminating ] is not found. PCRE has been made compatible with Perl in this
1789        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1790        cause memory overwriting.
1791    
1792    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1793        string has been matched (to stop an infinite loop). It was not recognizing
1794        a conditional subpattern that could match an empty string if that
1795        subpattern was within another subpattern. For example, it looped when
1796        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1797        condition was not nested. This bug has been fixed.
1798    
1799    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1800        past the start of the subject in the presence of bytes with the top bit
1801        set, for example "\x8aBCD".
1802    
1803    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1804        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1805    
1806    14. Optimized (?!) to (*FAIL).
1807    
1808    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1809        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1810        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1811        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1812        does: it's just the validity check that is more restrictive.
1813    
1814    16. Inserted checks for integer overflows during escape sequence (backslash)
1815        processing, and also fixed erroneous offset values for syntax errors during
1816        backslash processing.
1817    
1818    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1819        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1820    
1821    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1822        caused an overrun.
1823    
1824    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1825        something other than just ASCII characters) inside a group that had an
1826        unlimited repeat caused a loop at compile time (while checking to see
1827        whether the group could match an empty string).
1828    
1829    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1830        [\P{Any}] did so. (Error in the code for printing property names.)
1831    
1832    21. An orphan \E inside a character class could cause a crash.
1833    
1834    22. A repeated capturing bracket such as (A)? could cause a wild memory
1835        reference during compilation.
1836    
1837    23. There are several functions in pcre_compile() that scan along a compiled
1838        expression for various reasons (e.g. to see if it's fixed length for look
1839        behind). There were bugs in these functions when a repeated \p or \P was
1840        present in the pattern. These operators have additional parameters compared
1841        with \d, etc, and these were not being taken into account when moving along
1842        the compiled data. Specifically:
1843    
1844        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1845            length.
1846    
1847        (b) An item such as \pL+ within a repeated group could cause crashes or
1848            loops.
1849    
1850        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1851            "reference to non-existent subpattern" error.
1852    
1853        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1854    
1855    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1856        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1857    
1858    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1859    
1860    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1861        character were causing crashes (broken optimization).
1862    
1863    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1864        \p or \P) caused a compile-time loop.
1865    
1866    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1867        break. For example, the unstudied pattern [\r\n]A does not match the string
1868        "\r\nA" because change 7.0/46 below moves the current point on by two
1869        characters after failing to match at the start. However, the pattern \nA
1870        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1871        the same is true. There doesn't seem any very clean way out of this, but
1872        what I have chosen to do makes the common cases work: PCRE now takes note
1873        of whether there can be an explicit match for \r or \n anywhere in the
1874        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1875        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1876        pattern has explicit CR or LF references.
1877    
1878    29. Added (*CR) etc for changing newline setting at start of pattern.
1879    
1880    
1881    Version 7.2 19-Jun-07
1882    ---------------------
1883    
1884     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1885        which is apparently normally available under Windows.
1886    
1887     2. Re-jig the pcregrep tests with different newline settings in an attempt
1888        to make them independent of the local environment's newline setting.
1889    
1890     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1891    
1892     4. Some of the "internals" tests were previously cut out when the link size
1893        was not 2, because the output contained actual offsets. The recent new
1894        "Z" feature of pcretest means that these can be cut out, making the tests
1895        usable with all link sizes.
1896    
1897     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1898        stack recursion. This gives a massive performance boost under BSD, but just
1899        a small improvement under Linux. However, it saves one field in the frame
1900        in all cases.
1901    
1902     6. Added more features from the forthcoming Perl 5.10:
1903    
1904        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1905            recursion call. It refers to the nth most recently opened parentheses.
1906    
1907        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1908            to be opened parentheses.
1909    
1910        (c) Conditions that refer to capturing parentheses can be specified
1911            relatively, for example, (?(-2)... or (?(+3)...
1912    
1913        (d) \K resets the start of the current match so that everything before
1914            is not part of it.
1915    
1916        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1917    
1918        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1919            reference syntax.
1920    
1921        (g) (?| introduces a group in which the numbering of parentheses in each
1922            alternative starts with the same number.
1923    
1924        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1925    
1926     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1927        PCRE_INFO_JCHANGED.
1928    
1929     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1930        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1931        for detecting groups that can match an empty string.
1932    
1933     9. A pattern with a very large number of alternatives (more than several
1934        hundred) was running out of internal workspace during the pre-compile
1935        phase, where pcre_compile() figures out how much memory will be needed. A
1936        bit of new cunning has reduced the workspace needed for groups with
1937        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1938        workspace instead of running out of the 4096 that are available.
1939    
1940    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1941    
1942    11. Applied patch from Google to remove an optimization that didn't quite work.
1943        The report of the bug said:
1944    
1945          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1946          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1947          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1948    
1949    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1950        it matched the wrong number of bytes.
1951    
1952    
1953    Version 7.1 24-Apr-07
1954    ---------------------
1955    
1956     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1957        that is more "standard", making use of automake and other Autotools. There
1958        is some re-arrangement of the files and adjustment of comments consequent
1959        on this.
1960    
1961     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1962        for recursive directory scanning broke on some systems because the files
1963        are not scanned in any specific order and on different systems the order
1964        was different. A call to "sort" has been inserted into RunGrepTest for the
1965        approprate test as a short-term fix. In the longer term there may be an
1966        alternative.
1967    
1968     3. I had an email from Eric Raymond about problems translating some of PCRE's
1969        man pages to HTML (despite the fact that I distribute HTML pages, some
1970        people do their own conversions for various reasons). The problems
1971        concerned the use of low-level troff macros .br and .in. I have therefore
1972        removed all such uses from the man pages (some were redundant, some could
1973        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1974        HTML has been updated to handle .nf/.fi and to complain if it encounters
1975        .br or .in.
1976    
1977     4. Updated comments in configure.ac that get placed in config.h.in and also
1978        arranged for config.h to be included in the distribution, with the name
1979        config.h.generic, for the benefit of those who have to compile without
1980        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1981    
1982     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1983        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1984        makevp.bat for the current PCRE, using the additional files
1985        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1986    
1987     6. A Windows user reported a minor discrepancy with test 2, which turned out
1988        to be caused by a trailing space on an input line that had got lost in his
1989        copy. The trailing space was an accident, so I've just removed it.
1990    
1991     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1992        that is needed.
1993    
1994     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1995        as "const" (a) because they are and (b) because it helps the PHP
1996        maintainers who have recently made a script to detect big data structures
1997        in the php code that should be moved to the .rodata section. I remembered
1998        to update Builducptable as well, so it won't revert if ucptable.h is ever
1999        re-created.
2000    
2001     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2002        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2003        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2004        support is not required. This saves 1.5-2K of code, which is important in
2005        some applications.
2006    
2007        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2008        so as not to refer to the tables, even though these functions will never be
2009        called when UTF-8 support is disabled. Otherwise there are problems with a
2010        shared library.
2011    
2012    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2013    
2014        (a) It was defining its arguments as char * instead of void *.
2015    
2016        (b) It was assuming that all moves were upwards in memory; this was true
2017            a long time ago when I wrote it, but is no longer the case.
2018    
2019        The emulated memove() is provided for those environments that have neither
2020        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2021        is clearly not the case, as these two bugs were recently reported.
2022    
2023    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2024        and Detrail to create the HTML documentation, the .txt form of the man
2025        pages, and it removes trailing spaces from listed files. It also creates
2026        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2027        case, it wraps all the #defines with #ifndefs. This script should be run
2028        before "make dist".
2029    
2030    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2031        with Unicode property support.
2032    
2033        (a) For a maximizing quantifier, if the two different cases of the
2034            character were of different lengths in their UTF-8 codings (there are
2035            some cases like this - I found 11), and the matching function had to
2036            back up over a mixture of the two cases, it incorrectly assumed they
2037            were both the same length.
