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

Diff of /code/trunk/ChangeLog

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

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