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

Diff of /code/trunk/ChangeLog

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

revision 9 by nigel, Sat Feb 24 21:38:13 2007 UTC revision 733 by ph10, Tue Oct 11 10:29:36 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.20 10-Oct-2011
5    ------------------------
6    
7    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
8        a POSIX class. After further experiments with Perl, which convinced me that
9        Perl has bugs and confusions, a closing square bracket is no longer allowed
10        in a POSIX name. This bug also affected patterns with classes that started
11        with full stops.
12    
13    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
14        captured substring, but while checking the failing first alternative,
15        substring 1 is temporarily captured. If the output vector supplied to
16        pcre_exec() was not big enough for this capture, the yield of the function
17        was still zero ("insufficient space for captured substrings"). This cannot
18        be totally fixed without adding another stack variable, which seems a lot
19        of expense for a edge case. However, I have improved the situation in cases
20        such as /(a)(b)x|abc/ matched against "abc", where the return code
21        indicates that fewer than the maximum number of slots in the ovector have
22        been set.
23    
24    3.  Related to (2) above: when there are more back references in a pattern than
25        slots in the output vector, pcre_exec() uses temporary memory during
26        matching, and copies in the captures as far as possible afterwards. It was
27        using the entire output vector, but this conflicts with the specification
28        that only 2/3 is used for passing back captured substrings. Now it uses
29        only the first 2/3, for compatibility. This is, of course, another edge
30        case.
31    
32    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
33        main code base, and can be used by building with --enable-jit. When this is
34        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
35        runtime --no-jit option is given.
36    
37    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
38        ovector, the return from the function was zero, implying that there were
39        other matches that did not fit. The correct "exactly full" value is now
40        returned.
41    
42    6.  If a subpattern that was called recursively or as a subroutine contained
43        (*PRUNE) or any other control that caused it to give a non-standard return,
44        invalid errors such as "Error -26 (nested recursion at the same subject
45        position)" or even infinite loops could occur.
46    
47    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
48        computing the minimum length on reaching *ACCEPT, and so ended up with the
49        wrong value of 1 rather than 0. Further investigation indicates that
50        computing a minimum subject length in the presence of *ACCEPT is difficult
51        (think back references, subroutine calls), and so I have changed the code
52        so that no minimum is registered for a pattern that contains *ACCEPT.
53    
54    8.  If (*THEN) was present in the first (true) branch of a conditional group,
55        it was not handled as intended. [But see 16 below.]
56    
57    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
58        Sheri Pierce.
59    
60    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
61        the first byte in a match must be "a".
62    
63    11. Change 17 for 8.13 increased the recursion depth for patterns like
64        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
65        pattern contains any instances of (*THEN). If it does not, the old
66        optimizations are restored. It would be nice to do this on a per-group
67        basis, but at the moment that is not feasible.
68    
69    12. In some environments, the output of pcretest -C is CRLF terminated. This
70        broke RunTest's code that checks for the link size. A single white space
71        after the value is now allowed for.
72    
73    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
74        For "fr", it uses the Windows-specific input and output files.
75    
76    14. If (*THEN) appeared in a group that was called recursively or as a
77        subroutine, it did not work as intended. [But see next item.]
78    
79    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
80        pattern fragments (but not containing any | characters). If A and B are
81        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
82        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
83        D. In other words, Perl considers parentheses that do not contain any |
84        characters to be part of a surrounding alternative, whereas PCRE was
85        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
86        differently. PCRE now behaves in the same way as Perl, except in the case
87        of subroutine/recursion calls such as (?1) which have in any case always
88        been different (but PCRE had them first :-).
89    
90    16. Related to 15 above: Perl does not treat the | in a conditional group as
91        creating alternatives. Such a group is treated in the same way as an
92        ordinary group without any | characters when processing (*THEN). PCRE has
93        been changed to match Perl's behaviour.
94    
95    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
96        RunGrepTest script failed.
97    
98    18. Change 22 for version 13 caused atomic groups to use more stack. This is
99        inevitable for groups that contain captures, but it can lead to a lot of
100        stack use in large patterns. The old behaviour has been restored for atomic
101        groups that do not contain any capturing parentheses.