2038    
2039        (b) When PCRE was configured to use the heap rather than the stack for
2040            recursion during matching, it was not correctly preserving the data for
2041            the other case of a UTF-8 character when checking ahead for a match
2042            while processing a minimizing repeat. If the check also involved
2043            matching a wide character, but failed, corruption could cause an
2044            erroneous result when trying to check for a repeat of the original
2045            character.
2046    
2047    13. Some tidying changes to the testing mechanism:
2048    
2049        (a) The RunTest script now detects the internal link size and whether there
2050            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2051            values substituted by "configure". (The RunGrepTest script already did
2052            this for UTF-8.) The configure.ac script no longer substitutes the
2053            relevant variables.
2054    
2055        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2056            with length and offset values. This means that the output is different
2057            for different internal link sizes. Test 2 is skipped for link sizes
2058            other than 2 because of this, bypassing the problem. Unfortunately,
2059            there was also a test in test 3 (the locale tests) that used /B and
2060            failed for link sizes other than 2. Rather than cut the whole test out,
2061            I have added a new /Z option to pcretest that replaces the length and
2062            offset values with spaces. This is now used to make test 3 independent
2063            of link size. (Test 2 will be tidied up later.)
2064    
2065    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2066        segmentation fault instead of returning the appropriate error message.
2067    
2068    15. In multiline mode when the newline sequence was set to "any", the pattern
2069        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2070        This doesn't seem right; it now treats the CRLF combination as the line
2071        ending, and so does not match in that case. It's only a pattern such as ^$
2072        that would hit this one: something like ^ABC$ would have failed after \r
2073        and then tried again after \r\n.
2074    
2075    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2076        in an attempt to make files that differ only in their line terminators
2077        compare equal. This works on Linux.
2078    
2079    17. Under certain error circumstances pcregrep might try to free random memory
2080        as it exited. This is now fixed, thanks to valgrind.
2081    
2082    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2083        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2084        was because its rules for how to advance for /g after matching an empty
2085        string at the end of a line did not allow for this case. They now check for
2086        it specially.
2087    
2088    20. pcretest is supposed to handle patterns and data of any length, by
2089        extending its buffers when necessary. It was getting this wrong when the
2090        buffer for a data line had to be extended.
2091    
2092    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2093        CRLF as a newline sequence.
2094    
2095    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2096        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2097        I have nevertheless tidied it up.
2098    
2099    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2100    
2101    24. Added a man page for pcre-config.
2102    
2103    
2104    Version 7.0 19-Dec-06
2105    ---------------------
2106    
2107     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2108        moving to gcc 4.1.1.
2109    
2110     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2111        sys/time.h, which is documented as needed for this function. It doesn't
2112        seem to matter on Linux, but it showed up on some releases of OS X.
2113    
2114     3. It seems that there are systems where bytes whose values are greater than
2115        127 match isprint() in the "C" locale. The "C" locale should be the
2116        default when a C program starts up. In most systems, only ASCII printing
2117        characters match isprint(). This difference caused the output from pcretest
2118        to vary, making some of the tests fail. I have changed pcretest so that:
2119    
2120        (a) When it is outputting text in the compiled version of a pattern, bytes
2121            other than 32-126 are always shown as hex escapes.
2122    
2123        (b) When it is outputting text that is a matched part of a subject string,
2124            it does the same, unless a different locale has been set for the match
2125            (using the /L modifier). In this case, it uses isprint() to decide.
2126    
2127     4. Fixed a major bug that caused incorrect computation of the amount of memory
2128        required for a compiled pattern when options that changed within the
2129        pattern affected the logic of the preliminary scan that determines the
2130        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2131        that the computed length was too small. The symptoms of this bug were
2132        either the PCRE error "internal error: code overflow" from pcre_compile(),
2133        or a glibc crash with a message such as "pcretest: free(): invalid next
2134        size (fast)". Examples of patterns that provoked this bug (shown in
2135        pcretest format) are:
2136    
2137          /(?-x: )/x
2138          /(?x)(?-x: \s*#\s*)/
2139          /((?i)[\x{c0}])/8
2140          /(?i:[\x{c0}])/8
2141    
2142        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2143        is now done differently.
2144    
2145     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2146        wrapper classes; (b) implement a new function in the C++ scanner that is
2147        more efficient than the old way of doing things because it avoids levels of
2148        recursion in the regex matching; (c) add a paragraph to the documentation
2149        for the FullMatch() function.
2150    
2151     6. The escape sequence \n was being treated as whatever was defined as
2152        "newline". Not only was this contrary to the documentation, which states
2153        that \n is character 10 (hex 0A), but it also went horribly wrong when
2154        "newline" was defined as CRLF. This has been fixed.
2155    
2156     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2157        was being set to -1 for the "end of line" case (supposedly a value that no
2158        character can have). Though this value is never used (the check for end of
2159        line is "zero bytes in current character"), it caused compiler complaints.
2160        I've changed it to 0xffffffff.
2161    
2162     8. In pcre_version.c, the version string was being built by a sequence of
2163        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2164        string (as it is for production releases) called a macro with an empty
2165        argument. The C standard says the result of this is undefined. The gcc
2166        compiler treats it as an empty string (which was what was wanted) but it is
2167        reported that Visual C gives an error. The source has been hacked around to
2168        avoid this problem.
2169    
2170     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2171        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2172        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2173        of them did).
2174    
2175    10. Originally, pcretest opened its input and output without "b"; then I was
2176        told that "b" was needed in some environments, so it was added for release
2177        5.0 to both the input and output. (It makes no difference on Unix-like
2178        systems.) Later I was told that it is wrong for the input on Windows. I've
2179        now abstracted the modes into two macros, to make it easier to fiddle with
2180        them, and removed "b" from the input mode under Windows.
2181    
2182    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2183    
2184    12. Added -help and --help to pcretest as an official way of being reminded
2185        of the options.
2186    
2187    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2188        and pcrecpp.cc because they annoy compilers at high warning levels.
2189    
2190    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2191    
2192    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2193        scripts are not C programs :-) and which was not noticed because it works
2194        on Linux.
2195    
2196    16. pcretest is supposed to handle any length of pattern and data line (as one
2197        line or as a continued sequence of lines) by extending its input buffer if
2198        necessary. This feature was broken for very long pattern lines, leading to
2199        a string of junk being passed to pcre_compile() if the pattern was longer
2200        than about 50K.
2201    
2202    17. I have done a major re-factoring of the way pcre_compile() computes the
2203        amount of memory needed for a compiled pattern. Previously, there was code
2204        that made a preliminary scan of the pattern in order to do this. That was
2205        OK when PCRE was new, but as the facilities have expanded, it has become
2206        harder and harder to keep it in step with the real compile phase, and there
2207        have been a number of bugs (see for example, 4 above). I have now found a
2208        cunning way of running the real compile function in a "fake" mode that
2209        enables it to compute how much memory it would need, while actually only
2210        ever using a few hundred bytes of working memory and without too many
2211        tests of the mode. This should make future maintenance and development
2212        easier. A side effect of this work is that the limit of 200 on the nesting
2213        depth of parentheses has been removed (though this was never a serious
2214        limitation, I suspect). However, there is a downside: pcre_compile() now
2215        runs more slowly than before (30% or more, depending on the pattern). I
2216        hope this isn't a big issue. There is no effect on runtime performance.
2217    
2218    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2219        newline (only possible for the last line of a file) and it was a
2220        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2221    
2222    19. Added additional timing features to pcretest. (1) The -tm option now times
2223        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2224        separate command line item, by a number that specifies the number of
2225        repeats to use when timing. The default is 50000; this gives better
2226        precision, but takes uncomfortably long for very large patterns.
2227    
2228    20. Extended pcre_study() to be more clever in cases where a branch of a
2229        subpattern has no definite first character. For example, (a*|b*)[cd] would
2230        previously give no result from pcre_study(). Now it recognizes that the
2231        first character must be a, b, c, or d.
