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

Diff of /code/trunk/ChangeLog

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

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