102    
103    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
104        suppress the check for a minimum subject length at run time. (If it was
105        given to pcre_exec() or pcre_dfa_exec() it did work.)
106    
107    
108    Version 8.13 16-Aug-2011
109    ------------------------
110    
111    1.  The Unicode data tables have been updated to Unicode 6.0.0.
112    
113    2.  Two minor typos in pcre_internal.h have been fixed.
114    
115    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
116        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
117        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
118    
119    4.  There were a number of related bugs in the code for matching backrefences
120        caselessly in UTF-8 mode when codes for the characters concerned were
121        different numbers of bytes. For example, U+023A and U+2C65 are an upper
122        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
123        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
124        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
125        2-byte code at the end of the subject (it thought there wasn't enough data
126        left).
127    
128    5.  Comprehensive information about what went wrong is now returned by
129        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
130        as the output vector has at least 2 elements. The offset of the start of
131        the failing character and a reason code are placed in the vector.
132    
133    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
134        now returned is for the first byte of the failing character, instead of the
135        last byte inspected. This is an incompatible change, but I hope it is small
136        enough not to be a problem. It makes the returned offset consistent with
137        pcre_exec() and pcre_dfa_exec().
138    
139    7.  pcretest now gives a text phrase as well as the error number when
140        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
141        failure, the offset and reason code are output.
142    
143    8.  When \R was used with a maximizing quantifier it failed to skip backwards
144        over a \r\n pair if the subsequent match failed. Instead, it just skipped
145        back over a single character (\n). This seems wrong (because it treated the
146        two characters as a single entity when going forwards), conflicts with the
147        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
148        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
149        has been changed.
150    
151    9.  Some internal refactoring has changed the processing so that the handling
152        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
153        time (the PCRE_DOTALL option was changed this way some time ago: version
154        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
155        which was always a bit of a fudge. It also means that there is one less
156        argument for the match() function, which reduces its stack requirements
157        slightly. This change also fixes an incompatibility with Perl: the pattern
158        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
159    
160    10. More internal refactoring has drastically reduced the number of recursive
161        calls to match() for possessively repeated groups such as (abc)++ when
162        using pcre_exec().
163    
164    11. While implementing 10, a number of bugs in the handling of groups were
165        discovered and fixed:
166    
167        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
168        (a|)*(?1) gave a compile-time internal error.
169        ((a|)+)+  did not notice that the outer group could match an empty string.
170        (^a|^)+   was not marked as anchored.
171        (.*a|.*)+ was not marked as matching at start or after a newline.
172    
173    12. Yet more internal refactoring has removed another argument from the match()
174        function. Special calls to this function are now indicated by setting a
175        value in a variable in the "match data" data block.
176    
177    13. Be more explicit in pcre_study() instead of relying on "default" for
178        opcodes that mean there is no starting character; this means that when new
179        ones are added and accidentally left out of pcre_study(), testing should
180        pick them up.
181    
182    14. The -s option of pcretest has been documented for ages as being an old
183        synonym of -m (show memory usage). I have changed it to mean "force study
184        for every regex", that is, assume /S for every regex. This is similar to -i
185        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
186        using it. It makes it easier to run collections of tests with and without
187        study enabled, and thereby test pcre_study() more easily. All the standard
188        tests are now run with and without -s (but some patterns can be marked as
189        "never study" - see 20 below).
190    
191    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
192        restoration of the capturing data to the outer values was not happening
193        correctly.
194    
195    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
196        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
197        pattern had matched an empty string, and so incorrectly returned a no
198        match.
199    
200    17. There was optimizing code for the last branch of non-capturing parentheses,
201        and also for the obeyed branch of a conditional subexpression, which used
202        tail recursion to cut down on stack usage. Unfortunately, now that there is
203        the possibility of (*THEN) occurring in these branches, tail recursion is
204        no longer possible because the return has to be checked for (*THEN). These
205        two optimizations have therefore been removed. [But see 8.20/11 above.]