2232    
2233    21. There was an incorrect error "recursive call could loop indefinitely" if
2234        a subpattern (or the entire pattern) that was being tested for matching an
2235        empty string contained only one non-empty item after a nested subpattern.
2236        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2237        incorrectly, because the \d was being skipped in the check.
2238    
2239    22. The pcretest program now has a new pattern option /B and a command line
2240        option -b, which is equivalent to adding /B to every pattern. This causes
2241        it to show the compiled bytecode, without the additional information that
2242        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2243        is the same as /B/I).
2244    
2245    23. A new optimization is now able automatically to treat some sequences such
2246        as a*b as a*+b. More specifically, if something simple (such as a character
2247        or a simple class like \d) has an unlimited quantifier, and is followed by
2248        something that cannot possibly match the quantified thing, the quantifier
2249        is automatically "possessified".
2250    
2251    24. A recursive reference to a subpattern whose number was greater than 39
2252        went wrong under certain circumstances in UTF-8 mode. This bug could also
2253        have affected the operation of pcre_study().
2254    
2255    25. Realized that a little bit of performance could be had by replacing
2256        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2257    
2258    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2259    
2260    27. Possessive quantifiers such as a++ were previously implemented by turning
2261        them into atomic groups such as ($>a+). Now they have their own opcodes,
2262        which improves performance. This includes the automatically created ones
2263        from 23 above.
2264    
2265    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2266        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2267        the first matched character to be a colon. This applied both to named and
2268        numbered groups.
2269    
2270    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2271    
2272    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2273        building PCRE on Windows easier, so I have included it in the distribution.
2274    
2275    31. There is now a check in pcretest against a ridiculously large number being
2276        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2277        loop, the loop is abandoned.
2278    
2279    32. Forward references to subpatterns in conditions such as (?(2)...) where
2280        subpattern 2 is defined later cause pcre_compile() to search forwards in
2281        the pattern for the relevant set of parentheses. This search went wrong
2282        when there were unescaped parentheses in a character class, parentheses
2283        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2284    
2285    33. "Subroutine" calls and backreferences were previously restricted to
2286        referencing subpatterns earlier in the regex. This restriction has now
2287        been removed.
2288    
2289    34. Added a number of extra features that are going to be in Perl 5.10. On the
2290        whole, these are just syntactic alternatives for features that PCRE had
2291        previously implemented using the Python syntax or my own invention. The
2292        other formats are all retained for compatibility.
2293    
2294        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2295            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2296            also .NET compatible.
2297    
2298        (b) A recursion or subroutine call to a named group can now be defined as
2299            (?&name) as well as (?P>name).
2300    
2301        (c) A backreference to a named group can now be defined as \k<name> or
2302            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2303            5.10, are also .NET compatible.
2304    
2305        (d) A conditional reference to a named group can now use the syntax
2306            (?(<name>) or (?('name') as well as (?(name).
2307    
2308        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2309            groups (named and numbered) that are never evaluated inline, but can be
2310            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2311            is always false. There may be only one alternative in such a group.
2312    
2313        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2314            as the simple (?(R). The condition is true only if the most recent
2315            recursion is that of the given number or name. It does not search out
2316            through the entire recursion stack.
2317    
2318        (g) The escape \gN or \g{N} has been added, where N is a positive or
2319            negative number, specifying an absolute or relative reference.
2320    
2321    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2322        some "unreachable code" warnings.
2323    
2324    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2325        things, this adds five new scripts.
2326    
2327    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2328        There were also incompatibilities regarding the handling of \Q..\E inside
2329        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2330        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2331    
2332    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2333        matches an empty string, and forcibly breaks the loop. There were bugs in
2334        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2335        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2336        separate and independent bugs (that affected different cases) have been
2337        fixed.
2338    
2339    39. Refactored the code to abolish the use of different opcodes for small
2340        capturing bracket numbers. This is a tidy that I avoided doing when I
2341        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2342        The new approach is not only tidier, it makes it possible to reduce the
2343        memory needed to fix the previous bug (38).
2344    
2345    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2346        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2347        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2348        mode.
2349    
2350    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2351        report.
2352    
2353    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2354        copy construction and assignment in the C++ wrapper.
2355    
2356    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2357        couple of bugs that could have given wrong results in the "--newline=crlf"
2358        case.
2359    
2360    44. Added a number of casts and did some reorganization of signed/unsigned int
2361        variables following suggestions from Dair Grant. Also renamed the variable
2362        "this" as "item" because it is a C++ keyword.
2363    
2364    45. Arranged for dftables to add
2365    
2366          #include "pcre_internal.h"
2367    
2368        to pcre_chartables.c because without it, gcc 4.x may remove the array
2369        definition from the final binary if PCRE is built into a static library and
2370        dead code stripping is activated.
2371    
2372    46. For an unanchored pattern, if a match attempt fails at the start of a
2373        newline sequence, and the newline setting is CRLF or ANY, and the next two
2374        characters are CRLF, advance by two characters instead of one.
2375    
2376    
2377    Version 6.7 04-Jul-06
2378    ---------------------
2379    
2380     1. In order to handle tests when input lines are enormously long, pcretest has
2381        been re-factored so that it automatically extends its buffers when
2382        necessary. The code is crude, but this _is_ just a test program. The
2383        default size has been increased from 32K to 50K.
2384    
2385     2. The code in pcre_study() was using the value of the re argument before
2386        testing it for NULL. (Of course, in any sensible call of the function, it
2387        won't be NULL.)
2388    
2389     3. The memmove() emulation function in pcre_internal.h, which is used on
2390        systems that lack both memmove() and bcopy() - that is, hardly ever -
2391        was missing a "static" storage class specifier.
2392    
2393     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2394        containing an extended class (one that cannot be represented by a bitmap
2395        because it contains high-valued characters or Unicode property items, e.g.
2396        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2397        pattern, but PCRE should not loop if you do not (it no longer does).
2398        [Detail: two cases were found: (a) a repeated subpattern containing an
2399        extended class; (b) a recursive reference to a subpattern that followed a
2400        previous extended class. It wasn't skipping over the extended class
2401        correctly when UTF-8 mode was not set.]
2402    
2403     5. A negated single-character class was not being recognized as fixed-length
2404        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2405        compile error "lookbehind assertion is not fixed length".
2406    
2407     6. The RunPerlTest auxiliary script was showing an unexpected difference
2408        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2409        write a Perl script that can interpret lines of an input file either as
2410        byte characters or as UTF-8, which is what "perltest" was being required to
2411        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2412        can't do is switch easily at run time between having the "use utf8;" pragma
2413        or not. In the end, I fudged it by using the RunPerlTest script to insert
2414        "use utf8;" explicitly for the UTF-8 tests.
2415    
2416     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2417        the end of the subject string, contrary to the documentation and to what
2418        Perl does. This was true of both matching functions. Now it matches only at
2419        the start of the subject and immediately after *internal* newlines.
2420    
2421     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2422        a pointer to an int instead of a pointer to an unsigned long int. This
2423        caused problems on 64-bit systems.
2424    
2425     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2426        instance of the 'standard' template library not being so standard".
2427    
2428    10. There was no check on the number of named subpatterns nor the maximum
2429        length of a subpattern name. The product of these values is used to compute
2430        the size of the memory block for a compiled pattern. By supplying a very
2431        long subpattern name and a large number of named subpatterns, the size
2432        computation could be caused to overflow. This is now prevented by limiting
2433        the length of names to 32 characters, and the number of named subpatterns
2434        to 10,000.
2435    
2436    11. Subpatterns that are repeated with specific counts have to be replicated in
2437        the compiled pattern. The size of memory for this was computed from the
2438        length of the subpattern and the repeat count. The latter is limited to
2439        65535, but there was no limit on the former, meaning that integer overflow
2440        could in principle occur. The compiled length of a repeated subpattern is
2441        now limited to 30,000 bytes in order to prevent this.
