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

Diff of /code/trunk/ChangeLog

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

revision 7 by nigel, Sat Feb 24 21:38:09 2007 UTC revision 708 by ph10, Fri Sep 23 11:03:03 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.20 23-Sep-2011
5    ------------------------
6    
7    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
8        a POSIX class. After further experiments with Perl, which convinced me that
9        Perl has bugs and confusions, a closing square bracket is no longer allowed
10        in a POSIX name. This bug also affected patterns with classes that started
11        with full stops.
12    
13    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
14        captured substring, but while checking the failing first alternative,
15        substring 1 is temporarily captured. If the output vector supplied to
16        pcre_exec() was not big enough for this capture, the yield of the function
17        was still zero ("insufficient space for captured substrings"). This cannot
18        be totally fixed without adding another stack variable, which seems a lot
19        of expense for a edge case. However, I have improved the situation in cases
20        such as /(a)(b)x|abc/ matched against "abc", where the return code
21        indicates that fewer than the maximum number of slots in the ovector have
22        been set.
23    
24    3.  Related to (2) above: when there are more back references in a pattern than
25        slots in the output vector, pcre_exec() uses temporary memory during
26        matching, and copies in the captures as far as possible afterwards. It was
27        using the entire output vector, but this conflicts with the specification
28        that only 2/3 is used for passing back captured substrings. Now it uses
29        only the first 2/3, for compatibility. This is, of course, another edge
30        case.
31    
32    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
33        main code base, and can be used by building with --enable-jit. When this is
34        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
35        runtime --no-jit option is given.
36    
37    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
38        ovector, the return from the function was zero, implying that there were
39        other matches that did not fit. The correct "exactly full" value is now
40        returned.
41    
42    6.  If a subpattern that was called recursively or as a subroutine contained
43        (*PRUNE) or any other control that caused it to give a non-standard return,
44        invalid errors such as "Error -26 (nested recursion at the same subject
45        position)" or even infinite loops could occur.
46    
47    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
48        computing the minimum length on reaching *ACCEPT, and so ended up with the
49        wrong value of 1 rather than 0. Further investigation indicates that
50        computing a minimum subject length in the presence of *ACCEPT is difficult
51        (think back references, subroutine calls), and so I have changed the code
52        so that no minimum is registered for a pattern that contains *ACCEPT.
53    
54    8.  If (*THEN) was present in the first (true) branch of a conditional group,
55        it was not handled as intended.
56    
57    9.  Replaced RunTest.bat with the much improved version provided by Sheri
58        Pierce.
59    
60    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
61        the first byte in a match must be "a".
62    
63    11. Change 17 for 8.13 increased the recursion depth for patterns like
64        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
65        pattern contains any instances of (*THEN). If it does not, the old
66        optimizations are restored. It would be nice to do this on a per-group
67        basis, but at the moment that is not feasible.
68    
69    
70    Version 8.13 16-Aug-2011
71    ------------------------
72    
73    1.  The Unicode data tables have been updated to Unicode 6.0.0.
74    
75    2.  Two minor typos in pcre_internal.h have been fixed.
76    
77    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
78        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
79        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
80    
81    4.  There were a number of related bugs in the code for matching backrefences
82        caselessly in UTF-8 mode when codes for the characters concerned were
83        different numbers of bytes. For example, U+023A and U+2C65 are an upper
84        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
85        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
86        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
87        2-byte code at the end of the subject (it thought there wasn't enough data
88        left).
89    
90    5.  Comprehensive information about what went wrong is now returned by
91        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
92        as the output vector has at least 2 elements. The offset of the start of
93        the failing character and a reason code are placed in the vector.
94    
95    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
96        now returned is for the first byte of the failing character, instead of the
97        last byte inspected. This is an incompatible change, but I hope it is small
98        enough not to be a problem. It makes the returned offset consistent with
99        pcre_exec() and pcre_dfa_exec().
100    
101    7.  pcretest now gives a text phrase as well as the error number when
102        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
103        failure, the offset and reason code are output.
104    
105    8.  When \R was used with a maximizing quantifier it failed to skip backwards
106        over a \r\n pair if the subsequent match failed. Instead, it just skipped
107        back over a single character (\n). This seems wrong (because it treated the
108        two characters as a single entity when going forwards), conflicts with the
109        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
110        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
111        has been changed.
112    
113    9.  Some internal refactoring has changed the processing so that the handling
114        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
115        time (the PCRE_DOTALL option was changed this way some time ago: version
116        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
117        which was always a bit of a fudge. It also means that there is one less
118        argument for the match() function, which reduces its stack requirements
119        slightly. This change also fixes an incompatibility with Perl: the pattern
120        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
121    
122    10. More internal refactoring has drastically reduced the number of recursive
123        calls to match() for possessively repeated groups such as (abc)++ when
124        using pcre_exec().
125    
126    11. While implementing 10, a number of bugs in the handling of groups were
127        discovered and fixed:
128    
129        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
130        (a|)*(?1) gave a compile-time internal error.
131        ((a|)+)+  did not notice that the outer group could match an empty string.
132        (^a|^)+   was not marked as anchored.
133        (.*a|.*)+ was not marked as matching at start or after a newline.
134    
135    12. Yet more internal refactoring has removed another argument from the match()
136        function. Special calls to this function are now indicated by setting a
137        value in a variable in the "match data" data block.
138    
139    13. Be more explicit in pcre_study() instead of relying on "default" for
140        opcodes that mean there is no starting character; this means that when new
141        ones are added and accidentally left out of pcre_study(), testing should
142        pick them up.
143    
144    14. The -s option of pcretest has been documented for ages as being an old
145        synonym of -m (show memory usage). I have changed it to mean "force study
146        for every regex", that is, assume /S for every regex. This is similar to -i
147        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
148        using it. It makes it easier to run collections of tests with and without
149        study enabled, and thereby test pcre_study() more easily. All the standard
150        tests are now run with and without -s (but some patterns can be marked as
151        "never study" - see 20 below).
152    
153    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
154        restoration of the capturing data to the outer values was not happening
155        correctly.
156    
157    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
158        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
159        pattern had matched an empty string, and so incorrectly returned a no
160        match.
161    
162    17. There was optimizing code for the last branch of non-capturing parentheses,
163        and also for the obeyed branch of a conditional subexpression, which used
164        tail recursion to cut down on stack usage. Unfortunately, now that there is
165        the possibility of (*THEN) occurring in these branches, tail recursion is
166        no longer possible because the return has to be checked for (*THEN). These
167        two optimizations have therefore been removed. [But see 8.20/11 above.]
168    
169    18. If a pattern containing \R was studied, it was assumed that \R always
170        matched two bytes, thus causing the minimum subject length to be
171        incorrectly computed because \R can also match just one byte.
172    
173    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
174        was incorrectly computed.
175    
176    20. If /S is present twice on a test pattern in pcretest input, it now
177        *disables* studying, thereby overriding the use of -s on the command line
178        (see 14 above). This is necessary for one or two tests to keep the output
179        identical in both cases.
180    
181    21. When (*ACCEPT) was used in an assertion that matched an empty string and
182        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
183    
184    22. When an atomic group that contained a capturing parenthesis was
185        successfully matched, but the branch in which it appeared failed, the
186        capturing was not being forgotten if a higher numbered group was later
187        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
188        group 1 to "a", when in fact it should be unset. This applied to multi-
189        branched capturing and non-capturing groups, repeated or not, and also to
190        positive assertions (capturing in negative assertions does not happen
191        in PCRE) and also to nested atomic groups.
192    
193    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
194        subject after a captured substring, to make it easier to tell which of a
195        number of identical substrings has been captured.
196    
197    24. The way atomic groups are processed by pcre_exec() has been changed so that
198        if they are repeated, backtracking one repetition now resets captured
199        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
200        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
201        "aaa". Previously, it would have been "a". As part of this code
202        refactoring, the way recursive calls are handled has also been changed.
203    
204    25. If an assertion condition captured any substrings, they were not passed
205        back unless some other capturing happened later. For example, if
206        (?(?=(a))a) was matched against "a", no capturing was returned.
207    
208    26. When studying a pattern that contained subroutine calls or assertions,
209        the code for finding the minimum length of a possible match was handling
210        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
211        group 1 called group 2 while simultaneously a separate group 2 called group
212        1). A stack overflow occurred in this case. I have fixed this by limiting
213        the recursion depth to 10.
214    
215    27. Updated RunTest.bat in the distribution to the version supplied by Tom
216        Fortmann. This supports explicit test numbers on the command line, and has
217        argument validation and error reporting.
218    
219    28. An instance of \X with an unlimited repeat could fail if at any point the
220        first character it looked at was a mark character.
221    
222    29. Some minor code refactoring concerning Unicode properties and scripts
223        should reduce the stack requirement of match() slightly.
224    
225    30. Added the '=' option to pcretest to check the setting of unused capturing
226        slots at the end of the pattern, which are documented as being -1, but are
227        not included in the return count.
228    
229    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
230        compiled something random. Now it gives a compile-time error (as does
231        Perl).
232    
233    32. A *MARK encountered during the processing of a positive assertion is now
234        recorded and passed back (compatible with Perl).
235    
236    33. If --only-matching or --colour was set on a pcregrep call whose pattern
237        had alternative anchored branches, the search for a second match in a line
238        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
239        matched the line "0102" twice. The same bug affected patterns that started
240        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
241        twice.
242    
243    34. Previously, PCRE did not allow quantification of assertions. However, Perl
244        does, and because of capturing effects, quantifying parenthesized
245        assertions may at times be useful. Quantifiers are now allowed for
246        parenthesized assertions.
247    
248    35. A minor code tidy in pcre_compile() when checking options for \R usage.
249    
250    36. \g was being checked for fancy things in a character class, when it should
251        just be a literal "g".
252    
253    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
254        appearance of a nested POSIX class supersedes an apparent external class.
255        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
256        unescaped square brackets may also appear as part of class names. For
257        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
258        more like Perl. (But see 8.20/1 above.)
259    
260    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
261        was because it thought it was \N{name}, which is not supported).
262    
263    39. Add minix to OS list not supporting the -S option in pcretest.
264    
265    40. PCRE tries to detect cases of infinite recursion at compile time, but it
266        cannot analyze patterns in sufficient detail to catch mutual recursions
267        such as ((?1))((?2)). There is now a runtime test that gives an error if a
268        subgroup is called recursively as a subpattern for a second time at the
269        same position in the subject string. In previous releases this might have
270        been caught by the recursion limit, or it might have run out of stack.
271    
272    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
273        happen only once. PCRE was, however incorrectly giving a compile time error
274        "recursive call could loop indefinitely" because it cannot analyze the
275        pattern in sufficient detail. The compile time test no longer happens when
276        PCRE is compiling a conditional subpattern, but actual runaway loops are
277        now caught at runtime (see 40 above).
278    
279    42. It seems that Perl allows any characters other than a closing parenthesis
280        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
281        has been changed to be the same.
282    
283    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
284        as not to get warnings when autogen.sh is called. Also changed
285        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
286    
287    44. To help people who use pcregrep to scan files containing exceedingly long
288        lines, the following changes have been made:
289    
290        (a) The default value of the buffer size parameter has been increased from
291            8K to 20K. (The actual buffer used is three times this size.)
292    
293        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
294            PCRE is built.
295    
296        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
297            to be set at run time.
298    
299        (d) Numerical values in pcregrep options can be followed by K or M, for
300            example --buffer-size=50K.
301    
302        (e) If a line being scanned overflows pcregrep's buffer, an error is now
303            given and the return code is set to 2.
304    
305    45. Add a pointer to the latest mark to the callout data block.
306    
307    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
308        partial match of an empty string instead of no match. This was specific to
309        the use of ".".
310    
311    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
312        complete match instead of a partial match. This bug was dependent on both
313        the PCRE_UTF8 and PCRE_DOTALL options being set.
314    
315    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
316        starting byte set, because \b was not being ignored.
317    
318    
319    Version 8.12 15-Jan-2011
320    ------------------------
321    
322    1.  Fixed some typos in the markup of the man pages, and wrote a script that
323        checks for such things as part of the documentation building process.
324    
325    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
326        --match-limit and --recursion-limit options (added for 8.11). In
327        particular, this made one of the standard tests fail. (The integer value
328        went into the wrong half of a long int.)
329    
330    3.  If the --colour option was given to pcregrep with -v (invert match), it
331        did strange things, either producing crazy output, or crashing. It should,
332        of course, ignore a request for colour when reporting lines that do not
333        match.
334    
335    4.  Another pcregrep bug caused similar problems if --colour was specified with
336        -M (multiline) and the pattern match finished with a line ending.
337    
338    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
339        matched in multiline mode, the following line was shown as part of the
340        match. This seems wrong, so I have changed it.
341    
342    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
343        the check for further matches in the same line (so they could be coloured)
344        to overrun the end of the current line. If another match was found, it was
345        incorrectly shown (and then shown again when found in the next line).
346    
347    7.  If pcregrep was compiled under Windows, there was a reference to the
348        function pcregrep_exit() before it was defined. I am assuming this was
349        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
350        reported by a user. I've moved the definition above the reference.
351    
352    
353    Version 8.11 10-Dec-2010
354    ------------------------
355    
356    1.  (*THEN) was not working properly if there were untried alternatives prior
357        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
358        backtracked to try for "b" instead of moving to the next alternative branch
359        at the same level (in this case, to look for "c"). The Perl documentation
360        is clear that when (*THEN) is backtracked onto, it goes to the "next
361        alternative in the innermost enclosing group".
362    
363    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
364        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
365        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
366        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
367        (*THEN).
368    
369    3.  If \s appeared in a character class, it removed the VT character from
370        the class, even if it had been included by some previous item, for example
371        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
372        of \s, but is part of the POSIX "space" class.)
373    
374    4.  A partial match never returns an empty string (because you can always
375        match an empty string at the end of the subject); however the checking for
376        an empty string was starting at the "start of match" point. This has been
377        changed to the "earliest inspected character" point, because the returned
378        data for a partial match starts at this character. This means that, for
379        example, /(?<=abc)def/ gives a partial match for the subject "abc"
380        (previously it gave "no match").
381    
382    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
383        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
384        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
385        has an implication that the given string is incomplete (because a partial
386        match is preferred over a full match). For this reason, these items now
387        give a partial match in this situation. [Aside: previously, the one case
388        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
389        match rather than a full match, which was wrong by the old rules, but is
390        now correct.]
391    
392    6.  There was a bug in the handling of #-introduced comments, recognized when
393        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
394        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
395        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
396        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
397        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
398        places in pcre_compile().
399    
400    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
401        comments when looking ahead for named forward references to subpatterns,
402        the only newline sequence it recognized was NL. It now handles newlines
403        according to the set newline convention.
404    
405    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
406        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
407        cater for a lack of strerror(). These oversights have been fixed.
408    
409    9.  Added --match-limit and --recursion-limit to pcregrep.
410    
411    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
412    
413    11. When the -o option was used, pcregrep was setting a return code of 1, even
414        when matches were found, and --line-buffered was not being honoured.
415    
416    12. Added an optional parentheses number to the -o and --only-matching options
417        of pcregrep.
418    
419    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
420        can match an empty string. The code to do it in pcretest and pcredemo
421        needed fixing:
422    
423        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
424            only one byte after an empty string match just before CRLF (this case
425            just got forgotten; "any" and "anycrlf" were OK).
426    
427        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
428            mode when an empty string match preceded an ASCII character followed by
429            a non-ASCII character. (The code for advancing by one character rather
430            than one byte was nonsense.)
431    
432        (c) The pcredemo.c sample program did not have any code at all to handle
433            the cases when CRLF is a valid newline sequence.
434    
435    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
436        as a starting offset was within the subject string. There is now a new
437        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
438        negative or greater than the length of the string. In order to test this,
439        pcretest is extended to allow the setting of negative starting offsets.
440    
441    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
442        starting offset points to the beginning of a UTF-8 character was
443        unnecessarily clumsy. I tidied it up.
444    
445    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
446        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
447    
448    17. Nobody had reported that the --include_dir option, which was added in
449        release 7.7 should have been called --include-dir (hyphen, not underscore)
450        for compatibility with GNU grep. I have changed it to --include-dir, but
451        left --include_dir as an undocumented synonym, and the same for
452        --exclude-dir, though that is not available in GNU grep, at least as of
453        release 2.5.4.
454    
455    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
456        characters from a string of bytes) have been redefined so as not to use
457        loops, in order to improve performance in some environments. At the same
458        time, I abstracted some of the common code into auxiliary macros to save
459        repetition (this should not affect the compiled code).
