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

Diff of /code/trunk/ChangeLog

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

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