2442    
2443    12. Added the optional facility to have named substrings with the same name.
2444    
2445    13. Added the ability to use a named substring as a condition, using the
2446        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2447        are numbers (not recommended). Forward references are permitted.
2448    
2449    14. Added forward references in named backreferences (if you see what I mean).
2450    
2451    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2452        pattern could run off the end of the subject. For example, the pattern
2453        "(?s)(.{1,5})"8 did this with the subject "ab".
2454    
2455    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2456        PCRE_CASELESS was set when matching characters that were quantified with ?
2457        or *.
2458    
2459    17. A character class other than a single negated character that had a minimum
2460        but no maximum quantifier - for example [ab]{6,} - was not handled
2461        correctly by pce_dfa_exec(). It would match only one character.
2462    
2463    18. A valid (though odd) pattern that looked like a POSIX character
2464        class but used an invalid character after [ (for example [[,abc,]]) caused
2465        pcre_compile() to give the error "Failed: internal error: code overflow" or
2466        in some cases to crash with a glibc free() error. This could even happen if
2467        the pattern terminated after [[ but there just happened to be a sequence of
2468        letters, a binary zero, and a closing ] in the memory that followed.
2469    
2470    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2471        over the years. Originally (before any Unicode support), just the bottom 8
2472        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2473        output from "man perlunicode" includes this:
2474    
2475          The regular expression compiler produces polymorphic opcodes.  That
2476          is, the pattern adapts to the data and automatically switches to
2477          the Unicode character scheme when presented with Unicode data--or
2478          instead uses a traditional byte scheme when presented with byte
2479          data.
2480    
2481        Sadly, a wide octal escape does not cause a switch, and in a string with
2482        no other multibyte characters, these octal escapes are treated as before.
2483        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2484        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2485        Unicode string.
2486    
2487        I have not perpetrated such confusion in PCRE. Up till now, it took just
2488        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2489        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2490        translate to the appropriate multibyte character.
2491    
2492    29. Applied some refactoring to reduce the number of warnings from Microsoft
2493        and Borland compilers. This has included removing the fudge introduced
2494        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2495        a warning about an unused variable.
2496    
2497    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2498        characters since release 4.0, because Perl (from release 5.004) does not.
2499        [Or at least, is documented not to: some releases seem to be in conflict
2500        with the documentation.] However, when a pattern was studied with
2501        pcre_study() and all its branches started with \s, PCRE still included VT
2502        as a possible starting character. Of course, this did no harm; it just
2503        caused an unnecessary match attempt.
2504    
2505    22. Removed a now-redundant internal flag bit that recorded the fact that case
2506        dependency changed within the pattern. This was once needed for "required
2507        byte" processing, but is no longer used. This recovers a now-scarce options
2508        bit. Also moved the least significant internal flag bit to the most-
2509        significant bit of the word, which was not previously used (hangover from
2510        the days when it was an int rather than a uint) to free up another bit for
2511        the future.
2512    
2513    23. Added support for CRLF line endings as well as CR and LF. As well as the
2514        default being selectable at build time, it can now be changed at runtime
2515        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2516        specify that it is scanning data with non-default line endings.
2517    
2518    24. Changed the definition of CXXLINK to make it agree with the definition of
2519        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2520    
2521    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2522        recursions. This makes a big different to stack usage for some patterns.
2523    
2524    26. If a subpattern containing a named recursion or subroutine reference such
2525        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2526        the space required for the compiled pattern went wrong and gave too small a
2527        value. Depending on the environment, this could lead to "Failed: internal
2528        error: code overflow at offset 49" or "glibc detected double free or
2529        corruption" errors.
2530    
2531    27. Applied patches from Google (a) to support the new newline modes and (b) to
2532        advance over multibyte UTF-8 characters in GlobalReplace.
2533    
2534    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2535        difference for some implementation of PCRE in some Windows version.
2536    
2537    29. Added some extra testing facilities to pcretest:
2538    
2539        \q<number>   in a data line sets the "match limit" value
2540        \Q<number>   in a data line sets the "match recursion limt" value
2541        -S <number>  sets the stack size, where <number> is in megabytes
2542    
2543        The -S option isn't available for Windows.
2544    
2545    
2546    Version 6.6 06-Feb-06
2547    ---------------------
2548    
2549     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2550        in pcreposix.h. I have copied the definition from pcre.h.
2551    
2552     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2553        because pcre.h is no longer a built file.
2554    
2555     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2556        not normally included in the compiled code.
2557    
2558    
2559    Version 6.5 01-Feb-06
2560    ---------------------
2561    
2562     1. When using the partial match feature with pcre_dfa_exec(), it was not
2563        anchoring the second and subsequent partial matches at the new starting
2564        point. This could lead to incorrect results. For example, with the pattern
2565        /1234/, partially matching against "123" and then "a4" gave a match.
2566    
2567     2. Changes to pcregrep:
2568    
2569        (a) All non-match returns from pcre_exec() were being treated as failures
2570            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2571            error message is output. Some extra information is given for the
2572            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2573            probably the only errors that are likely to be caused by users (by
2574            specifying a regex that has nested indefinite repeats, for instance).
2575            If there are more than 20 of these errors, pcregrep is abandoned.
2576    
2577        (b) A binary zero was treated as data while matching, but terminated the
2578            output line if it was written out. This has been fixed: binary zeroes
2579            are now no different to any other data bytes.
2580    
2581        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2582            used to set a locale for matching. The --locale=xxxx long option has
2583            been added (no short equivalent) to specify a locale explicitly on the
2584            pcregrep command, overriding the environment variables.
2585    
2586        (d) When -B was used with -n, some line numbers in the output were one less
2587            than they should have been.
2588    
2589        (e) Added the -o (--only-matching) option.
2590    
2591        (f) If -A or -C was used with -c (count only), some lines of context were
2592            accidentally printed for the final match.
2593    
2594        (g) Added the -H (--with-filename) option.
2595    
2596        (h) The combination of options -rh failed to suppress file names for files
2597            that were found from directory arguments.
2598    
2599        (i) Added the -D (--devices) and -d (--directories) options.
2600    
2601        (j) Added the -F (--fixed-strings) option.
2602    
2603        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2604    
2605        (l) Added the --colo(u)r option.
2606    
2607        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2608            is not present by default.
2609    
2610     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2611        items such as (?R) or (?1), when the recursion could match a number of
2612        alternatives. If it matched one of the alternatives, but subsequently,
2613        outside the recursion, there was a failure, the code tried to back up into
2614        the recursion. However, because of the way PCRE is implemented, this is not
2615        possible, and the result was an incorrect result from the match.
2616    
2617        In order to prevent this happening, the specification of recursion has
2618        been changed so that all such subpatterns are automatically treated as
2619        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2620    
2621     4. I had overlooked the fact that, in some locales, there are characters for
2622        which isalpha() is true but neither isupper() nor islower() are true. In
2623        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2624        and ordfeminine) are like this. This affected the treatment of \w and \W
2625        when they appeared in character classes, but not when they appeared outside
2626        a character class. The bit map for "word" characters is now created
2627        separately from the results of isalnum() instead of just taking it from the
2628        upper, lower, and digit maps. (Plus the underscore character, of course.)
2629    
2630     5. The above bug also affected the handling of POSIX character classes such as
2631        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2632        permanent tables. Instead, the bit maps for such a class were previously
2633        created as the appropriate unions of the upper, lower, and digit bitmaps.
2634        Now they are created by subtraction from the [[:word:]] class, which has
2635        its own bitmap.
2636    
2637     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2638        It is created by subtracting the vertical space characters (\x09, \x0a,
2639        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2640        subtraction was done in the overall bitmap for a character class, meaning
2641        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2642        be recognized. This bug has been fixed.