460    
461    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
462        compile-time error is now given if \c is not followed by an ASCII
463        character, that is, a byte less than 128. (In EBCDIC mode, the code is
464        different, and any byte value is allowed.)
465    
466    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
467        START_OPTIMIZE option, which is now allowed at compile time - but just
468        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
469        to pcregrep and other applications that have no direct access to PCRE
470        options. The new /Y option in pcretest sets this option when calling
471        pcre_compile().
472    
473    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
474        back references. Groups containing recursive back references were forced to
475        be atomic by that change, but in the case of named groups, the amount of
476        memory required was incorrectly computed, leading to "Failed: internal
477        error: code overflow". This has been fixed.
478    
479    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
480        pcretest.c, to avoid build problems in some Borland environments.
481    
482    
483    Version 8.10 25-Jun-2010
484    ------------------------
485    
486    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
487        THEN.
488    
489    2.  (*ACCEPT) was not working when inside an atomic group.
490    
491    3.  Inside a character class, \B is treated as a literal by default, but
492        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
493        causes the error). The code is unchanged, but I tidied the documentation.
494    
495    4.  Inside a character class, PCRE always treated \R and \X as literals,
496        whereas Perl faults them if its -w option is set. I have changed PCRE so
497        that it faults them when PCRE_EXTRA is set.
498    
499    5.  Added support for \N, which always matches any character other than
500        newline. (It is the same as "." when PCRE_DOTALL is not set.)
501    
502    6.  When compiling pcregrep with newer versions of gcc which may have
503        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
504        declared with attribute warn_unused_result" were given. Just casting the
505        result to (void) does not stop the warnings; a more elaborate fudge is
506        needed. I've used a macro to implement this.
507    
508    7.  Minor change to pcretest.c to avoid a compiler warning.
509    
510    8.  Added four artifical Unicode properties to help with an option to make
511        \s etc use properties (see next item). The new properties are: Xan
512        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
513    
514    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
515        use Unicode properties. (*UCP) at the start of a pattern can be used to set
516        this option. Modified pcretest to add /W to test this facility. Added
517        REG_UCP to make it available via the POSIX interface.
518    
519    10. Added --line-buffered to pcregrep.
520    
521    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
522        studied, and the match started with a letter with a code point greater than
523        127 whose first byte was different to the first byte of the other case of
524        the letter, the other case of this starting letter was not recognized
525        (#976).
526    
527    12. If a pattern that was studied started with a repeated Unicode property
528        test, for example, \p{Nd}+, there was the theoretical possibility of
529        setting up an incorrect bitmap of starting bytes, but fortunately it could
530        not have actually happened in practice until change 8 above was made (it
531        added property types that matched character-matching opcodes).
532    
533    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
534        possible starting bytes for non-anchored patterns.
535    
536    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
537        \R, and also a number of cases that involve Unicode properties, both
538        explicit and implicit when PCRE_UCP is set.
539    
540    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
541        input, it could crash or give wrong results if characters with values
542        greater than 0xc0 were present in the subject string. (Detail: it assumed
543        UTF-8 input when processing these items.)
544    
545    16. Added a lot of (int) casts to avoid compiler warnings in systems where
546        size_t is 64-bit (#991).
547    
548    17. Added a check for running out of memory when PCRE is compiled with
549        --disable-stack-for-recursion (#990).
550    
551    18. If the last data line in a file for pcretest does not have a newline on
552        the end, a newline was missing in the output.
553    
554    19. The default pcre_chartables.c file recognizes only ASCII characters (values
555        less than 128) in its various bitmaps. However, there is a facility for
556        generating tables according to the current locale when PCRE is compiled. It
557        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
558        characters, are recognized by isspace() and therefore were getting set in
559        these tables, and indeed these tables seem to approximate to ISO 8859. This
560        caused a problem in UTF-8 mode when pcre_study() was used to create a list
561        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
562        which of course cannot start UTF-8 characters. I have changed the code so
563        that only real ASCII characters (less than 128) and the correct starting
564        bytes for UTF-8 encodings are set for characters greater than 127 when in
565        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
566        altogether.)
567    
568    20. Added the /T option to pcretest so as to be able to run tests with non-
569        standard character tables, thus making it possible to include the tests
570        used for 19 above in the standard set of tests.
571    
572    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
573        reference to a subpattern the other side of a comment that contains an
574        opening parenthesis caused either an internal compiling error, or a
575        reference to the wrong subpattern.
576    
577    
578    Version 8.02 19-Mar-2010
579    ------------------------
580    
581    1.  The Unicode data tables have been updated to Unicode 5.2.0.
582    
583    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
584        configured.
585    
586    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
587        original author of that file, following a query about its status.
588    
589    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
590        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
591    
592    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
593        quantifier applied to a forward-referencing subroutine call, could compile
594        incorrect code or give the error "internal error: previously-checked
595        referenced subpattern not found".
596    
597    6.  Both MS Visual Studio and Symbian OS have problems with initializing
598        variables to point to external functions. For these systems, therefore,
599        pcre_malloc etc. are now initialized to local functions that call the
600        relevant global functions.
601    
602    7.  There were two entries missing in the vectors called coptable and poptable
603        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
604        I've fixed the data, and added a kludgy way of testing at compile time that
605        the lengths are correct (equal to the number of opcodes).
606    
607    8.  Following on from 7, I added a similar kludge to check the length of the
608        eint vector in pcreposix.c.
609    
610    9.  Error texts for pcre_compile() are held as one long string to avoid too
611        much relocation at load time. To find a text, the string is searched,
612        counting zeros. There was no check for running off the end of the string,
613        which could happen if a new error number was added without updating the
614        string.
615    
616    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
617    
618    11. \K was not working if it appeared in an atomic group or in a group that
619        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
620        \K is "not well defined" if used in an assertion. PCRE now accepts it if
621        the assertion is positive, but not if it is negative.
622    
623    12. Change 11 fortuitously reduced the size of the stack frame used in the
624        "match()" function of pcre_exec.c by one pointer. Forthcoming
625        implementation of support for (*MARK) will need an extra pointer on the
626        stack; I have reserved it now, so that the stack frame size does not
627        decrease.
628    
629    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
630        item in branch that calls a recursion is a subroutine call - as in the
631        second branch in the above example - was incorrectly given the compile-
632        time error "recursive call could loop indefinitely" because pcre_compile()
633        was not correctly checking the subroutine for matching a non-empty string.
634    
635    14. The checks for overrunning compiling workspace could trigger after an
636        overrun had occurred. This is a "should never occur" error, but it can be
637        triggered by pathological patterns such as hundreds of nested parentheses.
638        The checks now trigger 100 bytes before the end of the workspace.
639    
640    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
641    
642    
643    Version 8.01 19-Jan-2010
644    ------------------------
645    
646    1.  If a pattern contained a conditional subpattern with only one branch (in
647        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
648        computed the wrong minimum data length (which is of course zero for such
649        subpatterns). This could cause incorrect "no match" results.
650    
651    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
652        the pattern is reset in the first branch, pcre_compile() failed with
653        "internal error: code overflow at offset...". This happened only when
654        the reset was to the original external option setting. (An optimization
655        abstracts leading options settings into an external setting, which was the
656        cause of this.)
657    
658    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
659        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
660        assertion pattern did not match (meaning that the assertion was true), it
661        was incorrectly treated as false if the SKIP had been reached during the
662        matching. This also applied to assertions used as conditions.
663    
664    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
665        assertion subpattern, including such a pattern used as a condition,
666        unpredictable results occurred, instead of the error return
667        PCRE_ERROR_DFA_UITEM.
668    
669    5.  The C++ GlobalReplace function was not working like Perl for the special
670        situation when an empty string is matched. It now does the fancy magic
671        stuff that is necessary.
672    
673    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
674        removed. (These were left over from very, very early versions of PCRE.)
675    
676    7.  Some cosmetic changes to the code to make life easier when compiling it
677        as part of something else:
678    
679        (a) Change DEBUG to PCRE_DEBUG.
680    
681        (b) In pcre_compile(), rename the member of the "branch_chain" structure
682            called "current" as "current_branch", to prevent a collision with the
683            Linux macro when compiled as a kernel module.
684    
685        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
686            prevent a collision with the Linux macro when compiled as a kernel
687            module.
688    
689    8.  In pcre_compile() there are some checks for integer overflows that used to
690        cast potentially large values to (double). This has been changed to that
691        when building, a check for int64_t is made, and if it is found, it is used
692        instead, thus avoiding the use of floating point arithmetic. (There is no
693        other use of FP in PCRE.) If int64_t is not found, the fallback is to
694        double.
695    
696    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
697        2005 (difference between two addresses compared to an unsigned value).
698    
699    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
700        custom one, because of the following reported problem in Windows:
701    
702          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
703              under Win32.
704          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
705              therefore missing the function definition.
706          - The compiler thus generates a "C" signature for the test function.
707          - The linker fails to find the "C" function.
708          - PCRE fails to configure if asked to do so against libbz2.
709    
710    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
711        messages were output:
712    
713          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
714          rerunning libtoolize, to keep the correct libtool macros in-tree.
715          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
716    
717        I have done both of these things.
718    
719    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
720        most of the time, it *can* run out if it is given a pattern that contains a
721        runaway infinite recursion. I updated the discussion in the pcrestack man
722        page.
723    
724    13. Now that we have gone to the x.xx style of version numbers, the minor
725        version may start with zero. Using 08 or 09 is a bad idea because users
726        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
727        interpreted as invalid octal numbers. I've updated the previous comment in
728        configure.ac, and also added a check that gives an error if 08 or 09 are
729        used.
730    
731    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
732        causing partial matching to fail when the end of the subject matched \W
733        in a UTF-8 pattern where \W was quantified with a minimum of 3.
734    
735    15. There were some discrepancies between the declarations in pcre_internal.h
736        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
737        their definitions. The declarations used "const uschar *" and the
738        definitions used USPTR. Even though USPTR is normally defined as "const
739        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
740        reported that: "This difference in casting confuses some C++ compilers, for
741        example, SunCC recognizes above declarations as different functions and
742        generates broken code for hbpcre." I have changed the declarations to use
743        USPTR.
744    
745    16. GNU libtool is named differently on some systems. The autogen.sh script now
746        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
747        (FreeBSD).
748    
749    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
750        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
751        comment: "Figure out how to create a longlong from a string: strtoll and
752        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
753        instance, but it only takes 2 args instead of 3!"
754    
755    18. A subtle bug concerned with back references has been fixed by a change of
756        specification, with a corresponding code fix. A pattern such as
757        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
758        refers, was giving matches when it shouldn't. For example, xa=xaaa would
759        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
760        same bug. Such groups have to be quantified to be useful, or contained
761        inside another quantified group. (If there's no repetition, the reference
762        can never match.) The problem arises because, having left the group and
763        moved on to the rest of the pattern, a later failure that backtracks into
764        the group uses the captured value from the final iteration of the group
765        rather than the correct earlier one. I have fixed this in PCRE by forcing
766        any group that contains a reference to itself to be an atomic group; that
767        is, there cannot be any backtracking into it once it has completed. This is
768        similar to recursive and subroutine calls.
769    
770    
771    Version 8.00 19-Oct-09
772    ----------------------
773    
774    1.  The table for translating pcre_compile() error codes into POSIX error codes
775        was out-of-date, and there was no check on the pcre_compile() error code
776        being within the table. This could lead to an OK return being given in
777        error.
778    
779    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
780        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
781        in a Windows environment.
782    
783    3.  The pcregrep --count option prints the count for each file even when it is
784        zero, as does GNU grep. However, pcregrep was also printing all files when
785        --files-with-matches was added. Now, when both options are given, it prints
786        counts only for those files that have at least one match. (GNU grep just
787        prints the file name in this circumstance, but including the count seems
788        more useful - otherwise, why use --count?) Also ensured that the
789        combination -clh just lists non-zero counts, with no names.
790    
791    4.  The long form of the pcregrep -F option was incorrectly implemented as
792        --fixed_strings instead of --fixed-strings. This is an incompatible change,
793        but it seems right to fix it, and I didn't think it was worth preserving
794        the old behaviour.
795    
796    5.  The command line items --regex=pattern and --regexp=pattern were not
797        recognized by pcregrep, which required --regex pattern or --regexp pattern
798        (with a space rather than an '='). The man page documented the '=' forms,
799        which are compatible with GNU grep; these now work.
800    
801    6.  No libpcreposix.pc file was created for pkg-config; there was just
802        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
803    
804    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
805        when UCP support is not needed, by modifying the Python script that
806        generates it from Unicode data files. This should not matter if the module
807        is correctly used as a library, but I received one complaint about 50K of
808        unwanted data. My guess is that the person linked everything into his
809        program rather than using a library. Anyway, it does no harm.
810    
811    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
812        was a minimum greater than 1 for a wide character in a possessive
813        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
814        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
815        character. Chaos in the form of incorrect output or a compiling loop could
816        result.
817    
818    9.  The restrictions on what a pattern can contain when partial matching is
819        requested for pcre_exec() have been removed. All patterns can now be
820        partially matched by this function. In addition, if there are at least two
821        slots in the offset vector, the offset of the earliest inspected character
822        for the match and the offset of the end of the subject are set in them when
823        PCRE_ERROR_PARTIAL is returned.
824    
825    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
826        synonymous with PCRE_PARTIAL, for backwards compatibility, and
827        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
828        and may be more useful for multi-segment matching.
829    
830    11. Partial matching with pcre_exec() is now more intuitive. A partial match
831        used to be given if ever the end of the subject was reached; now it is
832        given only if matching could not proceed because another character was
833        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
834        string "Z", which now yields "no match" instead of "partial match". In the
835        case of pcre_dfa_exec(), "no match" is given if every matching path for the
836        final character ended with (*FAIL).
837    
838    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
839        if the pattern had a "must contain" character that was already found in the
840        earlier partial match, unless partial matching was again requested. For
841        example, with the pattern /dog.(body)?/, the "must contain" character is
842        "g". If the first part-match was for the string "dog", restarting with
843        "sbody" failed. This bug has been fixed.
844    
845    13. The string returned by pcre_dfa_exec() after a partial match has been
846        changed so that it starts at the first inspected character rather than the
847        first character of the match. This makes a difference only if the pattern
848        starts with a lookbehind assertion or \b or \B (\K is not supported by
849        pcre_dfa_exec()). It's an incompatible change, but it makes the two
850        matching functions compatible, and I think it's the right thing to do.
851    
852    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
853        so that the demonstration program is easily available in environments where
854        PCRE has not been installed from source.
855    
856    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
857        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
858        library.
859    
860    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
861        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
862        is not the first non-POSIX option to be added. Clearly some people find
863        these options useful.
864    
865    17. If a caller to the POSIX matching function regexec() passes a non-zero
866        value for nmatch with a NULL value for pmatch, the value of
867        nmatch is forced to zero.
868    
869    18. RunGrepTest did not have a test for the availability of the -u option of
870        the diff command, as RunTest does. It now checks in the same way as
871        RunTest, and also checks for the -b option.
872    
873    19. If an odd number of negated classes containing just a single character
874        interposed, within parentheses, between a forward reference to a named
875        subpattern and the definition of the subpattern, compilation crashed with
876        an internal error, complaining that it could not find the referenced
877        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
878        [The bug was that it was starting one character too far in when skipping
879        over the character class, thus treating the ] as data rather than
880        terminating the class. This meant it could skip too much.]
881    
882    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
883        /g option in pcretest when the pattern contains \K, which makes it possible
884        to have an empty string match not at the start, even when the pattern is
885        anchored. Updated pcretest and pcredemo to use this option.
886    
887    21. If the maximum number of capturing subpatterns in a recursion was greater
888        than the maximum at the outer level, the higher number was returned, but
889        with unset values at the outer level. The correct (outer level) value is
890        now given.
891    
892    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
893        PCRE did not set those parentheses (unlike Perl). I have now found a way to
894        make it do so. The string so far is captured, making this feature
895        compatible with Perl.
896    
897    23. The tests have been re-organized, adding tests 11 and 12, to make it
898        possible to check the Perl 5.10 features against Perl 5.10.
899    
900    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
901        pattern matches a fixed length string. PCRE did not allow this; now it
902        does. Neither allows recursion.
903    
904    25. I finally figured out how to implement a request to provide the minimum
905        length of subject string that was needed in order to match a given pattern.
906        (It was back references and recursion that I had previously got hung up
907        on.) This code has now been added to pcre_study(); it finds a lower bound
908        to the length of subject needed. It is not necessarily the greatest lower
909        bound, but using it to avoid searching strings that are too short does give
910        some useful speed-ups. The value is available to calling programs via
911        pcre_fullinfo().