206    
207    18. If a pattern containing \R was studied, it was assumed that \R always
208        matched two bytes, thus causing the minimum subject length to be
209        incorrectly computed because \R can also match just one byte.
210    
211    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
212        was incorrectly computed.
213    
214    20. If /S is present twice on a test pattern in pcretest input, it now
215        *disables* studying, thereby overriding the use of -s on the command line
216        (see 14 above). This is necessary for one or two tests to keep the output
217        identical in both cases.
218    
219    21. When (*ACCEPT) was used in an assertion that matched an empty string and
220        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
221    
222    22. When an atomic group that contained a capturing parenthesis was
223        successfully matched, but the branch in which it appeared failed, the
224        capturing was not being forgotten if a higher numbered group was later
225        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
226        group 1 to "a", when in fact it should be unset. This applied to multi-
227        branched capturing and non-capturing groups, repeated or not, and also to
228        positive assertions (capturing in negative assertions does not happen
229        in PCRE) and also to nested atomic groups.
230    
231    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
232        subject after a captured substring, to make it easier to tell which of a
233        number of identical substrings has been captured.
234    
235    24. The way atomic groups are processed by pcre_exec() has been changed so that
236        if they are repeated, backtracking one repetition now resets captured
237        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
238        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
239        "aaa". Previously, it would have been "a". As part of this code
240        refactoring, the way recursive calls are handled has also been changed.
241    
242    25. If an assertion condition captured any substrings, they were not passed
243        back unless some other capturing happened later. For example, if
244        (?(?=(a))a) was matched against "a", no capturing was returned.
245    
246    26. When studying a pattern that contained subroutine calls or assertions,
247        the code for finding the minimum length of a possible match was handling
248        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
249        group 1 called group 2 while simultaneously a separate group 2 called group
250        1). A stack overflow occurred in this case. I have fixed this by limiting
251        the recursion depth to 10.
252    
253    27. Updated RunTest.bat in the distribution to the version supplied by Tom
254        Fortmann. This supports explicit test numbers on the command line, and has
255        argument validation and error reporting.
256    
257    28. An instance of \X with an unlimited repeat could fail if at any point the
258        first character it looked at was a mark character.
259    
260    29. Some minor code refactoring concerning Unicode properties and scripts
261        should reduce the stack requirement of match() slightly.
262    
263    30. Added the '=' option to pcretest to check the setting of unused capturing
264        slots at the end of the pattern, which are documented as being -1, but are
265        not included in the return count.
266    
267    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
268        compiled something random. Now it gives a compile-time error (as does
269        Perl).
270    
271    32. A *MARK encountered during the processing of a positive assertion is now
272        recorded and passed back (compatible with Perl).
273    
274    33. If --only-matching or --colour was set on a pcregrep call whose pattern
275        had alternative anchored branches, the search for a second match in a line
276        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
277        matched the line "0102" twice. The same bug affected patterns that started
278        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
279        twice.
280    
281    34. Previously, PCRE did not allow quantification of assertions. However, Perl
282        does, and because of capturing effects, quantifying parenthesized
283        assertions may at times be useful. Quantifiers are now allowed for
284        parenthesized assertions.
285    
286    35. A minor code tidy in pcre_compile() when checking options for \R usage.
287    
288    36. \g was being checked for fancy things in a character class, when it should
289        just be a literal "g".
290    
291    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
292        appearance of a nested POSIX class supersedes an apparent external class.
293        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
294        unescaped square brackets may also appear as part of class names. For
295        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
296        more like Perl. (But see 8.20/1 above.)
297    
298    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
299        was because it thought it was \N{name}, which is not supported).
300    
301    39. Add minix to OS list not supporting the -S option in pcretest.
302    
303    40. PCRE tries to detect cases of infinite recursion at compile time, but it
304        cannot analyze patterns in sufficient detail to catch mutual recursions
305        such as ((?1))((?2)). There is now a runtime test that gives an error if a
306        subgroup is called recursively as a subpattern for a second time at the
307        same position in the subject string. In previous releases this might have
308        been caught by the recursion limit, or it might have run out of stack.
