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

Diff of /code/trunk/ChangeLog

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

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