912    
913    26. While implementing 25, I discovered to my embarrassment that pcretest had
914        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
915        study optimizations had never been tested with that matching function.
916        Oops. What is worse, even when it was passed study data, there was a bug in
917        pcre_dfa_exec() that meant it never actually used it. Double oops. There
918        were also very few tests of studied patterns with pcre_dfa_exec().
919    
920    27. If (?| is used to create subpatterns with duplicate numbers, they are now
921        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
922        on the other side of the coin, they are no longer allowed to have different
923        names, because these cannot be distinguished in PCRE, and this has caused
924        confusion. (This is a difference from Perl.)
925    
926    28. When duplicate subpattern names are present (necessarily with different
927        numbers, as required by 27 above), and a test is made by name in a
928        conditional pattern, either for a subpattern having been matched, or for
929        recursion in such a pattern, all the associated numbered subpatterns are
930        tested, and the overall condition is true if the condition is true for any
931        one of them. This is the way Perl works, and is also more like the way
932        testing by number works.
933    
934    
935    Version 7.9 11-Apr-09
936    ---------------------
937    
938    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
939        (pcretest), all targets were linked against these libraries. This included
940        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
941        libraries. This caused unwanted dependencies to be created. This problem
942        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
943        pcretest is linked with readline.
944    
945    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
946        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
947        moved outside it again, because FALSE and TRUE are already defined in AIX,
948        but BOOL is not.
949    
950    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
951        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
952    
953    4.  The pcregrep documentation said spaces were inserted as well as colons (or
954        hyphens) following file names and line numbers when outputting matching
955        lines. This is not true; no spaces are inserted. I have also clarified the
956        wording for the --colour (or --color) option.
957    
958    5.  In pcregrep, when --colour was used with -o, the list of matching strings
959        was not coloured; this is different to GNU grep, so I have changed it to be
960        the same.
961    
962    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
963        each matching line was coloured. Now it goes on to look for further matches
964        of any of the test patterns, which is the same behaviour as GNU grep.
965    
966    7.  A pattern that could match an empty string could cause pcregrep to loop; it
967        doesn't make sense to accept an empty string match in pcregrep, so I have
968        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
969        seems to be how GNU grep behaves.
970    
971    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
972        start or after a newline", because the conditional assertion was not being
973        correctly handled. The rule now is that both the assertion and what follows
974        in the first alternative must satisfy the test.
975    
976    9.  If auto-callout was enabled in a pattern with a conditional group whose
977        condition was an assertion, PCRE could crash during matching, both with
978        pcre_exec() and pcre_dfa_exec().
979    
980    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
981        used for matching.
982    
983    11. Unicode property support in character classes was not working for
984        characters (bytes) greater than 127 when not in UTF-8 mode.
985    
986    12. Added the -M command line option to pcretest.
987    
988    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
989    
990    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
991    
992    16. Added comments and documentation about mis-use of no_arg in the C++
993        wrapper.
994    
995    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
996        from Martin Jerabek that uses macro names for all relevant character and
997        string constants.
998    
999    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1000        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1001        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1002        these, but not everybody uses configure.
1003    
1004    19. A conditional group that had only one branch was not being correctly
1005        recognized as an item that could match an empty string. This meant that an
1006        enclosing group might also not be so recognized, causing infinite looping
1007        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1008        with the subject "ab", where knowledge that the repeated group can match
1009        nothing is needed in order to break the loop.
1010    
1011    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1012        exec(), but without supplying a callout function, matching went wrong.
1013    
1014    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1015        leak if the size of the offset vector was greater than 30. When the vector
1016        is smaller, the saved offsets during recursion go onto a local stack
1017        vector, but for larger vectors malloc() is used. It was failing to free
1018        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1019        error, in fact).
1020    
1021    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1022        heapframe that is used only when UTF-8 support is enabled. This caused no
1023        problem, but was untidy.
1024    
1025    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1026        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1027        included within another project.
1028    
1029    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1030        slightly modified by me:
1031    
1032          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1033              not building pcregrep.
1034    
1035          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1036              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1037    
1038    25. Forward references, both numeric and by name, in patterns that made use of
1039        duplicate group numbers, could behave incorrectly or give incorrect errors,
1040        because when scanning forward to find the reference group, PCRE was not
1041        taking into account the duplicate group numbers. A pattern such as
1042        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1043    
1044    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1045        the feature of a custom pointer more persuasive (as requested by a user).
1046    
1047    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1048        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1049        pre-defined.
1050    
1051    28. Added support for (*UTF8) at the start of a pattern.
1052    
1053    29. Arrange for flags added by the "release type" setting in CMake to be shown
1054        in the configuration summary.
1055    
1056    
1057    Version 7.8 05-Sep-08
1058    ---------------------
1059    
1060    1.  Replaced UCP searching code with optimized version as implemented for Ad
1061        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1062        stage table and inline lookup instead of a function, giving speed ups of 2
1063        to 5 times on some simple patterns that I tested. Permission was given to
1064        distribute the MultiStage2.py script that generates the tables (it's not in
1065        the tarball, but is in the Subversion repository).
1066    
1067    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1068        scripts.
1069    
1070    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1071        a group with a zero qualifier. The result of the study could be incorrect,
1072        or the function might crash, depending on the pattern.
1073    
1074    4.  Caseless matching was not working for non-ASCII characters in back
1075        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1076        It now works when Unicode Property Support is available.
1077    
1078    5.  In pcretest, an escape such as \x{de} in the data was always generating
1079        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1080        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1081        truncation.
1082    
1083    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1084    
1085    7.  Added two (int) casts to pcregrep when printing the difference of two
1086        pointers, in case they are 64-bit values.
1087    
1088    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1089        test 2 if it fails.
1090    
1091    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1092        and a #define of that name to empty if it is not externally set. This is to
1093        allow users of MSVC to set it if necessary.
1094    
1095    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1096        the convenience functions in the pcre_get.c source file.
1097    
1098    11. An option change at the start of a pattern that had top-level alternatives
1099        could cause overwriting and/or a crash. This command provoked a crash in
1100        some environments:
1101    
1102          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1103    
1104        This potential security problem was recorded as CVE-2008-2371.
1105    
1106    12. For a pattern where the match had to start at the beginning or immediately
1107        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1108        pcre_dfa_exec() could read past the end of the passed subject if there was
1109        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1110        pcretest so that it places the subject at the end of its malloc-ed buffer.
1111    
1112    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1113        exec() might read past the end of the data buffer in UTF-8 mode.
1114    
1115    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1116        the data contained the byte 0x85 as part of a UTF-8 character within its
1117        first line. This applied both to normal and DFA matching.
1118    
1119    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1120        /^[^d]*?$/8 failed to match "abc".
1121    
1122    16. Added a missing copyright notice to pcrecpp_internal.h.
1123    
1124    17. Make it more clear in the documentation that values returned from
1125        pcre_exec() in ovector are byte offsets, not character counts.
1126    
1127    18. Tidied a few places to stop certain compilers from issuing warnings.
1128    
1129    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1130        supplied by Stefan Weber. I made a further small update for 7.8 because
1131        there is a change of source arrangements: the pcre_searchfuncs.c module is
1132        replaced by pcre_ucd.c.
1133    
1134    
1135    Version 7.7 07-May-08
1136    ---------------------
1137    
1138    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1139        a string to a long long, pretend we don't even have a long long." This is
1140        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1141    
1142    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1143        pre-7.6 versions, which defined a global no_arg variable instead of putting
1144        it in the RE class. (See also #8 below.)
1145    
1146    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1147        Lopes.
1148    
1149    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1150    
1151        (1) The include/exclude patterns were being applied to the whole pathnames
1152            of files, instead of just to the final components.
1153    
1154        (2) If there was more than one level of directory, the subdirectories were
1155            skipped unless they satisfied the include/exclude conditions. This is
1156            inconsistent with GNU grep (and could even be seen as contrary to the
1157            pcregrep specification - which I improved to make it absolutely clear).
1158            The action now is always to scan all levels of directory, and just
1159            apply the include/exclude patterns to regular files.
1160    
1161    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1162        --exclude_dir in the tests to avoid scanning .svn directories.
1163    
1164    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1165        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1166        doesn't support NULs in patterns.
1167    
1168    7.  Added some missing "const"s to declarations of static tables in
1169        pcre_compile.c and pcre_dfa_exec.c.
1170    
1171    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1172        caused by fix #2  above. (Subsequently also a second patch to fix the
1173        first patch. And a third patch - this was a messy problem.)
1174    
1175    9.  Applied Craig's patch to remove the use of push_back().
1176    
1177    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1178        matching function regexec().
1179    
1180    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1181        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1182        references. PCRE supports relative numbers with this syntax (I don't think
1183        Oniguruma does).
1184    
1185    12. Previously, a group with a zero repeat such as (...){0} was completely
1186        omitted from the compiled regex. However, this means that if the group
1187        was called as a subroutine from elsewhere in the pattern, things went wrong
1188        (an internal error was given). Such groups are now left in the compiled
1189        pattern, with a new opcode that causes them to be skipped at execution
1190        time.
1191    
1192    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1193        to the way PCRE behaves:
1194    
1195        (a) A lone ] character is dis-allowed (Perl treats it as data).
1196    
1197        (b) A back reference to an unmatched subpattern matches an empty string
1198            (Perl fails the current match path).
1199    
1200        (c) A data ] in a character class must be notated as \] because if the
1201            first data character in a class is ], it defines an empty class. (In
1202            Perl it is not possible to have an empty class.) The empty class []
1203            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1204            The negative empty class [^] matches any one character, independently
1205            of the DOTALL setting.
1206    
1207    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1208        non-existent subpattern following a character class starting with ']' and
1209        containing () gave an internal compiling error instead of "reference to
1210        non-existent subpattern". Fortunately, when the pattern did exist, the
1211        compiled code was correct. (When scanning forwards to check for the
1212        existencd of the subpattern, it was treating the data ']' as terminating
1213        the class, so got the count wrong. When actually compiling, the reference
1214        was subsequently set up correctly.)
1215    
1216    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1217        it was being rejected as not supported by pcre_dfa_exec(), even though
1218        other assertions are supported. I have made pcre_dfa_exec() support
1219        (*FAIL).
1220    
1221    16. The implementation of 13c above involved the invention of a new opcode,
1222        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1223        cannot be changed at match time, I realized I could make a small
1224        improvement to matching performance by compiling OP_ALLANY instead of
1225        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1226        on the OP_ANY path.
1227    
1228    17. Compiling pcretest on Windows with readline support failed without the
1229        following two fixes: (1) Make the unistd.h include conditional on
1230        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1231    
1232    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1233        ncurses library to be included for pcretest when ReadLine support is
1234        requested, but also to allow for it to be overridden. This patch came from
1235        Daniel Bergström.
1236    
1237    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1238        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1239        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1240        spotting this.
1241    
1242    
1243    Version 7.6 28-Jan-08
1244    ---------------------
1245    
1246    1.  A character class containing a very large number of characters with
1247        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1248        overflow.
1249    
1250    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1251        HAVE_LONG_LONG is not defined.
1252    
1253    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1254        bring them up to date and include new features. This patch includes:
1255    
1256        - Fixed PH's badly added libz and libbz2 support.
1257        - Fixed a problem with static linking.
1258        - Added pcredemo. [But later removed - see 7 below.]
1259        - Fixed dftables problem and added an option.
1260        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1261            HAVE_LONG_LONG.
1262        - Added readline support for pcretest.
1263        - Added an listing of the option settings after cmake has run.
1264    
1265    4.  A user submitted a patch to Makefile that makes it easy to create
1266        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1267        Makefile.am that cause it to include this special target, without
1268        affecting anything else. Note that the same mingw target plus all
1269        the other distribution libraries and programs are now supported
1270        when configuring with CMake (see 6 below) instead of with
1271        Configure/Make.
1272    
1273    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1274        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1275        exported in the Windows port". It has not yet been confirmed that the patch
1276        solves the problem, but it does no harm.
1277    
1278    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1279        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1280        with CMake, and also correct the comment about stack recursion.
1281    
1282    7.  Remove the automatic building of pcredemo from the ./configure system and
1283        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1284        of a program that users should build themselves after PCRE is installed, so
1285        building it automatically is not really right. What is more, it gave
1286        trouble in some build environments.
1287    
1288    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1289    
1290    
1291    Version 7.5 10-Jan-08
1292    ---------------------
1293    
1294    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1295        values in parens when parsing an RE using the C++ wrapper."
1296    
1297    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1298        Characters greater than 255 were excluded from the class instead of being
1299        included.
1300    
1301    3.  The same bug as (2) above applied to negated POSIX classes such as
1302        [:^space:].
1303    
1304    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1305        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1306        I have changed it.
1307    
1308    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1309        first named subpattern) and a construct such as (?&a) would reference the
1310        first named subpattern whose name started with "a" (in other words, the
1311        length check was missing). Both these problems are fixed. "Subpattern name
1312        expected" is now given for (?&) (a zero-length name), and this patch also
1313        makes it give the same error for \k'' (previously it complained that that
1314        was a reference to a non-existent subpattern).
1315    
1316    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1317        this is right because (?- can be followed by option settings as well as by
1318        digits. I have, however, made the messages clearer.
1319    
1320    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1321        than the number used in the conditional) now cause a compile-time error.
1322        This is actually not compatible with Perl, which accepts such patterns, but
1323        treats the conditional as always being FALSE (as PCRE used to), but it
1324        seems to me that giving a diagnostic is better.
1325    
1326    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1327        and messages.
1328    
1329    9.  Fix two occurrences of "backslash" in comments that should have been
1330        "backspace".
1331    
1332    10. Remove two redundant lines of code that can never be obeyed (their function
1333        was moved elsewhere).
1334    
1335    11. The program that makes PCRE's Unicode character property table had a bug
1336        which caused it to generate incorrect table entries for sequences of
1337        characters that have the same character type, but are in different scripts.
1338        It amalgamated them into a single range, with the script of the first of
1339        them. In other words, some characters were in the wrong script. There were
1340        thirteen such cases, affecting characters in the following ranges:
1341    
1342          U+002b0 - U+002c1
1343          U+0060c - U+0060d
1344          U+0061e - U+00612
1345          U+0064b - U+0065e
1346          U+0074d - U+0076d
1347          U+01800 - U+01805
1348          U+01d00 - U+01d77
1349          U+01d9b - U+01dbf
1350          U+0200b - U+0200f
1351          U+030fc - U+030fe
1352          U+03260 - U+0327f
1353          U+0fb46 - U+0fbb1
1354          U+10450 - U+1049d
1355    
1356    12. The -o option (show only the matching part of a line) for pcregrep was not
1357        compatible with GNU grep in that, if there was more than one match in a
1358        line, it showed only the first of them. It now behaves in the same way as
1359        GNU grep.
1360    
1361    13. If the -o and -v options were combined for pcregrep, it printed a blank
1362        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1363        does the same. The return code can be used to tell if there were any
1364        non-matching lines.
1365    
1366    14. Added --file-offsets and --line-offsets to pcregrep.
1367    
1368    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1369        infinitely looping recursion. The bug was that positive lookaheads were not
1370        being skipped when checking for a possible empty match (negative lookaheads
1371        and both kinds of lookbehind were skipped).
1372    
1373    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1374        inclusion of <windows.h> to before rather than after the definition of
1375        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1376    
1377    17. Specifying a possessive quantifier with a specific limit for a Unicode
1378        character property caused pcre_compile() to compile bad code, which led at
1379        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1380        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1381        caused the error; without that there was no problem.
1382    
1383    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1384    
1385    19. Added --enable-pcretest-libreadline.
1386    
1387    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1388        RE::GlobalReplace(). As a result, the number of replacements returned was
1389        double what it should be. I removed one of the increments, but Craig sent a
1390        later patch that removed the other one (the right fix) and added unit tests
1391        that check the return values (which was not done before).
1392    
1393    21. Several CMake things:
1394    
1395        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1396            the names libpcre and libpcreposix, not just pcre and pcreposix.
1397    
1398        (2) The above change means that pcretest and pcregrep are now correctly
1399            linked with the newly-built libraries, not previously installed ones.
1400    
1401        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1402    
1403    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1404        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1405        UTF-8 newline character). The key issue is that the pattern starts .*;
1406        this means that the match must be either at the beginning, or after a
1407        newline. The bug was in the code for advancing after a failed match and
1408        checking that the new position followed a newline. It was not taking
1409        account of UTF-8 characters correctly.