309    
310    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
311        happen only once. PCRE was, however incorrectly giving a compile time error
312        "recursive call could loop indefinitely" because it cannot analyze the
313        pattern in sufficient detail. The compile time test no longer happens when
314        PCRE is compiling a conditional subpattern, but actual runaway loops are
315        now caught at runtime (see 40 above).
316    
317    42. It seems that Perl allows any characters other than a closing parenthesis
318        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
319        has been changed to be the same.
320    
321    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
322        as not to get warnings when autogen.sh is called. Also changed
323        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
324    
325    44. To help people who use pcregrep to scan files containing exceedingly long
326        lines, the following changes have been made:
327    
328        (a) The default value of the buffer size parameter has been increased from
329            8K to 20K. (The actual buffer used is three times this size.)
330    
331        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
332            PCRE is built.
333    
334        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
335            to be set at run time.
336    
337        (d) Numerical values in pcregrep options can be followed by K or M, for
338            example --buffer-size=50K.
339    
340        (e) If a line being scanned overflows pcregrep's buffer, an error is now
341            given and the return code is set to 2.
342    
343    45. Add a pointer to the latest mark to the callout data block.
344    
345    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
346        partial match of an empty string instead of no match. This was specific to
347        the use of ".".
348    
349    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
350        complete match instead of a partial match. This bug was dependent on both
351        the PCRE_UTF8 and PCRE_DOTALL options being set.
352    
353    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
354        starting byte set, because \b was not being ignored.
355    
356    
357    Version 8.12 15-Jan-2011
358    ------------------------
359    
360    1.  Fixed some typos in the markup of the man pages, and wrote a script that
361        checks for such things as part of the documentation building process.
362    
363    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
364        --match-limit and --recursion-limit options (added for 8.11). In
365        particular, this made one of the standard tests fail. (The integer value
366        went into the wrong half of a long int.)
367    
368    3.  If the --colour option was given to pcregrep with -v (invert match), it
369        did strange things, either producing crazy output, or crashing. It should,
370        of course, ignore a request for colour when reporting lines that do not
371        match.
372    
373    4.  Another pcregrep bug caused similar problems if --colour was specified with
374        -M (multiline) and the pattern match finished with a line ending.
375    
376    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
377        matched in multiline mode, the following line was shown as part of the
378        match. This seems wrong, so I have changed it.
379    
380    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
381        the check for further matches in the same line (so they could be coloured)
382        to overrun the end of the current line. If another match was found, it was
383        incorrectly shown (and then shown again when found in the next line).
384    
385    7.  If pcregrep was compiled under Windows, there was a reference to the
386        function pcregrep_exit() before it was defined. I am assuming this was
387        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
388        reported by a user. I've moved the definition above the reference.
389    
390    
391    Version 8.11 10-Dec-2010
392    ------------------------
393    
394    1.  (*THEN) was not working properly if there were untried alternatives prior
395        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
396        backtracked to try for "b" instead of moving to the next alternative branch
397        at the same level (in this case, to look for "c"). The Perl documentation
398        is clear that when (*THEN) is backtracked onto, it goes to the "next
399        alternative in the innermost enclosing group".
400    
401    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
402        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
403        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
404        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
405        (*THEN).
406    
407    3.  If \s appeared in a character class, it removed the VT character from
408        the class, even if it had been included by some previous item, for example
409        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
410        of \s, but is part of the POSIX "space" class.)
411    
412    4.  A partial match never returns an empty string (because you can always
413        match an empty string at the end of the subject); however the checking for
414        an empty string was starting at the "start of match" point. This has been
415        changed to the "earliest inspected character" point, because the returned
416        data for a partial match starts at this character. This means that, for
417        example, /(?<=abc)def/ gives a partial match for the subject "abc"
418        (previously it gave "no match").
419    
420    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
421        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
422        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
423        has an implication that the given string is incomplete (because a partial
424        match is preferred over a full match). For this reason, these items now
425        give a partial match in this situation. [Aside: previously, the one case
426        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
427        match rather than a full match, which was wrong by the old rules, but is
428        now correct.]