2643    
2644     7. Patches from the folks at Google:
2645    
2646          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2647          real life, but is still worth protecting against".
2648    
2649          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2650          regular expressions".
2651    
2652          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2653          have it.
2654    
2655          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2656          "configure" and the latter not, in order to fix a problem somebody had
2657          with compiling the Arg class on HP-UX.
2658    
2659          (e) Improve the error-handling of the C++ wrapper a little bit.
2660    
2661          (f) New tests for checking recursion limiting.
2662    
2663     8. The pcre_memmove() function, which is used only if the environment does not
2664        have a standard memmove() function (and is therefore rarely compiled),
2665        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2666        returning a result (though PCRE never actually uses the result).
2667    
2668     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2669        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2670        returned instead of calling malloc() with an overflowing number that would
2671        most likely cause subsequent chaos.
2672    
2673    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2674    
2675    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2676        with this option is matched, the nmatch and pmatch options of regexec() are
2677        ignored.
2678    
2679    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2680        provided in case anyone wants to the the POSIX interface with UTF-8
2681        strings.
2682    
2683    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2684        C++ linking (needed for some HP-UX environments).
2685    
2686    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2687        (unused parameter) and in the pcre_printint() function (omitted "default"
2688        switch label when the default is to do nothing).
2689    
2690    15. Added some code to make it possible, when PCRE is compiled as a C++
2691        library, to replace subject pointers for pcre_exec() with a smart pointer
2692        class, thus making it possible to process discontinuous strings.
2693    
2694    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2695        much the same function. They were added by different people who were trying
2696        to make PCRE easy to compile on non-Unix systems. It has been suggested
2697        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2698        for compiling on Windows systems. I have therefore replaced it with
2699        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2700        defaults to "extern" for C or "extern C" for C++, which works fine on
2701        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2702        SCOPE with something explicit in config.h. In addition:
2703    
2704        (a) pcreposix.h still had just "extern" instead of either of these macros;
2705            I have replaced it with PCRE_DATA_SCOPE.
2706    
2707        (b) Functions such as _pcre_xclass(), which are internal to the library,
2708            but external in the C sense, all had PCRE_EXPORT in their definitions.
2709            This is apparently wrong for the Windows case, so I have removed it.
2710            (It makes no difference on Unix-like systems.)
2711    
2712    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2713        of recursive calls to match(). This is different to MATCH_LIMIT because
2714        that limits the total number of calls to match(), not all of which increase
2715        the depth of recursion. Limiting the recursion depth limits the amount of
2716        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2717        when PCRE is compiled, and changed at run time. A patch from Google adds
2718        this functionality to the C++ interface.
2719    
2720    18. Changes to the handling of Unicode character properties:
2721    
2722        (a) Updated the table to Unicode 4.1.0.
2723    
2724        (b) Recognize characters that are not in the table as "Cn" (undefined).
2725    
2726        (c) I revised the way the table is implemented to a much improved format
2727            which includes recognition of ranges. It now supports the ranges that
2728            are defined in UnicodeData.txt, and it also amalgamates other
2729            characters into ranges. This has reduced the number of entries in the
2730            table from around 16,000 to around 3,000, thus reducing its size
2731            considerably. I realized I did not need to use a tree structure after
2732            all - a binary chop search is just as efficient. Having reduced the
2733            number of entries, I extended their size from 6 bytes to 8 bytes to
2734            allow for more data.
2735    
2736        (d) Added support for Unicode script names via properties such as \p{Han}.
2737    
2738    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2739        matching that character.
2740    
2741    20. When matching a repeated Unicode property with a minimum greater than zero,
2742        (for example \pL{2,}), PCRE could look past the end of the subject if it
2743        reached it while seeking the minimum number of characters. This could
2744        happen only if some of the characters were more than one byte long, because
2745        there is a check for at least the minimum number of bytes.
2746    
2747    21. Refactored the implementation of \p and \P so as to be more general, to
2748        allow for more different types of property in future. This has changed the
2749        compiled form incompatibly. Anybody with saved compiled patterns that use
2750        \p or \P will have to recompile them.
2751    
2752    22. Added "Any" and "L&" to the supported property types.
2753    
2754    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2755        but give a compile time error if the value is greater than 0xff.
2756    
2757    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2758        accidentally not being installed or uninstalled.
2759    
2760    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2761        made were to insert the current release number. This seemed silly, because
2762        it made things harder for people building PCRE on systems that don't run
2763        "configure". I have turned pcre.h into a distributed file, no longer built
2764        by "configure", with the version identification directly included. There is
2765        no longer a pcre.h.in file.
2766    
2767        However, this change necessitated a change to the pcre-config script as
2768        well. It is built from pcre-config.in, and one of the substitutions was the
2769        release number. I have updated configure.ac so that ./configure now finds
2770        the release number by grepping pcre.h.
2771    
2772    26. Added the ability to run the tests under valgrind.
2773    
2774    
2775    Version 6.4 05-Sep-05
2776    ---------------------
2777    
2778     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2779        "--" to be printed when multiple files were scanned, even when none of the
2780        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2781        consider it to be a bug, and have restored the previous behaviour.
2782    
2783     2. A couple of code tidies to get rid of compiler warnings.
2784    
2785     3. The pcretest program used to cheat by referring to symbols in the library
2786        whose names begin with _pcre_. These are internal symbols that are not
2787        really supposed to be visible externally, and in some environments it is
2788        possible to suppress them. The cheating is now confined to including
2789        certain files from the library's source, which is a bit cleaner.
2790    
2791     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2792        file's purpose clearer.
2793    
2794     5. Reorganized pcre_ucp_findchar().
2795    
2796    
2797    Version 6.3 15-Aug-05
2798    ---------------------
2799    
2800     1. The file libpcre.pc.in did not have general read permission in the tarball.
2801    
2802     2. There were some problems when building without C++ support:
2803    
2804        (a) If C++ support was not built, "make install" and "make test" still
2805            tried to test it.
2806    
2807        (b) There were problems when the value of CXX was explicitly set. Some
2808            changes have been made to try to fix these, and ...
2809    
2810        (c) --disable-cpp can now be used to explicitly disable C++ support.
2811    
2812        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2813            backslash in a target when C++ was disabled. This confuses some
2814            versions of "make", apparently. Using an intermediate variable solves
2815            this. (Same for CPP_LOBJ.)
2816    
2817     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2818        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2819        necessary on certain architectures.
2820    
2821     4. Added a setting of -export-symbols-regex to the link command to remove
2822        those symbols that are exported in the C sense, but actually are local
2823        within the library, and not documented. Their names all begin with
2824        "_pcre_". This is not a perfect job, because (a) we have to except some
2825        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2826        available (and never for static libraries). I have made a note to try to
2827        find a way round (a) in the future.
2828    
2829    
2830    Version 6.2 01-Aug-05
2831    ---------------------
2832    
2833     1. There was no test for integer overflow of quantifier values. A construction
2834        such as {1111111111111111} would give undefined results. What is worse, if
2835        a minimum quantifier for a parenthesized subpattern overflowed and became
2836        negative, the calculation of the memory size went wrong. This could have
2837        led to memory overwriting.
2838    
2839     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2840    
2841     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2842        operating environments where this matters.
2843    
2844     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2845        PCRE options from within the C++ wrapper.
2846    
2847     5. Named capturing subpatterns were not being correctly counted when a pattern
2848        was compiled. This caused two problems: (a) If there were more than 100
2849        such subpatterns, the calculation of the memory needed for the whole
2850        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2851        back references of the form \12, where the number was greater than 9, were
2852        not recognized as back references, even though there were sufficient
2853        previous subpatterns.
2854    
2855     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2856        versions of gcc, e.g. 2.95.4.
2857    
2858    
2859    Version 6.1 21-Jun-05
2860    ---------------------
2861    
2862     1. There was one reference to the variable "posix" in pcretest.c that was not
2863        surrounded by "#if !defined NOPOSIX".