1410    
1411    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1412        character classes. PCRE was not treating the sequence [:...:] as a
1413        character class unless the ... were all letters. Perl, however, seems to
1414        allow any characters between [: and :], though of course it rejects as
1415        unknown any "names" that contain non-letters, because all the known class
1416        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1417        for example, whereas PCRE did not - it did not recognize a POSIX character
1418        class. This seemed a bit dangerous, so the code has been changed to be
1419        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1420        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1421        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1422        Perl does, and where it didn't before.
1423    
1424    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1425        Windows environments %n is disabled by default.
1426    
1427    
1428    Version 7.4 21-Sep-07
1429    ---------------------
1430    
1431    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1432        means that a class such as [\s] counted as "explicit reference to CR or
1433        LF". That isn't really right - the whole point of the change was to try to
1434        help when there was an actual mention of one of the two characters. So now
1435        the change happens only if \r or \n (or a literal CR or LF) character is
1436        encountered.
1437    
1438    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1439        of both had grown to the point where there were only 3 bits left.
1440        Fortunately, there was spare space in the data structure, and so I have
1441        moved the internal flags into a new 16-bit field to free up more option
1442        bits.
1443    
1444    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1445        but did not set the internal JCHANGED flag - either of these is enough to
1446        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1447        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1448        start sets both bits.
1449    
1450    4.  Added options (at build time, compile time, exec time) to change \R from
1451        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1452    
1453    5.  doc/pcresyntax.html was missing from the distribution.
1454    
1455    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1456        compatibility, even though it is no longer used.
1457    
1458    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1459        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1460        windows.h file is present (where different names are used). [This was
1461        reversed later after testing - see 16 below.]
1462    
1463    8.  Changed all #include <config.h> to #include "config.h". There were also
1464        some further <pcre.h> cases that I changed to "pcre.h".
1465    
1466    9.  When pcregrep was used with the --colour option, it missed the line ending
1467        sequence off the lines that it output.
1468    
1469    10. It was pointed out to me that arrays of string pointers cause lots of
1470        relocations when a shared library is dynamically loaded. A technique of
1471        using a single long string with a table of offsets can drastically reduce
1472        these. I have refactored PCRE in four places to do this. The result is
1473        dramatic:
1474    
1475          Originally:                          290
1476          After changing UCP table:            187
1477          After changing error message table:   43
1478          After changing table of "verbs"       36
1479          After changing table of Posix names   22
1480    
1481        Thanks to the folks working on Gregex for glib for this insight.
1482    
1483    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1484        unicode-properties was also set.
1485    
1486    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1487    
1488    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1489        checked only for CRLF.
1490    
1491    14. Added casts to pcretest.c to avoid compiler warnings.
1492    
1493    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1494    
1495    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1496        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1497        entirely. This removes changes made in 7 above.
1498    
1499    17. The CMake files have been updated, and there is now more information about
1500        building with CMake in the NON-UNIX-USE document.
1501    
1502    
1503    Version 7.3 28-Aug-07
1504    ---------------------
1505    
1506     1. In the rejigging of the build system that eventually resulted in 7.1, the
1507        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1508        brackets there is not right, since it causes compilers to look for an
1509        installed pcre.h, not the version that is in the source that is being
1510        compiled (which of course may be different). I have changed it back to:
1511    
1512          #include "pcre.h"
1513    
1514        I have a vague recollection that the change was concerned with compiling in
1515        different directories, but in the new build system, that is taken care of
1516        by the VPATH setting the Makefile.
1517    
1518     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1519        when the subject happened to end in the byte 0x85 (e.g. if the last
1520        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1521        characters but of course it shouldn't be taken as a newline when it is part
1522        of another character. The bug was that, for an unlimited repeat of . in
1523        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1524        characters when looking for a newline.
1525    
1526     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1527    
1528     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1529        in debug output.
1530    
1531     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1532        long printing in the pcrecpp unittest when running under MinGW.
1533    
1534     6. ESC_K was left out of the EBCDIC table.
1535    
1536     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1537        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1538        limit also applies to "virtual nesting" when a pattern is recursive, and in
1539        this case 1000 isn't so big. I have been able to remove this limit at the
1540        expense of backing off one optimization in certain circumstances. Normally,
1541        when pcre_exec() would call its internal match() function recursively and
1542        immediately return the result unconditionally, it uses a "tail recursion"
1543        feature to save stack. However, when a subpattern that can match an empty
1544        string has an unlimited repetition quantifier, it no longer makes this
1545        optimization. That gives it a stack frame in which to save the data for
1546        checking that an empty string has been matched. Previously this was taken
1547        from the 1000-entry workspace that had been reserved. So now there is no
1548        explicit limit, but more stack is used.
1549    
1550     8. Applied Daniel's patches to solve problems with the import/export magic
1551        syntax that is required for Windows, and which was going wrong for the
1552        pcreposix and pcrecpp parts of the library. These were overlooked when this
1553        problem was solved for the main library.
1554    
1555     9. There were some crude static tests to avoid integer overflow when computing
1556        the size of patterns that contain repeated groups with explicit upper
1557        limits. As the maximum quantifier is 65535, the maximum group length was
1558        set at 30,000 so that the product of these two numbers did not overflow a
1559        32-bit integer. However, it turns out that people want to use groups that
1560        are longer than 30,000 bytes (though not repeat them that many times).
1561        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1562        made it possible to implement the integer overflow checks in a much more
1563        dynamic way, which I have now done. The artificial limitation on group
1564        length has been removed - we now have only the limit on the total length of
1565        the compiled pattern, which depends on the LINK_SIZE setting.
1566    
1567    10. Fixed a bug in the documentation for get/copy named substring when
1568        duplicate names are permitted. If none of the named substrings are set, the
1569        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1570        empty string.
1571    
1572    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1573        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1574        because the ] is interpreted as the first data character and the
1575        terminating ] is not found. PCRE has been made compatible with Perl in this
1576        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1577        cause memory overwriting.
1578    
1579    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1580        string has been matched (to stop an infinite loop). It was not recognizing
1581        a conditional subpattern that could match an empty string if that
1582        subpattern was within another subpattern. For example, it looped when
1583        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1584        condition was not nested. This bug has been fixed.
1585    
1586    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1587        past the start of the subject in the presence of bytes with the top bit
1588        set, for example "\x8aBCD".
1589    
1590    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1591        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1592    
1593    14. Optimized (?!) to (*FAIL).
1594    
1595    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1596        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1597        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1598        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1599        does: it's just the validity check that is more restrictive.
1600    
1601    16. Inserted checks for integer overflows during escape sequence (backslash)
1602        processing, and also fixed erroneous offset values for syntax errors during
1603        backslash processing.
1604    
1605    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1606        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1607    
1608    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1609        caused an overrun.
1610    
1611    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1612        something other than just ASCII characters) inside a group that had an
1613        unlimited repeat caused a loop at compile time (while checking to see
1614        whether the group could match an empty string).
1615    
1616    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1617        [\P{Any}] did so. (Error in the code for printing property names.)
1618    
1619    21. An orphan \E inside a character class could cause a crash.
1620    
1621    22. A repeated capturing bracket such as (A)? could cause a wild memory
1622        reference during compilation.
1623    
1624    23. There are several functions in pcre_compile() that scan along a compiled
1625        expression for various reasons (e.g. to see if it's fixed length for look
1626        behind). There were bugs in these functions when a repeated \p or \P was
1627        present in the pattern. These operators have additional parameters compared
1628        with \d, etc, and these were not being taken into account when moving along
1629        the compiled data. Specifically:
1630    
1631        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1632            length.
1633    
1634        (b) An item such as \pL+ within a repeated group could cause crashes or
1635            loops.
1636    
1637        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1638            "reference to non-existent subpattern" error.
1639    
1640        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1641    
1642    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1643        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1644    
1645    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1646    
1647    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1648        character were causing crashes (broken optimization).
1649    
1650    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1651        \p or \P) caused a compile-time loop.
1652    
1653    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1654        break. For example, the unstudied pattern [\r\n]A does not match the string
1655        "\r\nA" because change 7.0/46 below moves the current point on by two
1656        characters after failing to match at the start. However, the pattern \nA
1657        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1658        the same is true. There doesn't seem any very clean way out of this, but
1659        what I have chosen to do makes the common cases work: PCRE now takes note
1660        of whether there can be an explicit match for \r or \n anywhere in the
1661        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1662        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1663        pattern has explicit CR or LF references.
1664    
1665    29. Added (*CR) etc for changing newline setting at start of pattern.
1666    
1667    
1668    Version 7.2 19-Jun-07
1669    ---------------------
1670    
1671     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1672        which is apparently normally available under Windows.
1673    
1674     2. Re-jig the pcregrep tests with different newline settings in an attempt
1675        to make them independent of the local environment's newline setting.
1676    
1677     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1678    
1679     4. Some of the "internals" tests were previously cut out when the link size
1680        was not 2, because the output contained actual offsets. The recent new
1681        "Z" feature of pcretest means that these can be cut out, making the tests
1682        usable with all link sizes.
1683    
1684     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1685        stack recursion. This gives a massive performance boost under BSD, but just
1686        a small improvement under Linux. However, it saves one field in the frame
1687        in all cases.
1688    
1689     6. Added more features from the forthcoming Perl 5.10:
1690    
1691        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1692            recursion call. It refers to the nth most recently opened parentheses.
1693    
1694        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1695            to be opened parentheses.
1696    
1697        (c) Conditions that refer to capturing parentheses can be specified
1698            relatively, for example, (?(-2)... or (?(+3)...
1699    
1700        (d) \K resets the start of the current match so that everything before
1701            is not part of it.
1702    
1703        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1704    
1705        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1706            reference syntax.
1707    
1708        (g) (?| introduces a group in which the numbering of parentheses in each
1709            alternative starts with the same number.
1710    
1711        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1712    
1713     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1714        PCRE_INFO_JCHANGED.
1715    
1716     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1717        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1718        for detecting groups that can match an empty string.
1719    
1720     9. A pattern with a very large number of alternatives (more than several
1721        hundred) was running out of internal workspace during the pre-compile
1722        phase, where pcre_compile() figures out how much memory will be needed. A
1723        bit of new cunning has reduced the workspace needed for groups with
1724        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1725        workspace instead of running out of the 4096 that are available.
1726    
1727    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1728    
1729    11. Applied patch from Google to remove an optimization that didn't quite work.
1730        The report of the bug said:
1731    
1732          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1733          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1734          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1735    
1736    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1737        it matched the wrong number of bytes.
1738    
1739    
1740    Version 7.1 24-Apr-07
1741    ---------------------
1742    
1743     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1744        that is more "standard", making use of automake and other Autotools. There
1745        is some re-arrangement of the files and adjustment of comments consequent
1746        on this.
1747    
1748     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1749        for recursive directory scanning broke on some systems because the files
1750        are not scanned in any specific order and on different systems the order
1751        was different. A call to "sort" has been inserted into RunGrepTest for the
1752        approprate test as a short-term fix. In the longer term there may be an
1753        alternative.
1754    
1755     3. I had an email from Eric Raymond about problems translating some of PCRE's
1756        man pages to HTML (despite the fact that I distribute HTML pages, some
1757        people do their own conversions for various reasons). The problems
1758        concerned the use of low-level troff macros .br and .in. I have therefore
1759        removed all such uses from the man pages (some were redundant, some could
1760        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1761        HTML has been updated to handle .nf/.fi and to complain if it encounters
1762        .br or .in.
1763    
1764     4. Updated comments in configure.ac that get placed in config.h.in and also
1765        arranged for config.h to be included in the distribution, with the name
1766        config.h.generic, for the benefit of those who have to compile without
1767        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1768    
1769     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1770        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1771        makevp.bat for the current PCRE, using the additional files
1772        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1773    
1774     6. A Windows user reported a minor discrepancy with test 2, which turned out
1775        to be caused by a trailing space on an input line that had got lost in his
1776        copy. The trailing space was an accident, so I've just removed it.
1777    
1778     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1779        that is needed.
1780    
1781     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1782        as "const" (a) because they are and (b) because it helps the PHP
1783        maintainers who have recently made a script to detect big data structures
1784        in the php code that should be moved to the .rodata section. I remembered
1785        to update Builducptable as well, so it won't revert if ucptable.h is ever
1786        re-created.
1787    
1788     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1789        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1790        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1791        support is not required. This saves 1.5-2K of code, which is important in
1792        some applications.
1793    
1794        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1795        so as not to refer to the tables, even though these functions will never be
1796        called when UTF-8 support is disabled. Otherwise there are problems with a
1797        shared library.
1798    
1799    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1800    
1801        (a) It was defining its arguments as char * instead of void *.
1802    
1803        (b) It was assuming that all moves were upwards in memory; this was true
1804            a long time ago when I wrote it, but is no longer the case.
1805    
1806        The emulated memove() is provided for those environments that have neither
1807        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1808        is clearly not the case, as these two bugs were recently reported.
1809    
1810    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1811        and Detrail to create the HTML documentation, the .txt form of the man
1812        pages, and it removes trailing spaces from listed files. It also creates
1813        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1814        case, it wraps all the #defines with #ifndefs. This script should be run
1815        before "make dist".
1816    
1817    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1818        with Unicode property support.
1819    
1820        (a) For a maximizing quantifier, if the two different cases of the
1821            character were of different lengths in their UTF-8 codings (there are
1822            some cases like this - I found 11), and the matching function had to
1823            back up over a mixture of the two cases, it incorrectly assumed they
1824            were both the same length.
1825    
1826        (b) When PCRE was configured to use the heap rather than the stack for
1827            recursion during matching, it was not correctly preserving the data for
1828            the other case of a UTF-8 character when checking ahead for a match
1829            while processing a minimizing repeat. If the check also involved
1830            matching a wide character, but failed, corruption could cause an
1831            erroneous result when trying to check for a repeat of the original
1832            character.
1833    
1834    13. Some tidying changes to the testing mechanism:
1835    
1836        (a) The RunTest script now detects the internal link size and whether there
1837            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1838            values substituted by "configure". (The RunGrepTest script already did
1839            this for UTF-8.) The configure.ac script no longer substitutes the
1840            relevant variables.
1841    
1842        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1843            with length and offset values. This means that the output is different
1844            for different internal link sizes. Test 2 is skipped for link sizes
1845            other than 2 because of this, bypassing the problem. Unfortunately,
1846            there was also a test in test 3 (the locale tests) that used /B and
1847            failed for link sizes other than 2. Rather than cut the whole test out,
1848            I have added a new /Z option to pcretest that replaces the length and
1849            offset values with spaces. This is now used to make test 3 independent
1850            of link size. (Test 2 will be tidied up later.)
1851    
1852    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1853        segmentation fault instead of returning the appropriate error message.
1854    
1855    15. In multiline mode when the newline sequence was set to "any", the pattern
1856        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1857        This doesn't seem right; it now treats the CRLF combination as the line
1858        ending, and so does not match in that case. It's only a pattern such as ^$
1859        that would hit this one: something like ^ABC$ would have failed after \r
1860        and then tried again after \r\n.
1861    
1862    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1863        in an attempt to make files that differ only in their line terminators
1864        compare equal. This works on Linux.
1865    
1866    17. Under certain error circumstances pcregrep might try to free random memory
1867        as it exited. This is now fixed, thanks to valgrind.
1868    
1869    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1870        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1871        was because its rules for how to advance for /g after matching an empty
1872        string at the end of a line did not allow for this case. They now check for
1873        it specially.
1874    
1875    20. pcretest is supposed to handle patterns and data of any length, by
1876        extending its buffers when necessary. It was getting this wrong when the
1877        buffer for a data line had to be extended.
1878    
1879    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1880        CRLF as a newline sequence.
1881    
1882    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1883        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1884        I have nevertheless tidied it up.
1885    
1886    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1887    
1888    24. Added a man page for pcre-config.
1889    
1890    
1891    Version 7.0 19-Dec-06
1892    ---------------------
1893    
1894     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1895        moving to gcc 4.1.1.
1896    
1897     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1898        sys/time.h, which is documented as needed for this function. It doesn't
1899        seem to matter on Linux, but it showed up on some releases of OS X.
1900    
1901     3. It seems that there are systems where bytes whose values are greater than
1902        127 match isprint() in the "C" locale. The "C" locale should be the
1903        default when a C program starts up. In most systems, only ASCII printing
1904        characters match isprint(). This difference caused the output from pcretest
1905        to vary, making some of the tests fail. I have changed pcretest so that:
1906    
1907        (a) When it is outputting text in the compiled version of a pattern, bytes
1908            other than 32-126 are always shown as hex escapes.