429    
430    6.  There was a bug in the handling of #-introduced comments, recognized when
431        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
432        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
433        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
434        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
435        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
436        places in pcre_compile().
437    
438    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
439        comments when looking ahead for named forward references to subpatterns,
440        the only newline sequence it recognized was NL. It now handles newlines
441        according to the set newline convention.
442    
443    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
444        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
445        cater for a lack of strerror(). These oversights have been fixed.
446    
447    9.  Added --match-limit and --recursion-limit to pcregrep.
448    
449    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
450    
451    11. When the -o option was used, pcregrep was setting a return code of 1, even
452        when matches were found, and --line-buffered was not being honoured.
453    
454    12. Added an optional parentheses number to the -o and --only-matching options
455        of pcregrep.
456    
457    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
458        can match an empty string. The code to do it in pcretest and pcredemo
459        needed fixing:
460    
461        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
462            only one byte after an empty string match just before CRLF (this case
463            just got forgotten; "any" and "anycrlf" were OK).
464    
465        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
466            mode when an empty string match preceded an ASCII character followed by
467            a non-ASCII character. (The code for advancing by one character rather
468            than one byte was nonsense.)
469    
470        (c) The pcredemo.c sample program did not have any code at all to handle
471            the cases when CRLF is a valid newline sequence.
472    
473    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
474        as a starting offset was within the subject string. There is now a new
475        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
476        negative or greater than the length of the string. In order to test this,
477        pcretest is extended to allow the setting of negative starting offsets.
478    
479    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
480        starting offset points to the beginning of a UTF-8 character was
481        unnecessarily clumsy. I tidied it up.
482    
483    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
484        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
485    
486    17. Nobody had reported that the --include_dir option, which was added in
487        release 7.7 should have been called --include-dir (hyphen, not underscore)
488        for compatibility with GNU grep. I have changed it to --include-dir, but
489        left --include_dir as an undocumented synonym, and the same for
490        --exclude-dir, though that is not available in GNU grep, at least as of
491        release 2.5.4.
492    
493    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
494        characters from a string of bytes) have been redefined so as not to use
495        loops, in order to improve performance in some environments. At the same
496        time, I abstracted some of the common code into auxiliary macros to save
497        repetition (this should not affect the compiled code).
498    
499    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
500        compile-time error is now given if \c is not followed by an ASCII
501        character, that is, a byte less than 128. (In EBCDIC mode, the code is
502        different, and any byte value is allowed.)
503    
504    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
505        START_OPTIMIZE option, which is now allowed at compile time - but just
506        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
507        to pcregrep and other applications that have no direct access to PCRE
508        options. The new /Y option in pcretest sets this option when calling
509        pcre_compile().
510    
511    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
512        back references. Groups containing recursive back references were forced to
513        be atomic by that change, but in the case of named groups, the amount of
514        memory required was incorrectly computed, leading to "Failed: internal
515        error: code overflow". This has been fixed.
516    
517    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
518        pcretest.c, to avoid build problems in some Borland environments.
519    
520    
521    Version 8.10 25-Jun-2010
522    ------------------------
523    
524    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
525        THEN.
526    
527    2.  (*ACCEPT) was not working when inside an atomic group.
528    
529    3.  Inside a character class, \B is treated as a literal by default, but
530        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
531        causes the error). The code is unchanged, but I tidied the documentation.
532    
533    4.  Inside a character class, PCRE always treated \R and \X as literals,
534        whereas Perl faults them if its -w option is set. I have changed PCRE so
535        that it faults them when PCRE_EXTRA is set.
536    
537    5.  Added support for \N, which always matches any character other than
538        newline. (It is the same as "." when PCRE_DOTALL is not set.)
539    
540    6.  When compiling pcregrep with newer versions of gcc which may have
541        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
542        declared with attribute warn_unused_result" were given. Just casting the
543        result to (void) does not stop the warnings; a more elaborate fudge is
544        needed. I've used a macro to implement this.
545    
546    7.  Minor change to pcretest.c to avoid a compiler warning.