2864    
2865     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2866        the cross-check on the old pcre_info() function, for the benefit of the
2867        cut-down version of PCRE that is currently imported into Exim.
2868    
2869     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2870        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2871        patterns that start (?i) but I don't think that matters. The use of (?i) is
2872        just an example; this all applies to the other options as well.
2873    
2874     4. Since libtool seems to echo the compile commands it is issuing, the output
2875        from "make" can be reduced a bit by putting "@" in front of each libtool
2876        compile command.
2877    
2878     5. Patch from the folks at Google for configure.in to be a bit more thorough
2879        in checking for a suitable C++ installation before trying to compile the
2880        C++ stuff. This should fix a reported problem when a compiler was present,
2881        but no suitable headers.
2882    
2883     6. The man pages all had just "PCRE" as their title. I have changed them to
2884        be the relevant file name. I have also arranged that these names are
2885        retained in the file doc/pcre.txt, which is a concatenation in text format
2886        of all the man pages except the little individual ones for each function.
2887    
2888     7. The NON-UNIX-USE file had not been updated for the different set of source
2889        files that come with release 6. I also added a few comments about the C++
2890        wrapper.
2891    
2892    
2893    Version 6.0 07-Jun-05
2894    ---------------------
2895    
2896     1. Some minor internal re-organization to help with my DFA experiments.
2897    
2898     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2899        didn't matter for the library itself when fully configured, but did matter
2900        when compiling without UCP support, or within Exim, where the ucp files are
2901        not imported.
2902    
2903     3. Refactoring of the library code to split up the various functions into
2904        different source modules. The addition of the new DFA matching code (see
2905        below) to a single monolithic source would have made it really too
2906        unwieldy, quite apart from causing all the code to be include in a
2907        statically linked application, when only some functions are used. This is
2908        relevant even without the DFA addition now that patterns can be compiled in
2909        one application and matched in another.
2910    
2911        The downside of splitting up is that there have to be some external
2912        functions and data tables that are used internally in different modules of
2913        the library but which are not part of the API. These have all had their
2914        names changed to start with "_pcre_" so that they are unlikely to clash
2915        with other external names.
2916    
2917     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2918        a different (DFA) algorithm. Although it is slower than the original
2919        function, it does have some advantages for certain types of matching
2920        problem.
2921    
2922     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2923        including restarting after a partial match.
2924    
2925     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2926        defined when compiling for Windows was sent to me. I have put it into the
2927        code, though I have no means of testing or verifying it.
2928    
2929     7. Added the pcre_refcount() auxiliary function.
2930    
2931     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2932        match before or at the first newline in the subject string. In pcretest,
2933        the /f option on a pattern can be used to set this.
2934    
2935     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2936        would behave wrongly. This has been present in PCRE since release 4.0.
2937    
2938    10. A number of changes to the pcregrep command:
2939    
2940        (a) Refactored how -x works; insert ^(...)$ instead of setting
2941            PCRE_ANCHORED and checking the length, in preparation for adding
2942            something similar for -w.
2943    
2944        (b) Added the -w (match as a word) option.
2945    
2946        (c) Refactored the way lines are read and buffered so as to have more
2947            than one at a time available.
2948    
2949        (d) Implemented a pcregrep test script.
2950    
2951        (e) Added the -M (multiline match) option. This allows patterns to match
2952            over several lines of the subject. The buffering ensures that at least
2953            8K, or the rest of the document (whichever is the shorter) is available
2954            for matching (and similarly the previous 8K for lookbehind assertions).
2955    
2956        (f) Changed the --help output so that it now says
2957    
2958              -w, --word-regex(p)
2959    
2960            instead of two lines, one with "regex" and the other with "regexp"
2961            because that confused at least one person since the short forms are the
2962            same. (This required a bit of code, as the output is generated
2963            automatically from a table. It wasn't just a text change.)
2964    
2965        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2966            option but starts with a hyphen. Could be a pattern or a path name
2967            starting with a hyphen, for instance.
2968    
2969        (h) "-" can be given as a file name to represent stdin.
2970    
2971        (i) When file names are being printed, "(standard input)" is used for
2972            the standard input, for compatibility with GNU grep. Previously
2973            "<stdin>" was used.
2974    
2975        (j) The option --label=xxx can be used to supply a name to be used for
2976            stdin when file names are being printed. There is no short form.
2977    
2978        (k) Re-factored the options decoding logic because we are going to add
2979            two more options that take data. Such options can now be given in four
2980            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2981    
2982        (l) Added the -A, -B, and -C options for requesting that lines of context
2983            around matches be printed.
2984    
2985        (m) Added the -L option to print the names of files that do not contain
2986            any matching lines, that is, the complement of -l.
2987    
2988        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2989            continue to scan other files.
2990    
2991        (o) The -s option was incorrectly implemented. For compatibility with other
2992            greps, it now suppresses the error message for a non-existent or non-
2993            accessible file (but not the return code). There is a new option called
2994            -q that suppresses the output of matching lines, which was what -s was
2995            previously doing.
2996    
2997        (p) Added --include and --exclude options to specify files for inclusion
2998            and exclusion when recursing.
2999    
3000    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3001        Hopefully, it now does.
3002    
3003    12. Missing cast in pcre_study().
3004    
3005    13. Added an "uninstall" target to the makefile.
3006    
3007    14. Replaced "extern" in the function prototypes in Makefile.in with
3008        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3009        world, but is set differently for Windows.
3010    
3011    15. Added a second compiling function called pcre_compile2(). The only
3012        difference is that it has an extra argument, which is a pointer to an
3013        integer error code. When there is a compile-time failure, this is set
3014        non-zero, in addition to the error test pointer being set to point to an
3015        error message. The new argument may be NULL if no error number is required
3016        (but then you may as well call pcre_compile(), which is now just a
3017        wrapper). This facility is provided because some applications need a
3018        numeric error indication, but it has also enabled me to tidy up the way
3019        compile-time errors are handled in the POSIX wrapper.
3020    
3021    16. Added VPATH=.libs to the makefile; this should help when building with one
3022        prefix path and installing with another. (Or so I'm told by someone who
3023        knows more about this stuff than I do.)
3024    
3025    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3026        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3027        match everything, including newlines. This is not POSIX-compatible, but
3028        somebody wanted the feature. From pcretest it can be activated by using
3029        both the P and the s flags.
3030    
3031    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3032    
3033    19. libpcre.pc was being incorrectly installed as executable.
3034    
3035    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3036        it now also looks for '\r' so that it will work unmodified on Windows.
3037    
3038    21. Added Google's contributed C++ wrapper to the distribution.
3039    
3040    22. Added some untidy missing memory free() calls in pcretest, to keep
3041        Electric Fence happy when testing.
3042    
3043    
3044    
3045    Version 5.0 13-Sep-04
3046    ---------------------
3047    
3048     1. Internal change: literal characters are no longer packed up into items
3049        containing multiple characters in a single byte-string. Each character
3050        is now matched using a separate opcode. However, there may be more than one
3051        byte in the character in UTF-8 mode.
3052    
3053     2. The pcre_callout_block structure has two new fields: pattern_position and
3054        next_item_length. These contain the offset in the pattern to the next match
3055        item, and its length, respectively.
3056    
3057     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3058        insertion of callouts before each pattern item. Added the /C option to
3059        pcretest to make use of this.
3060    
3061     4. On the advice of a Windows user, the lines
3062    
3063          #if defined(_WIN32) || defined(WIN32)
3064          _setmode( _fileno( stdout ), 0x8000 );
3065          #endif  /* defined(_WIN32) || defined(WIN32) */
3066    
3067        have been added to the source of pcretest. This apparently does useful
3068        magic in relation to line terminators.
3069    
3070     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3071        for the benefit of those environments where the "b" makes a difference.