1909    
1910        (b) When it is outputting text that is a matched part of a subject string,
1911            it does the same, unless a different locale has been set for the match
1912            (using the /L modifier). In this case, it uses isprint() to decide.
1913    
1914     4. Fixed a major bug that caused incorrect computation of the amount of memory
1915        required for a compiled pattern when options that changed within the
1916        pattern affected the logic of the preliminary scan that determines the
1917        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1918        that the computed length was too small. The symptoms of this bug were
1919        either the PCRE error "internal error: code overflow" from pcre_compile(),
1920        or a glibc crash with a message such as "pcretest: free(): invalid next
1921        size (fast)". Examples of patterns that provoked this bug (shown in
1922        pcretest format) are:
1923    
1924          /(?-x: )/x
1925          /(?x)(?-x: \s*#\s*)/
1926          /((?i)[\x{c0}])/8
1927          /(?i:[\x{c0}])/8
1928    
1929        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1930        is now done differently.
1931    
1932     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1933        wrapper classes; (b) implement a new function in the C++ scanner that is
1934        more efficient than the old way of doing things because it avoids levels of
1935        recursion in the regex matching; (c) add a paragraph to the documentation
1936        for the FullMatch() function.
1937    
1938     6. The escape sequence \n was being treated as whatever was defined as
1939        "newline". Not only was this contrary to the documentation, which states
1940        that \n is character 10 (hex 0A), but it also went horribly wrong when
1941        "newline" was defined as CRLF. This has been fixed.
1942    
1943     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1944        was being set to -1 for the "end of line" case (supposedly a value that no
1945        character can have). Though this value is never used (the check for end of
1946        line is "zero bytes in current character"), it caused compiler complaints.
1947        I've changed it to 0xffffffff.
1948    
1949     8. In pcre_version.c, the version string was being built by a sequence of
1950        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1951        string (as it is for production releases) called a macro with an empty
1952        argument. The C standard says the result of this is undefined. The gcc
1953        compiler treats it as an empty string (which was what was wanted) but it is
1954        reported that Visual C gives an error. The source has been hacked around to
1955        avoid this problem.
1956    
1957     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1958        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1959        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1960        of them did).
1961    
1962    10. Originally, pcretest opened its input and output without "b"; then I was
1963        told that "b" was needed in some environments, so it was added for release
1964        5.0 to both the input and output. (It makes no difference on Unix-like
1965        systems.) Later I was told that it is wrong for the input on Windows. I've
1966        now abstracted the modes into two macros, to make it easier to fiddle with
1967        them, and removed "b" from the input mode under Windows.
1968    
1969    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1970    
1971    12. Added -help and --help to pcretest as an official way of being reminded
1972        of the options.
1973    
1974    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1975        and pcrecpp.cc because they annoy compilers at high warning levels.
1976    
1977    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1978    
1979    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1980        scripts are not C programs :-) and which was not noticed because it works
1981        on Linux.
1982    
1983    16. pcretest is supposed to handle any length of pattern and data line (as one
1984        line or as a continued sequence of lines) by extending its input buffer if
1985        necessary. This feature was broken for very long pattern lines, leading to
1986        a string of junk being passed to pcre_compile() if the pattern was longer
1987        than about 50K.
1988    
1989    17. I have done a major re-factoring of the way pcre_compile() computes the
1990        amount of memory needed for a compiled pattern. Previously, there was code
1991        that made a preliminary scan of the pattern in order to do this. That was
1992        OK when PCRE was new, but as the facilities have expanded, it has become
1993        harder and harder to keep it in step with the real compile phase, and there
1994        have been a number of bugs (see for example, 4 above). I have now found a
1995        cunning way of running the real compile function in a "fake" mode that
1996        enables it to compute how much memory it would need, while actually only
1997        ever using a few hundred bytes of working memory and without too many
1998        tests of the mode. This should make future maintenance and development
1999        easier. A side effect of this work is that the limit of 200 on the nesting
2000        depth of parentheses has been removed (though this was never a serious
2001        limitation, I suspect). However, there is a downside: pcre_compile() now
2002        runs more slowly than before (30% or more, depending on the pattern). I
2003        hope this isn't a big issue. There is no effect on runtime performance.
2004    
2005    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2006        newline (only possible for the last line of a file) and it was a
2007        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2008    
2009    19. Added additional timing features to pcretest. (1) The -tm option now times
2010        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2011        separate command line item, by a number that specifies the number of
2012        repeats to use when timing. The default is 50000; this gives better
2013        precision, but takes uncomfortably long for very large patterns.
2014    
2015    20. Extended pcre_study() to be more clever in cases where a branch of a
2016        subpattern has no definite first character. For example, (a*|b*)[cd] would
2017        previously give no result from pcre_study(). Now it recognizes that the
2018        first character must be a, b, c, or d.
2019    
2020    21. There was an incorrect error "recursive call could loop indefinitely" if
2021        a subpattern (or the entire pattern) that was being tested for matching an
2022        empty string contained only one non-empty item after a nested subpattern.
2023        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2024        incorrectly, because the \d was being skipped in the check.
2025    
2026    22. The pcretest program now has a new pattern option /B and a command line
2027        option -b, which is equivalent to adding /B to every pattern. This causes
2028        it to show the compiled bytecode, without the additional information that
2029        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2030        is the same as /B/I).
2031    
2032    23. A new optimization is now able automatically to treat some sequences such
2033        as a*b as a*+b. More specifically, if something simple (such as a character
2034        or a simple class like \d) has an unlimited quantifier, and is followed by
2035        something that cannot possibly match the quantified thing, the quantifier
2036        is automatically "possessified".
2037    
2038    24. A recursive reference to a subpattern whose number was greater than 39
2039        went wrong under certain circumstances in UTF-8 mode. This bug could also
2040        have affected the operation of pcre_study().
2041    
2042    25. Realized that a little bit of performance could be had by replacing
2043        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2044    
2045    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2046    
2047    27. Possessive quantifiers such as a++ were previously implemented by turning
2048        them into atomic groups such as ($>a+). Now they have their own opcodes,
2049        which improves performance. This includes the automatically created ones
2050        from 23 above.
2051    
2052    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2053        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2054        the first matched character to be a colon. This applied both to named and
2055        numbered groups.
2056    
2057    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2058    
2059    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2060        building PCRE on Windows easier, so I have included it in the distribution.
2061    
2062    31. There is now a check in pcretest against a ridiculously large number being
2063        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2064        loop, the loop is abandoned.
2065    
2066    32. Forward references to subpatterns in conditions such as (?(2)...) where
2067        subpattern 2 is defined later cause pcre_compile() to search forwards in
2068        the pattern for the relevant set of parentheses. This search went wrong
2069        when there were unescaped parentheses in a character class, parentheses
2070        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2071    
2072    33. "Subroutine" calls and backreferences were previously restricted to
2073        referencing subpatterns earlier in the regex. This restriction has now
2074        been removed.
2075    
2076    34. Added a number of extra features that are going to be in Perl 5.10. On the
2077        whole, these are just syntactic alternatives for features that PCRE had
2078        previously implemented using the Python syntax or my own invention. The
2079        other formats are all retained for compatibility.
2080    
2081        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2082            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2083            also .NET compatible.
2084    
2085        (b) A recursion or subroutine call to a named group can now be defined as
2086            (?&name) as well as (?P>name).
2087    
2088        (c) A backreference to a named group can now be defined as \k<name> or
2089            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2090            5.10, are also .NET compatible.
2091    
2092        (d) A conditional reference to a named group can now use the syntax
2093            (?(<name>) or (?('name') as well as (?(name).
2094    
2095        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2096            groups (named and numbered) that are never evaluated inline, but can be
2097            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2098            is always false. There may be only one alternative in such a group.
2099    
2100        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2101            as the simple (?(R). The condition is true only if the most recent
2102            recursion is that of the given number or name. It does not search out
2103            through the entire recursion stack.
2104    
2105        (g) The escape \gN or \g{N} has been added, where N is a positive or
2106            negative number, specifying an absolute or relative reference.
2107    
2108    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2109        some "unreachable code" warnings.
2110    
2111    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2112        things, this adds five new scripts.
2113    
2114    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2115        There were also incompatibilities regarding the handling of \Q..\E inside
2116        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2117        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2118    
2119    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2120        matches an empty string, and forcibly breaks the loop. There were bugs in
2121        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2122        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2123        separate and independent bugs (that affected different cases) have been
2124        fixed.
2125    
2126    39. Refactored the code to abolish the use of different opcodes for small
2127        capturing bracket numbers. This is a tidy that I avoided doing when I
2128        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2129        The new approach is not only tidier, it makes it possible to reduce the
2130        memory needed to fix the previous bug (38).
2131    
2132    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2133        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2134        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2135        mode.
2136    
2137    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2138        report.
2139    
2140    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2141        copy construction and assignment in the C++ wrapper.
2142    
2143    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2144        couple of bugs that could have given wrong results in the "--newline=crlf"
2145        case.
2146    
2147    44. Added a number of casts and did some reorganization of signed/unsigned int
2148        variables following suggestions from Dair Grant. Also renamed the variable
2149        "this" as "item" because it is a C++ keyword.
2150    
2151    45. Arranged for dftables to add
2152    
2153          #include "pcre_internal.h"
2154    
2155        to pcre_chartables.c because without it, gcc 4.x may remove the array
2156        definition from the final binary if PCRE is built into a static library and
2157        dead code stripping is activated.
2158    
2159    46. For an unanchored pattern, if a match attempt fails at the start of a
2160        newline sequence, and the newline setting is CRLF or ANY, and the next two
2161        characters are CRLF, advance by two characters instead of one.
2162    
2163    
2164    Version 6.7 04-Jul-06
2165    ---------------------
2166    
2167     1. In order to handle tests when input lines are enormously long, pcretest has
2168        been re-factored so that it automatically extends its buffers when
2169        necessary. The code is crude, but this _is_ just a test program. The
2170        default size has been increased from 32K to 50K.
2171    
2172     2. The code in pcre_study() was using the value of the re argument before
2173        testing it for NULL. (Of course, in any sensible call of the function, it
2174        won't be NULL.)
2175    
2176     3. The memmove() emulation function in pcre_internal.h, which is used on
2177        systems that lack both memmove() and bcopy() - that is, hardly ever -
2178        was missing a "static" storage class specifier.
2179    
2180     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2181        containing an extended class (one that cannot be represented by a bitmap
2182        because it contains high-valued characters or Unicode property items, e.g.
2183        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2184        pattern, but PCRE should not loop if you do not (it no longer does).
2185        [Detail: two cases were found: (a) a repeated subpattern containing an
2186        extended class; (b) a recursive reference to a subpattern that followed a
2187        previous extended class. It wasn't skipping over the extended class
2188        correctly when UTF-8 mode was not set.]
2189    
2190     5. A negated single-character class was not being recognized as fixed-length
2191        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2192        compile error "lookbehind assertion is not fixed length".
2193    
2194     6. The RunPerlTest auxiliary script was showing an unexpected difference
2195        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2196        write a Perl script that can interpret lines of an input file either as
2197        byte characters or as UTF-8, which is what "perltest" was being required to
2198        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2199        can't do is switch easily at run time between having the "use utf8;" pragma
2200        or not. In the end, I fudged it by using the RunPerlTest script to insert
2201        "use utf8;" explicitly for the UTF-8 tests.
2202    
2203     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2204        the end of the subject string, contrary to the documentation and to what
2205        Perl does. This was true of both matching functions. Now it matches only at
2206        the start of the subject and immediately after *internal* newlines.
2207    
2208     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2209        a pointer to an int instead of a pointer to an unsigned long int. This
2210        caused problems on 64-bit systems.
2211    
2212     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2213        instance of the 'standard' template library not being so standard".
2214    
2215    10. There was no check on the number of named subpatterns nor the maximum
2216        length of a subpattern name. The product of these values is used to compute
2217        the size of the memory block for a compiled pattern. By supplying a very
2218        long subpattern name and a large number of named subpatterns, the size
2219        computation could be caused to overflow. This is now prevented by limiting
2220        the length of names to 32 characters, and the number of named subpatterns
2221        to 10,000.
2222    
2223    11. Subpatterns that are repeated with specific counts have to be replicated in
2224        the compiled pattern. The size of memory for this was computed from the
2225        length of the subpattern and the repeat count. The latter is limited to
2226        65535, but there was no limit on the former, meaning that integer overflow
2227        could in principle occur. The compiled length of a repeated subpattern is
2228        now limited to 30,000 bytes in order to prevent this.
2229    
2230    12. Added the optional facility to have named substrings with the same name.
2231    
2232    13. Added the ability to use a named substring as a condition, using the
2233        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2234        are numbers (not recommended). Forward references are permitted.
2235    
2236    14. Added forward references in named backreferences (if you see what I mean).
2237    
2238    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2239        pattern could run off the end of the subject. For example, the pattern
2240        "(?s)(.{1,5})"8 did this with the subject "ab".
2241    
2242    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2243        PCRE_CASELESS was set when matching characters that were quantified with ?
2244        or *.
2245    
2246    17. A character class other than a single negated character that had a minimum
2247        but no maximum quantifier - for example [ab]{6,} - was not handled
2248        correctly by pce_dfa_exec(). It would match only one character.
2249    
2250    18. A valid (though odd) pattern that looked like a POSIX character
2251        class but used an invalid character after [ (for example [[,abc,]]) caused
2252        pcre_compile() to give the error "Failed: internal error: code overflow" or
2253        in some cases to crash with a glibc free() error. This could even happen if
2254        the pattern terminated after [[ but there just happened to be a sequence of
2255        letters, a binary zero, and a closing ] in the memory that followed.
2256    
2257    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2258        over the years. Originally (before any Unicode support), just the bottom 8
2259        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2260        output from "man perlunicode" includes this:
2261    
2262          The regular expression compiler produces polymorphic opcodes.  That
2263          is, the pattern adapts to the data and automatically switches to
2264          the Unicode character scheme when presented with Unicode data--or
2265          instead uses a traditional byte scheme when presented with byte
2266          data.
2267    
2268        Sadly, a wide octal escape does not cause a switch, and in a string with
2269        no other multibyte characters, these octal escapes are treated as before.
2270        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2271        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2272        Unicode string.
2273    
2274        I have not perpetrated such confusion in PCRE. Up till now, it took just
2275        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2276        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2277        translate to the appropriate multibyte character.
2278    
2279    29. Applied some refactoring to reduce the number of warnings from Microsoft
2280        and Borland compilers. This has included removing the fudge introduced
2281        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2282        a warning about an unused variable.
2283    
2284    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2285        characters since release 4.0, because Perl (from release 5.004) does not.
2286        [Or at least, is documented not to: some releases seem to be in conflict
2287        with the documentation.] However, when a pattern was studied with
2288        pcre_study() and all its branches started with \s, PCRE still included VT
2289        as a possible starting character. Of course, this did no harm; it just
2290        caused an unnecessary match attempt.
2291    
2292    22. Removed a now-redundant internal flag bit that recorded the fact that case
2293        dependency changed within the pattern. This was once needed for "required
2294        byte" processing, but is no longer used. This recovers a now-scarce options
2295        bit. Also moved the least significant internal flag bit to the most-
2296        significant bit of the word, which was not previously used (hangover from
2297        the days when it was an int rather than a uint) to free up another bit for
2298        the future.
2299    
2300    23. Added support for CRLF line endings as well as CR and LF. As well as the
2301        default being selectable at build time, it can now be changed at runtime
2302        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2303        specify that it is scanning data with non-default line endings.
2304    
2305    24. Changed the definition of CXXLINK to make it agree with the definition of
2306        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2307    
2308    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2309        recursions. This makes a big different to stack usage for some patterns.
2310    
2311    26. If a subpattern containing a named recursion or subroutine reference such
2312        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2313        the space required for the compiled pattern went wrong and gave too small a
2314        value. Depending on the environment, this could lead to "Failed: internal
2315        error: code overflow at offset 49" or "glibc detected double free or
2316        corruption" errors.
2317    
2318    27. Applied patches from Google (a) to support the new newline modes and (b) to
2319        advance over multibyte UTF-8 characters in GlobalReplace.
2320    
2321    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2322        difference for some implementation of PCRE in some Windows version.
2323    
2324    29. Added some extra testing facilities to pcretest:
2325    
2326        \q<number>   in a data line sets the "match limit" value
2327        \Q<number>   in a data line sets the "match recursion limt" value
2328        -S <number>  sets the stack size, where <number> is in megabytes
2329    
2330        The -S option isn't available for Windows.
2331    
2332    
2333    Version 6.6 06-Feb-06
2334    ---------------------
2335    
2336     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2337        in pcreposix.h. I have copied the definition from pcre.h.