547    
548    8.  Added four artifical Unicode properties to help with an option to make
549        \s etc use properties (see next item). The new properties are: Xan
550        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
551    
552    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
553        use Unicode properties. (*UCP) at the start of a pattern can be used to set
554        this option. Modified pcretest to add /W to test this facility. Added
555        REG_UCP to make it available via the POSIX interface.
556    
557    10. Added --line-buffered to pcregrep.
558    
559    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
560        studied, and the match started with a letter with a code point greater than
561        127 whose first byte was different to the first byte of the other case of
562        the letter, the other case of this starting letter was not recognized
563        (#976).
564    
565    12. If a pattern that was studied started with a repeated Unicode property
566        test, for example, \p{Nd}+, there was the theoretical possibility of
567        setting up an incorrect bitmap of starting bytes, but fortunately it could
568        not have actually happened in practice until change 8 above was made (it
569        added property types that matched character-matching opcodes).
570    
571    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
572        possible starting bytes for non-anchored patterns.
573    
574    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
575        \R, and also a number of cases that involve Unicode properties, both
576        explicit and implicit when PCRE_UCP is set.
577    
578    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
579        input, it could crash or give wrong results if characters with values
580        greater than 0xc0 were present in the subject string. (Detail: it assumed
581        UTF-8 input when processing these items.)
582    
583    16. Added a lot of (int) casts to avoid compiler warnings in systems where
584        size_t is 64-bit (#991).
585    
586    17. Added a check for running out of memory when PCRE is compiled with
587        --disable-stack-for-recursion (#990).
588    
589    18. If the last data line in a file for pcretest does not have a newline on
590        the end, a newline was missing in the output.
591    
592    19. The default pcre_chartables.c file recognizes only ASCII characters (values
593        less than 128) in its various bitmaps. However, there is a facility for
594        generating tables according to the current locale when PCRE is compiled. It
595        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
596        characters, are recognized by isspace() and therefore were getting set in
597        these tables, and indeed these tables seem to approximate to ISO 8859. This
598        caused a problem in UTF-8 mode when pcre_study() was used to create a list
599        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
600        which of course cannot start UTF-8 characters. I have changed the code so
601        that only real ASCII characters (less than 128) and the correct starting
602        bytes for UTF-8 encodings are set for characters greater than 127 when in
603        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
604        altogether.)
605    
606    20. Added the /T option to pcretest so as to be able to run tests with non-
607        standard character tables, thus making it possible to include the tests
608        used for 19 above in the standard set of tests.
609    
610    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
611        reference to a subpattern the other side of a comment that contains an
612        opening parenthesis caused either an internal compiling error, or a
613        reference to the wrong subpattern.
614    
615    
616    Version 8.02 19-Mar-2010
617    ------------------------
618    
619    1.  The Unicode data tables have been updated to Unicode 5.2.0.
620    
621    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
622        configured.
623    
624    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
625        original author of that file, following a query about its status.
626    
627    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
628        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
629    
630    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
631        quantifier applied to a forward-referencing subroutine call, could compile
632        incorrect code or give the error "internal error: previously-checked
633        referenced subpattern not found".
634    
635    6.  Both MS Visual Studio and Symbian OS have problems with initializing
636        variables to point to external functions. For these systems, therefore,
637        pcre_malloc etc. are now initialized to local functions that call the
638        relevant global functions.
639    
640    7.  There were two entries missing in the vectors called coptable and poptable
641        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
642        I've fixed the data, and added a kludgy way of testing at compile time that
643        the lengths are correct (equal to the number of opcodes).
644    
645    8.  Following on from 7, I added a similar kludge to check the length of the
646        eint vector in pcreposix.c.
647    
648    9.  Error texts for pcre_compile() are held as one long string to avoid too
649        much relocation at load time. To find a text, the string is searched,
650        counting zeros. There was no check for running off the end of the string,
651        which could happen if a new error number was added without updating the
652        string.
653    
654    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
655    
656    11. \K was not working if it appeared in an atomic group or in a group that
657        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
658        \K is "not well defined" if used in an assertion. PCRE now accepts it if
659        the assertion is positive, but not if it is negative.