3072    
3073     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3074        to know about it. I have put a hack into configure.in that adds in code
3075        to set GCC=yes if CC=icc. This seems to end up at a point in the
3076        generated configure script that is early enough to affect the setting of
3077        compiler options, which is what is needed, but I have no means of testing
3078        whether it really works. (The user who reported this had patched the
3079        generated configure script, which of course I cannot do.)
3080    
3081        LATER: After change 22 below (new libtool files), the configure script
3082        seems to know about icc (and also ecc). Therefore, I have commented out
3083        this hack in configure.in.
3084    
3085     7. Added support for pkg-config (2 patches were sent in).
3086    
3087     8. Negated POSIX character classes that used a combination of internal tables
3088        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3089        [[:^ascii]]. Typically, they would match almost any characters. The other
3090        POSIX classes were not broken in this way.
3091    
3092     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3093        to find the match, as PCRE was deluded into thinking that the match had to
3094        start at the start point or following a newline. The same bug applied to
3095        patterns with negative forward assertions or any backward assertions
3096        preceding ".*" at the start, unless the pattern required a fixed first
3097        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3098    
3099    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3100        starting at the last subject character, bytes beyond the end of the subject
3101        string were read.
3102    
3103    11. Renamed the variable "class" as "classbits" to make life easier for C++
3104        users. (Previously there was a macro definition, but it apparently wasn't
3105        enough.)
3106    
3107    12. Added the new field "tables" to the extra data so that tables can be passed
3108        in at exec time, or the internal tables can be re-selected. This allows
3109        a compiled regex to be saved and re-used at a later time by a different
3110        program that might have everything at different addresses.
3111    
3112    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3113        -R library as well as a -L library.
3114    
3115    14. The debugging options of pcretest (-d on the command line or D on a
3116        pattern) showed incorrect output for anything following an extended class
3117        that contained multibyte characters and which was followed by a quantifier.
3118    
3119    15. Added optional support for general category Unicode character properties
3120        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3121        support. It adds about 90K to the size of the library. The meanings of the
3122        inbuilt class escapes such as \d and \s have NOT been changed.
3123    
3124    16. Updated pcredemo.c to include calls to free() to release the memory for the
3125        compiled pattern.
3126    
3127    17. The generated file chartables.c was being created in the source directory
3128        instead of in the building directory. This caused the build to fail if the
3129        source directory was different from the building directory, and was
3130        read-only.
3131    
3132    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3133        file. No doubt somebody will tell me if they don't make sense... Also added
3134        Dan Mooney's comments about building on OpenVMS.
3135    
3136    19. Added support for partial matching via the PCRE_PARTIAL option for
3137        pcre_exec() and the \P data escape in pcretest.
3138    
3139    20. Extended pcretest with 3 new pattern features:
3140    
3141        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3142              write the compiled pattern to the file whose name is "rest-of-line".
3143              This is a straight binary dump of the data, with the saved pointer to
3144              the character tables forced to be NULL. The study data, if any, is
3145              written too. After writing, pcretest reads a new pattern.
3146    
3147        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3148              compiled pattern from the given file. There must not be any
3149              occurrences of "<" in the file name (pretty unlikely); if there are,
3150              pcretest will instead treat the initial "<" as a pattern delimiter.
3151              After reading in the pattern, pcretest goes on to read data lines as
3152              usual.
3153    
3154        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3155              and 16-bit fields in a compiled pattern, to simulate a pattern that
3156              was compiled on a host of opposite endianness.
3157    
3158    21. The pcre-exec() function can now cope with patterns that were compiled on
3159        hosts of opposite endianness, with this restriction:
3160    
3161          As for any compiled expression that is saved and used later, the tables
3162          pointer field cannot be preserved; the extra_data field in the arguments
3163          to pcre_exec() should be used to pass in a tables address if a value
3164          other than the default internal tables were used at compile time.
3165    
3166    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3167        now diagnosed as an error. Previously, most of the time, a negative number
3168        would have been treated as zero, but if in addition "ovector" was passed as
3169        NULL, a crash could occur.
3170    
3171    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3172        new versions from the libtool 1.5 distribution (the last one is a copy of
3173        a file called libtool.m4). This seems to have fixed the need to patch
3174        "configure" to support Darwin 1.3 (which I used to do). However, I still
3175        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3176        workstation).
3177    
3178    24. Changed the PCRE licence to be the more standard "BSD" licence.
3179    
3180    
3181    Version 4.5 01-Dec-03
3182    ---------------------
3183    
3184     1. There has been some re-arrangement of the code for the match() function so
3185        that it can be compiled in a version that does not call itself recursively.
3186        Instead, it keeps those local variables that need separate instances for
3187        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3188        needs to "recurse". Keeping track of where control must go is done by means
3189        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3190        hide most of the details from the main code, and operates only if
3191        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3192        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3193        operating.
3194    
3195        To make it easier for callers to provide specially tailored get/free
3196        functions for this usage, two new functions, pcre_stack_malloc, and
3197        pcre_stack_free, are used. They are always called in strict stacking order,
3198        and the size of block requested is always the same.
3199    
3200        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3201        PCRE has been compiled to use the stack or the heap for recursion. The
3202        -C option of pcretest uses this to show which version is compiled.
3203    
3204        A new data escape \S, is added to pcretest; it causes the amounts of store
3205        obtained and freed by both kinds of malloc/free at match time to be added
3206        to the output.
3207    
3208     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3209        what's available on my current Linux desktop machine.
3210    
3211     3. When matching a UTF-8 string, the test for a valid string at the start has
3212        been extended. If start_offset is not zero, PCRE now checks that it points
3213        to a byte that is the start of a UTF-8 character. If not, it returns
3214        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3215        this is necessary because there may be backward assertions in the pattern.
3216        When matching the same subject several times, it may save resources to use
3217        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3218    
3219     4. The code for checking the validity of UTF-8 strings has been tightened so
3220        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3221        containing "overlong sequences".
3222    
3223     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3224        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3225        should have been "&", but it just so happened that all the cases this let
3226        through by mistake were picked up later in the function.
3227    
3228     6. I had used a variable called "isblank" - this is a C99 function, causing
3229        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3230    
3231     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3232        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3233        so the version is shown; (c) stop "make test" echoing "./RunTest".
3234    
3235     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3236    
3237     9. The prototype for memmove() for systems that don't have it was using
3238        size_t, but the inclusion of the header that defines size_t was later. I've
3239        moved the #includes for the C headers earlier to avoid this.
3240    
3241    10. Added some adjustments to the code to make it easier to compiler on certain
3242        special systems:
3243    
3244          (a) Some "const" qualifiers were missing.
3245          (b) Added the macro EXPORT before all exported functions; by default this
3246              is defined to be empty.
3247          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3248              that it reads its output file name as an argument instead of writing
3249              to the standard output and assuming this can be redirected.
3250    
3251    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3252        class containing characters with values greater than 255, PCRE compilation
3253        went into a loop.
3254    
3255    12. A recursive reference to a subpattern that was within another subpattern
3256        that had a minimum quantifier of zero caused PCRE to crash. For example,
3257        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3258        recursion. If the recursively-called subpattern itself had a zero repeat,
3259        that was OK.
3260    
3261    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3262        buffer into which it was copied (for escape processing) was still set at
3263        1024, so long lines caused crashes.
3264    
3265    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3266        "internal error: code overflow...". This applied to any character class
3267        that was followed by a possessive quantifier.
3268    
3269    15. Modified the Makefile to add libpcre.la as a prerequisite for
3270        libpcreposix.la because I was told this is needed for a parallel build to
3271        work.
3272    
3273    16. If a pattern that contained .* following optional items at the start was
3274        studied, the wrong optimizing data was generated, leading to matching
3275        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3276        matching string must start with a or b or c. The correct conclusion for
3277        this pattern is that a match can start with any character.