2338    
2339     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2340        because pcre.h is no longer a built file.
2341    
2342     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2343        not normally included in the compiled code.
2344    
2345    
2346    Version 6.5 01-Feb-06
2347    ---------------------
2348    
2349     1. When using the partial match feature with pcre_dfa_exec(), it was not
2350        anchoring the second and subsequent partial matches at the new starting
2351        point. This could lead to incorrect results. For example, with the pattern
2352        /1234/, partially matching against "123" and then "a4" gave a match.
2353    
2354     2. Changes to pcregrep:
2355    
2356        (a) All non-match returns from pcre_exec() were being treated as failures
2357            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2358            error message is output. Some extra information is given for the
2359            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2360            probably the only errors that are likely to be caused by users (by
2361            specifying a regex that has nested indefinite repeats, for instance).
2362            If there are more than 20 of these errors, pcregrep is abandoned.
2363    
2364        (b) A binary zero was treated as data while matching, but terminated the
2365            output line if it was written out. This has been fixed: binary zeroes
2366            are now no different to any other data bytes.
2367    
2368        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2369            used to set a locale for matching. The --locale=xxxx long option has
2370            been added (no short equivalent) to specify a locale explicitly on the
2371            pcregrep command, overriding the environment variables.
2372    
2373        (d) When -B was used with -n, some line numbers in the output were one less
2374            than they should have been.
2375    
2376        (e) Added the -o (--only-matching) option.
2377    
2378        (f) If -A or -C was used with -c (count only), some lines of context were
2379            accidentally printed for the final match.
2380    
2381        (g) Added the -H (--with-filename) option.
2382    
2383        (h) The combination of options -rh failed to suppress file names for files
2384            that were found from directory arguments.
2385    
2386        (i) Added the -D (--devices) and -d (--directories) options.
2387    
2388        (j) Added the -F (--fixed-strings) option.
2389    
2390        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2391    
2392        (l) Added the --colo(u)r option.
2393    
2394        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2395            is not present by default.
2396    
2397     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2398        items such as (?R) or (?1), when the recursion could match a number of
2399        alternatives. If it matched one of the alternatives, but subsequently,
2400        outside the recursion, there was a failure, the code tried to back up into
2401        the recursion. However, because of the way PCRE is implemented, this is not
2402        possible, and the result was an incorrect result from the match.
2403    
2404        In order to prevent this happening, the specification of recursion has
2405        been changed so that all such subpatterns are automatically treated as
2406        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2407    
2408     4. I had overlooked the fact that, in some locales, there are characters for
2409        which isalpha() is true but neither isupper() nor islower() are true. In
2410        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2411        and ordfeminine) are like this. This affected the treatment of \w and \W
2412        when they appeared in character classes, but not when they appeared outside
2413        a character class. The bit map for "word" characters is now created
2414        separately from the results of isalnum() instead of just taking it from the
2415        upper, lower, and digit maps. (Plus the underscore character, of course.)
2416    
2417     5. The above bug also affected the handling of POSIX character classes such as
2418        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2419        permanent tables. Instead, the bit maps for such a class were previously
2420        created as the appropriate unions of the upper, lower, and digit bitmaps.
2421        Now they are created by subtraction from the [[:word:]] class, which has
2422        its own bitmap.
2423    
2424     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2425        It is created by subtracting the vertical space characters (\x09, \x0a,
2426        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2427        subtraction was done in the overall bitmap for a character class, meaning
2428        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2429        be recognized. This bug has been fixed.
2430    
2431     7. Patches from the folks at Google:
2432    
2433          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2434          real life, but is still worth protecting against".
2435    
2436          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2437          regular expressions".
2438    
2439          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2440          have it.
2441    
2442          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2443          "configure" and the latter not, in order to fix a problem somebody had
2444          with compiling the Arg class on HP-UX.
2445    
2446          (e) Improve the error-handling of the C++ wrapper a little bit.
2447    
2448          (f) New tests for checking recursion limiting.
2449    
2450     8. The pcre_memmove() function, which is used only if the environment does not
2451        have a standard memmove() function (and is therefore rarely compiled),
2452        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2453        returning a result (though PCRE never actually uses the result).
2454    
2455     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2456        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2457        returned instead of calling malloc() with an overflowing number that would
2458        most likely cause subsequent chaos.
2459    
2460    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2461    
2462    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2463        with this option is matched, the nmatch and pmatch options of regexec() are
2464        ignored.
2465    
2466    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2467        provided in case anyone wants to the the POSIX interface with UTF-8
2468        strings.
2469    
2470    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2471        C++ linking (needed for some HP-UX environments).
2472    
2473    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2474        (unused parameter) and in the pcre_printint() function (omitted "default"
2475        switch label when the default is to do nothing).
2476    
2477    15. Added some code to make it possible, when PCRE is compiled as a C++
2478        library, to replace subject pointers for pcre_exec() with a smart pointer
2479        class, thus making it possible to process discontinuous strings.
2480    
2481    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2482        much the same function. They were added by different people who were trying
2483        to make PCRE easy to compile on non-Unix systems. It has been suggested
2484        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2485        for compiling on Windows systems. I have therefore replaced it with
2486        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2487        defaults to "extern" for C or "extern C" for C++, which works fine on
2488        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2489        SCOPE with something explicit in config.h. In addition:
2490    
2491        (a) pcreposix.h still had just "extern" instead of either of these macros;
2492            I have replaced it with PCRE_DATA_SCOPE.
2493    
2494        (b) Functions such as _pcre_xclass(), which are internal to the library,
2495            but external in the C sense, all had PCRE_EXPORT in their definitions.
2496            This is apparently wrong for the Windows case, so I have removed it.
2497            (It makes no difference on Unix-like systems.)
2498    
2499    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2500        of recursive calls to match(). This is different to MATCH_LIMIT because
2501        that limits the total number of calls to match(), not all of which increase
2502        the depth of recursion. Limiting the recursion depth limits the amount of
2503        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2504        when PCRE is compiled, and changed at run time. A patch from Google adds
2505        this functionality to the C++ interface.
2506    
2507    18. Changes to the handling of Unicode character properties:
2508    
2509        (a) Updated the table to Unicode 4.1.0.
2510    
2511        (b) Recognize characters that are not in the table as "Cn" (undefined).
2512    
2513        (c) I revised the way the table is implemented to a much improved format
2514            which includes recognition of ranges. It now supports the ranges that
2515            are defined in UnicodeData.txt, and it also amalgamates other
2516            characters into ranges. This has reduced the number of entries in the
2517            table from around 16,000 to around 3,000, thus reducing its size
2518            considerably. I realized I did not need to use a tree structure after
2519            all - a binary chop search is just as efficient. Having reduced the
2520            number of entries, I extended their size from 6 bytes to 8 bytes to
2521            allow for more data.
2522    
2523        (d) Added support for Unicode script names via properties such as \p{Han}.
2524    
2525    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2526        matching that character.
2527    
2528    20. When matching a repeated Unicode property with a minimum greater than zero,
2529        (for example \pL{2,}), PCRE could look past the end of the subject if it
2530        reached it while seeking the minimum number of characters. This could
2531        happen only if some of the characters were more than one byte long, because
2532        there is a check for at least the minimum number of bytes.
2533    
2534    21. Refactored the implementation of \p and \P so as to be more general, to
2535        allow for more different types of property in future. This has changed the
2536        compiled form incompatibly. Anybody with saved compiled patterns that use
2537        \p or \P will have to recompile them.
2538    
2539    22. Added "Any" and "L&" to the supported property types.
2540    
2541    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2542        but give a compile time error if the value is greater than 0xff.
2543    
2544    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2545        accidentally not being installed or uninstalled.
2546    
2547    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2548        made were to insert the current release number. This seemed silly, because
2549        it made things harder for people building PCRE on systems that don't run
2550        "configure". I have turned pcre.h into a distributed file, no longer built
2551        by "configure", with the version identification directly included. There is
2552        no longer a pcre.h.in file.
2553    
2554        However, this change necessitated a change to the pcre-config script as
2555        well. It is built from pcre-config.in, and one of the substitutions was the
2556        release number. I have updated configure.ac so that ./configure now finds
2557        the release number by grepping pcre.h.
2558    
2559    26. Added the ability to run the tests under valgrind.
2560    
2561    
2562    Version 6.4 05-Sep-05
2563    ---------------------
2564    
2565     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2566        "--" to be printed when multiple files were scanned, even when none of the
2567        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2568        consider it to be a bug, and have restored the previous behaviour.
2569    
2570     2. A couple of code tidies to get rid of compiler warnings.
2571    
2572     3. The pcretest program used to cheat by referring to symbols in the library
2573        whose names begin with _pcre_. These are internal symbols that are not
2574        really supposed to be visible externally, and in some environments it is
2575        possible to suppress them. The cheating is now confined to including
2576        certain files from the library's source, which is a bit cleaner.
2577    
2578     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2579        file's purpose clearer.
2580    
2581     5. Reorganized pcre_ucp_findchar().
2582    
2583    
2584    Version 6.3 15-Aug-05
2585    ---------------------
2586    
2587     1. The file libpcre.pc.in did not have general read permission in the tarball.
2588    
2589     2. There were some problems when building without C++ support:
2590    
2591        (a) If C++ support was not built, "make install" and "make test" still
2592            tried to test it.
2593    
2594        (b) There were problems when the value of CXX was explicitly set. Some
2595            changes have been made to try to fix these, and ...
2596    
2597        (c) --disable-cpp can now be used to explicitly disable C++ support.
2598    
2599        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2600            backslash in a target when C++ was disabled. This confuses some
2601            versions of "make", apparently. Using an intermediate variable solves
2602            this. (Same for CPP_LOBJ.)
2603    
2604     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2605        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2606        necessary on certain architectures.
2607    
2608     4. Added a setting of -export-symbols-regex to the link command to remove
2609        those symbols that are exported in the C sense, but actually are local
2610        within the library, and not documented. Their names all begin with
2611        "_pcre_". This is not a perfect job, because (a) we have to except some
2612        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2613        available (and never for static libraries). I have made a note to try to
2614        find a way round (a) in the future.
2615    
2616    
2617    Version 6.2 01-Aug-05
2618    ---------------------
2619    
2620     1. There was no test for integer overflow of quantifier values. A construction
2621        such as {1111111111111111} would give undefined results. What is worse, if
2622        a minimum quantifier for a parenthesized subpattern overflowed and became
2623        negative, the calculation of the memory size went wrong. This could have
2624        led to memory overwriting.
2625    
2626     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2627    
2628     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2629        operating environments where this matters.
2630    
2631     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2632        PCRE options from within the C++ wrapper.
2633    
2634     5. Named capturing subpatterns were not being correctly counted when a pattern
2635        was compiled. This caused two problems: (a) If there were more than 100
2636        such subpatterns, the calculation of the memory needed for the whole
2637        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2638        back references of the form \12, where the number was greater than 9, were
2639        not recognized as back references, even though there were sufficient
2640        previous subpatterns.
2641    
2642     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2643        versions of gcc, e.g. 2.95.4.
2644    
2645    
2646    Version 6.1 21-Jun-05
2647    ---------------------
2648    
2649     1. There was one reference to the variable "posix" in pcretest.c that was not
2650        surrounded by "#if !defined NOPOSIX".
2651    
2652     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2653        the cross-check on the old pcre_info() function, for the benefit of the
2654        cut-down version of PCRE that is currently imported into Exim.
2655    
2656     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2657        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2658        patterns that start (?i) but I don't think that matters. The use of (?i) is
2659        just an example; this all applies to the other options as well.
2660    
2661     4. Since libtool seems to echo the compile commands it is issuing, the output
2662        from "make" can be reduced a bit by putting "@" in front of each libtool
2663        compile command.
2664    
2665     5. Patch from the folks at Google for configure.in to be a bit more thorough
2666        in checking for a suitable C++ installation before trying to compile the
2667        C++ stuff. This should fix a reported problem when a compiler was present,
2668        but no suitable headers.
2669    
2670     6. The man pages all had just "PCRE" as their title. I have changed them to
2671        be the relevant file name. I have also arranged that these names are
2672        retained in the file doc/pcre.txt, which is a concatenation in text format
2673        of all the man pages except the little individual ones for each function.
2674    
2675     7. The NON-UNIX-USE file had not been updated for the different set of source
2676        files that come with release 6. I also added a few comments about the C++
2677        wrapper.
2678    
2679    
2680    Version 6.0 07-Jun-05
2681    ---------------------
2682    
2683     1. Some minor internal re-organization to help with my DFA experiments.
2684    
2685     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2686        didn't matter for the library itself when fully configured, but did matter
2687        when compiling without UCP support, or within Exim, where the ucp files are
2688        not imported.
2689    
2690     3. Refactoring of the library code to split up the various functions into
2691        different source modules. The addition of the new DFA matching code (see
2692        below) to a single monolithic source would have made it really too
2693        unwieldy, quite apart from causing all the code to be include in a
2694        statically linked application, when only some functions are used. This is
2695        relevant even without the DFA addition now that patterns can be compiled in
2696        one application and matched in another.
2697    
2698        The downside of splitting up is that there have to be some external
2699        functions and data tables that are used internally in different modules of
2700        the library but which are not part of the API. These have all had their
2701        names changed to start with "_pcre_" so that they are unlikely to clash
2702        with other external names.
2703    
2704     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2705        a different (DFA) algorithm. Although it is slower than the original
2706        function, it does have some advantages for certain types of matching
2707        problem.
2708    
2709     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2710        including restarting after a partial match.
2711    
2712     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2713        defined when compiling for Windows was sent to me. I have put it into the
2714        code, though I have no means of testing or verifying it.
2715    
2716     7. Added the pcre_refcount() auxiliary function.
2717    
2718     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2719        match before or at the first newline in the subject string. In pcretest,
2720        the /f option on a pattern can be used to set this.
2721    
2722     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2723        would behave wrongly. This has been present in PCRE since release 4.0.
2724    
2725    10. A number of changes to the pcregrep command:
2726    
2727        (a) Refactored how -x works; insert ^(...)$ instead of setting
2728            PCRE_ANCHORED and checking the length, in preparation for adding
2729            something similar for -w.
2730    
2731        (b) Added the -w (match as a word) option.
2732    
2733        (c) Refactored the way lines are read and buffered so as to have more
2734            than one at a time available.
2735    
2736        (d) Implemented a pcregrep test script.
2737    
2738        (e) Added the -M (multiline match) option. This allows patterns to match
2739            over several lines of the subject. The buffering ensures that at least
2740            8K, or the rest of the document (whichever is the shorter) is available
2741            for matching (and similarly the previous 8K for lookbehind assertions).
2742    
2743        (f) Changed the --help output so that it now says
2744    
2745              -w, --word-regex(p)
2746    
2747            instead of two lines, one with "regex" and the other with "regexp"
2748            because that confused at least one person since the short forms are the
2749            same. (This required a bit of code, as the output is generated
2750            automatically from a table. It wasn't just a text change.)
2751    
2752        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2753            option but starts with a hyphen. Could be a pattern or a path name
2754            starting with a hyphen, for instance.
2755    
2756        (h) "-" can be given as a file name to represent stdin.
2757    
2758        (i) When file names are being printed, "(standard input)" is used for
2759            the standard input, for compatibility with GNU grep. Previously
2760            "<stdin>" was used.
2761    
2762        (j) The option --label=xxx can be used to supply a name to be used for
2763            stdin when file names are being printed. There is no short form.
2764    
2765        (k) Re-factored the options decoding logic because we are going to add
2766            two more options that take data. Such options can now be given in four
2767            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2768    
2769        (l) Added the -A, -B, and -C options for requesting that lines of context
2770            around matches be printed.
2771    
2772        (m) Added the -L option to print the names of files that do not contain
2773            any matching lines, that is, the complement of -l.
2774    
2775        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2776            continue to scan other files.
2777    
2778        (o) The -s option was incorrectly implemented. For compatibility with other
2779            greps, it now suppresses the error message for a non-existent or non-
2780            accessible file (but not the return code). There is a new option called
2781            -q that suppresses the output of matching lines, which was what -s was
2782            previously doing.
2783    
2784        (p) Added --include and --exclude options to specify files for inclusion
2785            and exclusion when recursing.
2786    
2787    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2788        Hopefully, it now does.
2789    
2790    12. Missing cast in pcre_study().
2791    
2792    13. Added an "uninstall" target to the makefile.
