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

Diff of /code/trunk/ChangeLog

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

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