3278    
3279    
3280    Version 4.4 13-Aug-03
3281    ---------------------
3282    
3283     1. In UTF-8 mode, a character class containing characters with values between
3284        127 and 255 was not handled correctly if the compiled pattern was studied.
3285        In fixing this, I have also improved the studying algorithm for such
3286        classes (slightly).
3287    
3288     2. Three internal functions had redundant arguments passed to them. Removal
3289        might give a very teeny performance improvement.
3290    
3291     3. Documentation bug: the value of the capture_top field in a callout is *one
3292        more than* the number of the hightest numbered captured substring.
3293    
3294     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3295        in incorrectly linking with a previously installed version. They now link
3296        explicitly with libpcre.la.
3297    
3298     5. configure.in no longer needs to recognize Cygwin specially.
3299    
3300     6. A problem in pcre.in for Windows platforms is fixed.
3301    
3302     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3303        pcretest, it used to include the size of the study block as part of its
3304        output. Unfortunately, the structure contains a field that has a different
3305        size on different hardware architectures. This meant that the tests that
3306        showed this size failed. As the block is currently always of a fixed size,
3307        this information isn't actually particularly useful in pcretest output, so
3308        I have just removed it.
3309    
3310     8. Three pre-processor statements accidentally did not start in column 1.
3311        Sadly, there are *still* compilers around that complain, even though
3312        standard C has not required this for well over a decade. Sigh.
3313    
3314     9. In pcretest, the code for checking callouts passed small integers in the
3315        callout_data field, which is a void * field. However, some picky compilers
3316        complained about the casts involved for this on 64-bit systems. Now
3317        pcretest passes the address of the small integer instead, which should get
3318        rid of the warnings.
3319    
3320    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3321        both compile and run time, and gives an error if an invalid UTF-8 sequence
3322        is found. There is a option for disabling this check in cases where the
3323        string is known to be correct and/or the maximum performance is wanted.
3324    
3325    11. In response to a bug report, I changed one line in Makefile.in from
3326    
3327            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3328        to
3329            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3330    
3331        to look similar to other lines, but I have no way of telling whether this
3332        is the right thing to do, as I do not use Windows. No doubt I'll get told
3333        if it's wrong...
3334    
3335    
3336    Version 4.3 21-May-03
3337    ---------------------
3338    
3339    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3340       Makefile.
3341    
3342    2. Some refactoring to improve the quality of the code:
3343    
3344       (i)   The utf8_table... variables are now declared "const".
3345    
3346       (ii)  The code for \cx, which used the "case flipping" table to upper case
3347             lower case letters, now just substracts 32. This is ASCII-specific,
3348             but the whole concept of \cx is ASCII-specific, so it seems
3349             reasonable.
3350    
3351       (iii) PCRE was using its character types table to recognize decimal and
3352             hexadecimal digits in the pattern. This is silly, because it handles
3353             only 0-9, a-f, and A-F, but the character types table is locale-
3354             specific, which means strange things might happen. A private
3355             table is now used for this - though it costs 256 bytes, a table is
3356             much faster than multiple explicit tests. Of course, the standard
3357             character types table is still used for matching digits in subject
3358             strings against \d.
3359    
3360       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3361             ending in _t are). So I've renamed it as ESC_tee.
3362    
3363    3. The first argument for regexec() in the POSIX wrapper should have been
3364       defined as "const".
3365    
3366    4. Changed pcretest to use malloc() for its buffers so that they can be
3367       Electric Fenced for debugging.
3368    
3369    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3370       to read one or more bytes before the start of the subject string. Often this
3371       had no effect on PCRE's behaviour, but in some circumstances it could
3372       provoke a segmentation fault.
3373    
3374    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3375       to try to read one or more bytes before the start of the subject string.
3376    
3377    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3378       UTF-8 support could misbehave in various ways if the subject string
3379       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3380       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3381       back over UTF-8 characters.)
3382    
3383    
3384    Version 4.2 14-Apr-03
3385    ---------------------
3386    
3387    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3388    
3389    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3390         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3391         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3392         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3393         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3394           and BUILD_EXEEXT
3395         Note: automatic setting of the BUILD variables is not yet working
3396         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3397           compile-time but not at link-time
3398         [LINK]: use for linking executables only
3399         make different versions for Windows and non-Windows
3400         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3401           libraries
3402         [LINK_FOR_BUILD]: new variable
3403         [OBJEXT]: use throughout
3404         [EXEEXT]: use throughout
3405         <winshared>: new target
3406         <wininstall>: new target
3407         <dftables.o>: use native compiler
3408         <dftables>: use native linker
3409         <install>: handle Windows platform correctly
3410         <clean>: ditto
3411         <check>: ditto
3412         copy DLL to top builddir before testing
3413    
3414       As part of these changes, -no-undefined was removed again. This was reported
3415       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3416       in any case.
3417    
3418    3. Some tidies to get rid of compiler warnings:
3419    
3420       . In the match_data structure, match_limit was an unsigned long int, whereas
3421         match_call_count was an int. I've made them both unsigned long ints.
3422    
3423       . In pcretest the fact that a const uschar * doesn't automatically cast to
3424         a void * provoked a warning.
3425    
3426       . Turning on some more compiler warnings threw up some "shadow" variables
3427         and a few more missing casts.
3428    
3429    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3430       option, a class that contained a single character with a value between 128
3431       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3432    
3433    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3434       option, a class that contained several characters, but with at least one
3435       whose value was between 128 and 255 caused PCRE to crash.
3436    
3437    
3438    Version 4.1 12-Mar-03
3439    ---------------------
3440    
3441    1. Compiling with gcc -pedantic found a couple of places where casts were
3442    needed, and a string in dftables.c that was longer than standard compilers are
3443    required to support.
3444    
3445    2. Compiling with Sun's compiler found a few more places where the code could
3446    be tidied up in order to avoid warnings.
3447    
3448    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3449    first of these names is deprecated in the latest Autoconf in favour of the name
3450    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3451    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3452    analogy I have changed it to CFLAGS_FOR_BUILD.
3453    
3454    4. Added -no-undefined to the linking command in the Makefile, because this is
3455    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3456    linking step for the pcreposix library.
3457    
3458    5. PCRE was failing to diagnose the case of two named groups with the same
3459    name.
3460    
3461    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3462    literal character that is needed in the subject for a match, and scans along to
3463    ensure that it is present before embarking on the full matching process. This
3464    saves time in cases of nested unlimited repeats that are never going to match.
3465    Problem: the scan can take a lot of time if the subject is very long (e.g.
3466    megabytes), thus penalizing straightforward matches. It is now done only if the
3467    amount of subject to be scanned is less than 1000 bytes.
3468    
3469    7. A lesser problem with the same optimization is that it was recording the
3470    first character of an anchored pattern as "needed", thus provoking a search
3471    right along the subject, even when the first match of the pattern was going to
3472    fail. The "needed" character is now not set for anchored patterns, unless it
3473    follows something in the pattern that is of non-fixed length. Thus, it still
3474    fulfils its original purpose of finding quick non-matches in cases of nested
3475    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3476    
3477    
3478    Version 4.0 17-Feb-03
3479    ---------------------
3480    
3481    1. If a comment in an extended regex that started immediately after a meta-item
3482    extended to the end of string, PCRE compiled incorrect data. This could lead to
3483    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3484    
3485    2. Moved to autoconf 2.53 and libtool 1.4.2.
3486    
3487    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3488    the special perltest8 script is no longer needed - all the tests can be run
3489    from a single perltest script.
3490    
3491    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3492    by \s. It has now been removed in PCRE. This means it isn't recognized as
3493    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3494    class [:space:] *does* include VT, thereby creating a mess.
3495    
3496    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3497    space and tab.
3498    
3499    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3500    its new features into the main test script, reducing the number of scripts.
3501    
3502    7. Perl 5.8 has changed the meaning of pattern