2793    
2794    14. Replaced "extern" in the function prototypes in Makefile.in with
2795        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2796        world, but is set differently for Windows.
2797    
2798    15. Added a second compiling function called pcre_compile2(). The only
2799        difference is that it has an extra argument, which is a pointer to an
2800        integer error code. When there is a compile-time failure, this is set
2801        non-zero, in addition to the error test pointer being set to point to an
2802        error message. The new argument may be NULL if no error number is required
2803        (but then you may as well call pcre_compile(), which is now just a
2804        wrapper). This facility is provided because some applications need a
2805        numeric error indication, but it has also enabled me to tidy up the way
2806        compile-time errors are handled in the POSIX wrapper.
2807    
2808    16. Added VPATH=.libs to the makefile; this should help when building with one
2809        prefix path and installing with another. (Or so I'm told by someone who
2810        knows more about this stuff than I do.)
2811    
2812    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2813        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2814        match everything, including newlines. This is not POSIX-compatible, but
2815        somebody wanted the feature. From pcretest it can be activated by using
2816        both the P and the s flags.
2817    
2818    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2819    
2820    19. libpcre.pc was being incorrectly installed as executable.
2821    
2822    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2823        it now also looks for '\r' so that it will work unmodified on Windows.
2824    
2825    21. Added Google's contributed C++ wrapper to the distribution.
2826    
2827    22. Added some untidy missing memory free() calls in pcretest, to keep
2828        Electric Fence happy when testing.
2829    
2830    
2831    
2832    Version 5.0 13-Sep-04
2833    ---------------------
2834    
2835     1. Internal change: literal characters are no longer packed up into items
2836        containing multiple characters in a single byte-string. Each character
2837        is now matched using a separate opcode. However, there may be more than one
2838        byte in the character in UTF-8 mode.
2839    
2840     2. The pcre_callout_block structure has two new fields: pattern_position and
2841        next_item_length. These contain the offset in the pattern to the next match
2842        item, and its length, respectively.
2843    
2844     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2845        insertion of callouts before each pattern item. Added the /C option to
2846        pcretest to make use of this.
2847    
2848     4. On the advice of a Windows user, the lines
2849    
2850          #if defined(_WIN32) || defined(WIN32)
2851          _setmode( _fileno( stdout ), 0x8000 );
2852          #endif  /* defined(_WIN32) || defined(WIN32) */
2853    
2854        have been added to the source of pcretest. This apparently does useful
2855        magic in relation to line terminators.
2856    
2857     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2858        for the benefit of those environments where the "b" makes a difference.
2859    
2860     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2861        to know about it. I have put a hack into configure.in that adds in code
2862        to set GCC=yes if CC=icc. This seems to end up at a point in the
2863        generated configure script that is early enough to affect the setting of
2864        compiler options, which is what is needed, but I have no means of testing
2865        whether it really works. (The user who reported this had patched the
2866        generated configure script, which of course I cannot do.)
2867    
2868        LATER: After change 22 below (new libtool files), the configure script
2869        seems to know about icc (and also ecc). Therefore, I have commented out
2870        this hack in configure.in.
2871    
2872     7. Added support for pkg-config (2 patches were sent in).
2873    
2874     8. Negated POSIX character classes that used a combination of internal tables
2875        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2876        [[:^ascii]]. Typically, they would match almost any characters. The other
2877        POSIX classes were not broken in this way.
2878    
2879     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2880        to find the match, as PCRE was deluded into thinking that the match had to
2881        start at the start point or following a newline. The same bug applied to
2882        patterns with negative forward assertions or any backward assertions
2883        preceding ".*" at the start, unless the pattern required a fixed first
2884        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2885    
2886    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2887        starting at the last subject character, bytes beyond the end of the subject
2888        string were read.
2889    
2890    11. Renamed the variable "class" as "classbits" to make life easier for C++
2891        users. (Previously there was a macro definition, but it apparently wasn't
2892        enough.)
2893    
2894    12. Added the new field "tables" to the extra data so that tables can be passed
2895        in at exec time, or the internal tables can be re-selected. This allows
2896        a compiled regex to be saved and re-used at a later time by a different
2897        program that might have everything at different addresses.
2898    
2899    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2900        -R library as well as a -L library.
2901    
2902    14. The debugging options of pcretest (-d on the command line or D on a
2903        pattern) showed incorrect output for anything following an extended class
2904        that contained multibyte characters and which was followed by a quantifier.
2905    
2906    15. Added optional support for general category Unicode character properties
2907        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2908        support. It adds about 90K to the size of the library. The meanings of the
2909        inbuilt class escapes such as \d and \s have NOT been changed.
2910    
2911    16. Updated pcredemo.c to include calls to free() to release the memory for the
2912        compiled pattern.
2913    
2914    17. The generated file chartables.c was being created in the source directory
2915        instead of in the building directory. This caused the build to fail if the
2916        source directory was different from the building directory, and was
2917        read-only.
2918    
2919    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2920        file. No doubt somebody will tell me if they don't make sense... Also added
2921        Dan Mooney's comments about building on OpenVMS.
2922    
2923    19. Added support for partial matching via the PCRE_PARTIAL option for
2924        pcre_exec() and the \P data escape in pcretest.
2925    
2926    20. Extended pcretest with 3 new pattern features:
2927    
2928        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2929              write the compiled pattern to the file whose name is "rest-of-line".
2930              This is a straight binary dump of the data, with the saved pointer to
2931              the character tables forced to be NULL. The study data, if any, is
2932              written too. After writing, pcretest reads a new pattern.
2933    
2934        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2935              compiled pattern from the given file. There must not be any
2936              occurrences of "<" in the file name (pretty unlikely); if there are,
2937              pcretest will instead treat the initial "<" as a pattern delimiter.
2938              After reading in the pattern, pcretest goes on to read data lines as
2939              usual.
2940    
2941        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2942              and 16-bit fields in a compiled pattern, to simulate a pattern that
2943              was compiled on a host of opposite endianness.
2944    
2945    21. The pcre-exec() function can now cope with patterns that were compiled on
2946        hosts of opposite endianness, with this restriction:
2947    
2948          As for any compiled expression that is saved and used later, the tables
2949          pointer field cannot be preserved; the extra_data field in the arguments
2950          to pcre_exec() should be used to pass in a tables address if a value
2951          other than the default internal tables were used at compile time.
2952    
2953    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2954        now diagnosed as an error. Previously, most of the time, a negative number
2955        would have been treated as zero, but if in addition "ovector" was passed as
2956        NULL, a crash could occur.
2957    
2958    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2959        new versions from the libtool 1.5 distribution (the last one is a copy of
2960        a file called libtool.m4). This seems to have fixed the need to patch
2961        "configure" to support Darwin 1.3 (which I used to do). However, I still
2962        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2963        workstation).
2964    
2965    24. Changed the PCRE licence to be the more standard "BSD" licence.
2966    
2967    
2968    Version 4.5 01-Dec-03
2969    ---------------------
2970    
2971     1. There has been some re-arrangement of the code for the match() function so
2972        that it can be compiled in a version that does not call itself recursively.
2973        Instead, it keeps those local variables that need separate instances for
2974        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2975        needs to "recurse". Keeping track of where control must go is done by means
2976        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2977        hide most of the details from the main code, and operates only if
2978        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2979        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2980        operating.
2981    
2982        To make it easier for callers to provide specially tailored get/free
2983        functions for this usage, two new functions, pcre_stack_malloc, and
2984        pcre_stack_free, are used. They are always called in strict stacking order,
2985        and the size of block requested is always the same.
2986    
2987        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2988        PCRE has been compiled to use the stack or the heap for recursion. The
2989        -C option of pcretest uses this to show which version is compiled.
2990    
2991        A new data escape \S, is added to pcretest; it causes the amounts of store
2992        obtained and freed by both kinds of malloc/free at match time to be added
2993        to the output.
2994    
2995     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2996        what's available on my current Linux desktop machine.
2997    
2998     3. When matching a UTF-8 string, the test for a valid string at the start has
2999        been extended. If start_offset is not zero, PCRE now checks that it points
3000        to a byte that is the start of a UTF-8 character. If not, it returns
3001        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3002        this is necessary because there may be backward assertions in the pattern.
3003        When matching the same subject several times, it may save resources to use
3004        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3005    
3006     4. The code for checking the validity of UTF-8 strings has been tightened so
3007        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3008        containing "overlong sequences".
3009    
3010     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3011        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3012        should have been "&", but it just so happened that all the cases this let
3013        through by mistake were picked up later in the function.
3014    
3015     6. I had used a variable called "isblank" - this is a C99 function, causing
3016        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3017    
3018     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3019        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3020        so the version is shown; (c) stop "make test" echoing "./RunTest".
3021    
3022     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3023    
3024     9. The prototype for memmove() for systems that don't have it was using
3025        size_t, but the inclusion of the header that defines size_t was later. I've
3026        moved the #includes for the C headers earlier to avoid this.
3027    
3028    10. Added some adjustments to the code to make it easier to compiler on certain
3029        special systems:
3030    
3031          (a) Some "const" qualifiers were missing.
3032          (b) Added the macro EXPORT before all exported functions; by default this
3033              is defined to be empty.
3034          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3035              that it reads its output file name as an argument instead of writing
3036              to the standard output and assuming this can be redirected.
3037    
3038    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3039        class containing characters with values greater than 255, PCRE compilation
3040        went into a loop.
3041    
3042    12. A recursive reference to a subpattern that was within another subpattern
3043        that had a minimum quantifier of zero caused PCRE to crash. For example,
3044        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3045        recursion. If the recursively-called subpattern itself had a zero repeat,
3046        that was OK.
3047    
3048    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3049        buffer into which it was copied (for escape processing) was still set at
3050        1024, so long lines caused crashes.
3051    
3052    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3053        "internal error: code overflow...". This applied to any character class
3054        that was followed by a possessive quantifier.
3055    
3056    15. Modified the Makefile to add libpcre.la as a prerequisite for
3057        libpcreposix.la because I was told this is needed for a parallel build to
3058        work.
3059    
3060    16. If a pattern that contained .* following optional items at the start was
3061        studied, the wrong optimizing data was generated, leading to matching
3062        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3063        matching string must start with a or b or c. The correct conclusion for
3064        this pattern is that a match can start with any character.
3065    
3066    
3067    Version 4.4 13-Aug-03
3068    ---------------------
3069    
3070     1. In UTF-8 mode, a character class containing characters with values between
3071        127 and 255 was not handled correctly if the compiled pattern was studied.
3072        In fixing this, I have also improved the studying algorithm for such
3073        classes (slightly).
3074    
3075     2. Three internal functions had redundant arguments passed to them. Removal
3076        might give a very teeny performance improvement.
3077    
3078     3. Documentation bug: the value of the capture_top field in a callout is *one
3079        more than* the number of the hightest numbered captured substring.
3080    
3081     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3082        in incorrectly linking with a previously installed version. They now link
3083        explicitly with libpcre.la.
3084    
3085     5. configure.in no longer needs to recognize Cygwin specially.
3086    
3087     6. A problem in pcre.in for Windows platforms is fixed.
3088    
3089     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3090        pcretest, it used to include the size of the study block as part of its
3091        output. Unfortunately, the structure contains a field that has a different
3092        size on different hardware architectures. This meant that the tests that
3093        showed this size failed. As the block is currently always of a fixed size,
3094        this information isn't actually particularly useful in pcretest output, so
3095        I have just removed it.
3096    
3097     8. Three pre-processor statements accidentally did not start in column 1.
3098        Sadly, there are *still* compilers around that complain, even though
3099        standard C has not required this for well over a decade. Sigh.
3100    
3101     9. In pcretest, the code for checking callouts passed small integers in the
3102        callout_data field, which is a void * field. However, some picky compilers
3103        complained about the casts involved for this on 64-bit systems. Now
3104        pcretest passes the address of the small integer instead, which should get
3105        rid of the warnings.
3106    
3107    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3108        both compile and run time, and gives an error if an invalid UTF-8 sequence
3109        is found. There is a option for disabling this check in cases where the
3110        string is known to be correct and/or the maximum performance is wanted.
3111    
3112    11. In response to a bug report, I changed one line in Makefile.in from
3113    
3114            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3115        to
3116            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3117    
3118        to look similar to other lines, but I have no way of telling whether this
3119        is the right thing to do, as I do not use Windows. No doubt I'll get told
3120        if it's wrong...
3121    
3122    
3123    Version 4.3 21-May-03
3124    ---------------------
3125    
3126    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3127       Makefile.
3128    
3129    2. Some refactoring to improve the quality of the code:
3130    
3131       (i)   The utf8_table... variables are now declared "const".
3132    
3133       (ii)  The code for \cx, which used the "case flipping" table to upper case
3134             lower case letters, now just substracts 32. This is ASCII-specific,
3135             but the whole concept of \cx is ASCII-specific, so it seems
3136             reasonable.
3137    
3138       (iii) PCRE was using its character types table to recognize decimal and
3139             hexadecimal digits in the pattern. This is silly, because it handles
3140             only 0-9, a-f, and A-F, but the character types table is locale-
3141             specific, which means strange things might happen. A private
3142             table is now used for this - though it costs 256 bytes, a table is
3143             much faster than multiple explicit tests. Of course, the standard
3144             character types table is still used for matching digits in subject
3145             strings against \d.
3146    
3147       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3148             ending in _t are). So I've renamed it as ESC_tee.
3149    
3150    3. The first argument for regexec() in the POSIX wrapper should have been
3151       defined as "const".
3152    
3153    4. Changed pcretest to use malloc() for its buffers so that they can be
3154       Electric Fenced for debugging.
3155    
3156    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3157       to read one or more bytes before the start of the subject string. Often this
3158       had no effect on PCRE's behaviour, but in some circumstances it could
3159       provoke a segmentation fault.
3160    
3161    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3162       to try to read one or more bytes before the start of the subject string.
3163    
3164    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3165       UTF-8 support could misbehave in various ways if the subject string
3166       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3167       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3168       back over UTF-8 characters.)
3169    
3170    
3171    Version 4.2 14-Apr-03
3172    ---------------------
3173    
3174    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3175    
3176    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3177         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3178         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3179         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3180         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3181           and BUILD_EXEEXT
3182         Note: automatic setting of the BUILD variables is not yet working
3183         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3184           compile-time but not at link-time
3185         [LINK]: use for linking executables only
3186         make different versions for Windows and non-Windows
3187         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3188           libraries
3189         [LINK_FOR_BUILD]: new variable
3190         [OBJEXT]: use throughout
3191         [EXEEXT]: use throughout
3192         <winshared>: new target
3193         <wininstall>: new target
3194         <dftables.o>: use native compiler
3195         <dftables>: use native linker
3196         <install>: handle Windows platform correctly
3197         <clean>: ditto
3198         <check>: ditto
3199         copy DLL to top builddir before testing
3200    
3201       As part of these changes, -no-undefined was removed again. This was reported
3202       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3203       in any case.
3204    
3205    3. Some tidies to get rid of compiler warnings:
3206    
3207       . In the match_data structure, match_limit was an unsigned long int, whereas
3208         match_call_count was an int. I've made them both unsigned long ints.
3209    
3210       . In pcretest the fact that a const uschar * doesn't automatically cast to
3211         a void * provoked a warning.
3212    
3213       . Turning on some more compiler warnings threw up some "shadow" variables
3214         and a few more missing casts.
3215    
3216    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3217       option, a class that contained a single character with a value between 128
3218       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3219    
3220    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3221       option, a class that contained several characters, but with at least one
3222       whose value was between 128 and 255 caused PCRE to crash.
3223    
3224    
3225    Version 4.1 12-Mar-03
3226    ---------------------
3227    
3228    1. Compiling with gcc -pedantic found a couple of places where casts were
3229    needed, and a string in dftables.c that was longer than standard compilers are
3230    required to support.
3231    
3232    2. Compiling with Sun's compiler found a few more places where the code could
3233    be tidied up in order to avoid warnings.
3234    
3235    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3236    first of these names is deprecated in the latest Autoconf in favour of the name
3237    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3238    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3239    analogy I have changed it to CFLAGS_FOR_BUILD.
3240    
3241    4. Added -no-undefined to the linking command in the Makefile, because this is
3242    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3243    linking step for the pcreposix library.
3244    
3245    5. PCRE was failing to diagnose the case of two named groups with the same
3246    name.
3247    
3248    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3249    literal character that is needed in the subject for a match, and scans along to
3250    ensure that it is present before embarking on the full matching process. This
3251    saves time in cases of nested unlimited repeats that are never going to match.
3252    Problem: the scan can take a lot of time if the subject is very long (e.g.
3253    megabytes), thus penalizing straightforward matches. It is now done only if the
3254    amount of subject to be scanned is less than 1000 bytes.
3255    
3256    7. A lesser problem with the same optimization is that it was recording the
3257    first character of an anchored pattern as "needed", thus provoking a search
3258    right along the subject, even when the first match of the pattern was going to
3259    fail. The "needed" character is now not set for anchored patterns, unless it
3260    follows something in the pattern that is of non-fixed length. Thus, it still
3261    fulfils its original purpose of finding quick non-matches in cases of nested
3262    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3263    
3264    
3265    Version 4.0 17-Feb-03
3266    ---------------------
3267    
3268    1. If a comment in an extended regex that started immediately after a meta-item
3269    extended to the end of string, PCRE compiled incorrect data. This could lead to
3270    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3271    
3272    2. Moved to autoconf 2.53 and libtool 1.4.2.
3273    
3274    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3275    the special perltest8 script is no longer needed - all the tests can be run
3276    from a single perltest script.
3277    
3278    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3279    by \s. It has now been removed in PCRE. This means it isn't recognized as
3280    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3281    class [:space:] *does* include VT, thereby creating a mess.
3282    
3283    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3284    space and tab.
3285    
3286    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3287    its new features into the main test script, reducing the number of scripts.
3288    
3289    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3290    were backward compatible, and made the (?i) apply to the whole pattern, as if
3291    /i were given. Now it behaves more logically, and applies the option setting
3292    only to what follows. PCRE has been changed to follow suit. However, if it
3293    finds options settings right at the start of the pattern, it extracts them into
3294    the global options, as before. Thus, they show up in the info data.
3295    
3296    8. Added support for the \Q...\E escape sequence. Characters in between are
3297    treated as literals. This is slightly different from Perl in that $ and @ are
3298    also handled as literals inside the quotes. In Perl, they will cause variable
3299    interpolation. Note the following examples:
3300    
3301        Pattern            PCRE matches      Perl matches
3302    
3303        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3304        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3305        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3306    
3307    For compatibility with Perl, \Q...\E sequences are recognized inside character
3308    classes as well as outside them.
3309    
3310    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3311    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3312    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3313    signed/unsigned warnings.
3314    
3315    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3316    option for pcretest, so I've replaced it by a simple function that does just
3317    that job.
3318    
3319    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3320    "pcregrep -".
3321    
3322    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3323    Java package. This provides some syntactic sugar for simple cases of what my
3324    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3325    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3326    item, you can use this simplified notation. Note that only makes sense with
3327    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3328    greediness, whatever the setting of the PCRE_UNGREEDY option.
3329    
3330    13. A change of greediness default within a pattern was not taking effect at
3331    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3332    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3333    was abstracted outside.
3334    
3335    14. PCRE now supports the \G assertion. It is true when the current matching
3336    position is at the start point of the match. This differs from \A when the
3337    starting offset is non-zero. Used with the /g option of pcretest (or similar
3338    code), it works in the same way as it does for Perl's /g option. If all
3339    alternatives of a regex begin with \G, the expression is anchored to the start
3340    match position, and the "anchored" flag is set in the compiled expression.
3341    
3342    15. Some bugs concerning the handling of certain option changes within patterns
3343    have been fixed. These applied to options other than (?ims). For example,
3344    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3345    been the other way round. Some of this was related to change 7 above.
3346    
3347    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3348    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3349    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3350    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3351    
3352    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3353    mode. Unlike ".", it always matches newline, whatever the setting of
3354    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3355    assertions. Perl allows it, but it doesn't (in general) work because it can't
3356    calculate the length of the lookbehind. At least, that's the case for Perl
3357    5.8.0 - I've been told they are going to document that it doesn't work in
3358    future.
3359    
3360    18. Added an error diagnosis for escapes that PCRE does not support: these are
3361    \L, \l, \N, \P, \p, \U, \u, and \X.
3362    
3363    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3364    reading past the end of the pattern in cases such as /[abcd/.
3365    
3366    20. PCRE was getting more memory than necessary for patterns with classes that
3367    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3368    
3369    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3370    compiling PCRE for use with Virtual Pascal.
3371    
3372    22. Small fix to the Makefile to make it work properly if the build is done
3373    outside the source tree.
3374    
3375    23. Added a new extension: a condition to go with recursion. If a conditional
3376    subpattern starts with (?(R) the "true" branch is used if recursion has
3377    happened, whereas the "false" branch is used only at the top level.
3378    
3379    24. When there was a very long string of literal characters (over 255 bytes
3380    without UTF support, over 250 bytes with UTF support), the computation of how
3381    much memory was required could be incorrect, leading to segfaults or other
3382    strange effects.
3383    
3384    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3385    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3386    there was a subsequent back reference to those brackets. This meant that, for
3387    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3388    possible to check for precisely this case. All we can do is abandon the
3389    optimization if .* occurs inside capturing brackets when there are any back
3390    references whatsoever. (See below for a better fix that came later.)
3391    
3392    26. The handling of the optimization for finding the first character of a
3393    non-anchored pattern, and for finding a character that is required later in the
3394    match were failing in some cases. This didn't break the matching; it just
3395    failed to optimize when it could. The way this is done has been re-implemented.
3396    
3397    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3398    
3399    28. Added a new feature that provides some of the functionality that Perl
3400    provides with (?{...}). The facility is termed a "callout". The way it is done
3401    in PCRE is for the caller to provide an optional function, by setting
3402    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3403    global variable. By default it is unset, which disables all calling out. To get
3404    the function called, the regex must include (?C) at appropriate points. This
3405    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3406    This provides a means of identifying different callout points. When PCRE
3407    reaches such a point in the regex, if pcre_callout has been set, the external
3408    function is called. It is provided with data in a structure called
3409    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3410    matching continues; if it returns a non-zero value, the match at the current
3411    point fails. However, backtracking will occur if possible. [This was changed
3412    later and other features added - see item 49 below.]
3413    
3414    29. pcretest is upgraded to test the callout functionality. It provides a
3415    callout function that displays information. By default, it shows the start of
3416    the match and the current position in the text. There are some new data escapes
3417    to vary what happens:
3418    
3419        \C+         in addition, show current contents of captured substrings
3420        \C-         do not supply a callout function
3421        \C!n        return 1 when callout number n is reached
3422        \C!n!m      return 1 when callout number n is reached for the mth time
3423    
3424    30. If pcregrep was called with the -l option and just a single file name, it
3425    output "<stdin>" if a match was found, instead of the file name.
3426    
3427    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3428    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3429    pcre_exec(). This saves a malloc/free per call. The default value of
3430    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3431    when configuring.
3432    
3433    32. The default maximum size of a compiled pattern is 64K. There have been a
3434    few cases of people hitting this limit. The code now uses macros to handle the
3435    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3436    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3437    configuring. Tests 2 and 5 work only with 2-byte links because they output
3438    debugging information about compiled patterns.
3439    
3440    33. Internal code re-arrangements:
3441    
3442    (a) Moved the debugging function for printing out a compiled regex into
3443        its own source file (printint.c) and used #include to pull it into
3444        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3445        separate copies.
3446    
3447    (b) Defined the list of op-code names for debugging as a macro in
3448        internal.h so that it is next to the definition of the opcodes.
3449    
3450    (c) Defined a table of op-code lengths for simpler skipping along compiled
3451        code. This is again a macro in internal.h so that it is next to the
3452        definition of the opcodes.
3453    
3454    34. Added support for recursive calls to individual subpatterns, along the
3455    lines of Robin Houston's patch (but implemented somewhat differently).
3456    
3457    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3458    allow it to read and process whole directories in Win32. This code was
3459    contributed by Lionel Fourquaux; it has not been tested by me.
3460    
3461    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3462    used to name a group. Names consist of alphanumerics and underscores, and must
3463    be unique. Back references use the syntax (?P=name) and recursive calls use
3464    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3465    numbers. The function pcre_fullinfo() can be used after compilation to extract
3466    a name/number map. There are three relevant calls:
3467    
3468      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3469      PCRE_INFO_NAMECOUNT            yields the number of entries
3470      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3471    
3472    The map is a vector of fixed-size entries. The size of each entry depends on
3473    the length of the longest name used. The first two bytes of each entry are the
3474    group number, most significant byte first. There follows the corresponding
3475    name, zero terminated. The names are in alphabetical order.
3476    
3477    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3478    case instead of 255. Making it the same both with and without UTF-8 support
3479    means that the same test output works with both.
3480    
3481    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3482    calling malloc() with a zero argument.
3483    
3484    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3485    optimization. I've improved things by keeping a bitmap of backreferences with
3486    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3487    fact referenced, the optimization can be applied. It is unlikely that a
3488    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3489    the match to follow \n) will appear inside brackets with a number greater than
3490    31, but if it does, any back reference > 31 suppresses the optimization.
3491    
3492    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3493    of disabling numbered capturing parentheses. Any opening parenthesis that is
3494    not followed by ? behaves as if it were followed by ?: but named parentheses
3495    can still be used for capturing (and they will acquire numbers in the usual
3496    way).
3497    
3498    41. Redesigned the return codes from the match() function into yes/no/error so
3499    that errors can be passed back from deep inside the nested calls. A malloc
3500    failure while inside a recursive subpattern call now causes the
3501    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3502    
3503    42. It is now possible to set a limit on the number of times the match()
3504    function is called in a call to pcre_exec(). This facility makes it possible to
3505    limit the amount of recursion and backtracking, though not in a directly
3506    obvious way, because the match() function is used in a number of different
3507    circumstances. The count starts from zero for each position in the subject
3508    string (for non-anchored patterns). The default limit is, for compatibility, a
3509    large number, namely 10 000 000. You can change this in two ways:
3510    
3511    (a) When configuring PCRE before making, you can use --with-match-limit=n
3512        to set a default value for the compiled library.
3513    
3514    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3515        a different value is set. See 45 below.
3516    
3517    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3518    
3519    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3520    of things that can be changed at compile time. The first argument specifies
3521    what is wanted and the second points to where the information is to be placed.
3522    The current list of available information is:
3523    
3524      PCRE_CONFIG_UTF8
3525    
3526    The output is an integer that is set to one if UTF-8 support is available;
3527    otherwise it is set to zero.
3528    
3529      PCRE_CONFIG_NEWLINE
3530    
3531    The output is an integer that it set to the value of the code that is used for
3532    newline. It is either LF (10) or CR (13).
3533    
3534      PCRE_CONFIG_LINK_SIZE
3535    
3536    The output is an integer that contains the number of bytes used for internal
3537    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3538    
3539      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3540    
3541    The output is an integer that contains the threshold above which the POSIX
3542    interface uses malloc() for output vectors. See item 31 above.
3543    
3544      PCRE_CONFIG_MATCH_LIMIT
3545    
3546    The output is an unsigned integer that contains the default limit of the number
3547    of match() calls in a pcre_exec() execution. See 42 above.
3548    
3549    44. pcretest has been upgraded by the addition of the -C option. This causes it
3550    to extract all the available output from the new pcre_config() function, and to
3551    output it. The program then exits immediately.
3552    
3553    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3554    order to support additional features. One way would have been to define
3555    pcre_exec2() (for example) with extra arguments, but this would not have been
3556    extensible, and would also have required all calls to the original function to
3557    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3558    is used for passing in "extra" data from pcre_study().
3559    
3560    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3561    contains the following fields:
3562    
3563      flags         a bitmap indicating which of the following fields are set
3564      study_data    opaque data from pcre_study()
3565      match_limit   a way of specifying a limit on match() calls for a specific
3566                      call to pcre_exec()
3567      callout_data  data for callouts (see 49 below)
3568    
3569    The flag bits are also defined in pcre.h, and are
3570    
3571      PCRE_EXTRA_STUDY_DATA
3572      PCRE_EXTRA_MATCH_LIMIT
3573      PCRE_EXTRA_CALLOUT_DATA
3574    
3575    The pcre_study() function now returns one of these new pcre_extra blocks, with
3576    the actual study data pointed to by the study_data field, and the
3577    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3578    before. That is, this change is entirely upwards-compatible and requires no
3579    change to existing code.
3580    
3581    If you want to pass in additional data to pcre_exec(), you can either place it
3582    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3583    block.
3584    
3585    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3586    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3587    times with different match limits, until it finds the minimum value needed for
3588    pcre_exec() to complete. The value is then output. This can be instructive; for
3589    most simple matches the number is quite small, but for pathological cases it
3590    gets very large very quickly.
3591    
3592    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3593    returns the size of the data block pointed to by the study_data field in a
3594    pcre_extra block, that is, the value that was passed as the argument to
3595    pcre_malloc() when PCRE was getting memory in which to place the information
3596    created by pcre_study(). The fourth argument should point to a size_t variable.
3597    pcretest has been extended so that this information is shown after a successful
3598    pcre_study() call when information about the compiled regex is being displayed.
3599    
3600    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3601    because what follows is always an absolute path. (Later: it turns out that this
3602    is more than cosmetic for MinGW, because it doesn't like empty path
3603    components.)
3604    
3605    49. Some changes have been made to the callout feature (see 28 above):
3606    
3607    (i)  A callout function now has three choices for what it returns:
3608    
3609           0  =>  success, carry on matching
3610         > 0  =>  failure at this point, but backtrack if possible
3611         < 0  =>  serious error, return this value from pcre_exec()
3612    
3613         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3614         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3615         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3616         use by callout functions. It will never be used by PCRE itself.
3617    
3618    (ii) The pcre_extra structure (see 45 above) has a void * field called
3619         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3620         pcre_callout_block structure has a field of the same name. The contents of
3621         the field passed in the pcre_extra structure are passed to the callout
3622         function in the corresponding field in the callout block. This makes it
3623         easier to use the same callout-containing regex from multiple threads. For
3624         testing, the pcretest program has a new data escape
3625    
3626           \C*n        pass the number n (may be negative) as callout_data
3627    
3628         If the callout function in pcretest receives a non-zero value as
3629         callout_data, it returns that value.
3630    
3631    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3632    there were some redundant $(CFLAGS) in commands that are now specified as
3633    $(LINK), which already includes $(CFLAGS).
3634    
3635    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3636    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3637    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3638    one-byte characters throughout. Note that case-insensitive matching applies
3639    only to characters whose values are less than 256. PCRE doesn't support the
3640    notion of cases for higher-valued characters.
3641    
3642    (i)   A character class whose characters are all within 0-255 is handled as
3643          a bit map, and the map is inverted for negative classes. Previously, a
3644          character > 255 always failed to match such a class; however it should
3645          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3646    
3647    (ii)  A negated character class with a single character < 255 is coded as
3648          "not this character" (OP_NOT). This wasn't working properly when the test
3649          character was multibyte, either singly or repeated.
3650    
3651    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3652          mode, for example: \x{100}{2,3}.
3653    
3654    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3655          singly or repeated) now correctly test multibyte characters. However,
3656          PCRE doesn't recognize any characters with values greater than 255 as
3657          digits, spaces, or word characters. Such characters always match \D, \S,
3658          and \W, and never match \d, \s, or \w.
3659    
3660    (v)   Classes may now contain characters and character ranges with values
3661          greater than 255. For example: [ab\x{100}-\x{400}].
3662    
3663    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3664          PCRE in UTF-8 mode.
3665    
3666    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3667    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3668    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3669    value.)
3670    
3671    53. The single man page has become too large. I have therefore split it up into
3672    a number of separate man pages. These also give rise to individual HTML pages;
3673    these are now put in a separate directory, and there is an index.html page that
3674    lists them all. Some hyperlinking between the pages has been installed.
3675    
3676    54. Added convenience functions for handling named capturing parentheses.
3677    
3678    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3679    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3680    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3681    are faulted.
3682    
3683    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3684    calling configure. These values are used when compiling the dftables.c program
3685    which is run to generate the source of the default character tables. They
3686    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3687    you will need to set these values.
3688    
3689    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3690    
3691    
3692    Version 3.9 02-Jan-02
3693    ---------------------
3694    
3695    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3696    
3697    2. If --disable-static was given, the building process failed when trying to
3698    build pcretest and pcregrep. (For some reason it was using libtool to compile
3699    them, which is not right, as they aren't part of the library.)
3700    
3701    
3702    Version 3.8 18-Dec-01
3703    ---------------------
3704    
3705    1. The experimental UTF-8 code was completely screwed up. It was packing the
3706    bytes in the wrong order. How dumb can you get?
3707    
3708    
3709    Version 3.7 29-Oct-01