660    
661    12. Change 11 fortuitously reduced the size of the stack frame used in the
662        "match()" function of pcre_exec.c by one pointer. Forthcoming
663        implementation of support for (*MARK) will need an extra pointer on the
664        stack; I have reserved it now, so that the stack frame size does not
665        decrease.
666    
667    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
668        item in branch that calls a recursion is a subroutine call - as in the
669        second branch in the above example - was incorrectly given the compile-
670        time error "recursive call could loop indefinitely" because pcre_compile()
671        was not correctly checking the subroutine for matching a non-empty string.
672    
673    14. The checks for overrunning compiling workspace could trigger after an
674        overrun had occurred. This is a "should never occur" error, but it can be
675        triggered by pathological patterns such as hundreds of nested parentheses.
676        The checks now trigger 100 bytes before the end of the workspace.
677    
678    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
679    
680    
681    Version 8.01 19-Jan-2010
682    ------------------------
683    
684    1.  If a pattern contained a conditional subpattern with only one branch (in
685        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
686        computed the wrong minimum data length (which is of course zero for such
687        subpatterns). This could cause incorrect "no match" results.
688    
689    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
690        the pattern is reset in the first branch, pcre_compile() failed with
691        "internal error: code overflow at offset...". This happened only when
692        the reset was to the original external option setting. (An optimization
693        abstracts leading options settings into an external setting, which was the
694        cause of this.)
695    
696    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
697        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
698        assertion pattern did not match (meaning that the assertion was true), it
699        was incorrectly treated as false if the SKIP had been reached during the
700        matching. This also applied to assertions used as conditions.
701    
702    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
703        assertion subpattern, including such a pattern used as a condition,
704        unpredictable results occurred, instead of the error return
705        PCRE_ERROR_DFA_UITEM.
706    
707    5.  The C++ GlobalReplace function was not working like Perl for the special
708        situation when an empty string is matched. It now does the fancy magic
709        stuff that is necessary.
710    
711    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
712        removed. (These were left over from very, very early versions of PCRE.)
713    
714    7.  Some cosmetic changes to the code to make life easier when compiling it
715        as part of something else:
716    
717        (a) Change DEBUG to PCRE_DEBUG.
718    
719        (b) In pcre_compile(), rename the member of the "branch_chain" structure
720            called "current" as "current_branch", to prevent a collision with the
721            Linux macro when compiled as a kernel module.
722    
723        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
724            prevent a collision with the Linux macro when compiled as a kernel
725            module.
726    
727    8.  In pcre_compile() there are some checks for integer overflows that used to
728        cast potentially large values to (double). This has been changed to that
729        when building, a check for int64_t is made, and if it is found, it is used
730        instead, thus avoiding the use of floating point arithmetic. (There is no
731        other use of FP in PCRE.) If int64_t is not found, the fallback is to
732        double.
733    
734    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
735        2005 (difference between two addresses compared to an unsigned value).
736    
737    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
738        custom one, because of the following reported problem in Windows:
739    
740          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
741              under Win32.
742          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
743              therefore missing the function definition.
744          - The compiler thus generates a "C" signature for the test function.
745          - The linker fails to find the "C" function.
746          - PCRE fails to configure if asked to do so against libbz2.
747    
748    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
749        messages were output:
750    
751          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
752          rerunning libtoolize, to keep the correct libtool macros in-tree.
753          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
754    
755        I have done both of these things.
756    
757    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
758        most of the time, it *can* run out if it is given a pattern that contains a
759        runaway infinite recursion. I updated the discussion in the pcrestack man
760        page.
761    
762    13. Now that we have gone to the x.xx style of version numbers, the minor
763        version may start with zero. Using 08 or 09 is a bad idea because users
764        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
765        interpreted as invalid octal numbers. I've updated the previous comment in
766        configure.ac, and also added a check that gives an error if 08 or 09 are
767        used.
768    
769    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
770        causing partial matching to fail when the end of the subject matched \W
771        in a UTF-8 pattern where \W was quantified with a minimum of 3.
772    
773    15. There were some discrepancies between the declarations in pcre_internal.h
774        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
775        their definitions. The declarations used "const uschar *" and the
776        definitions used USPTR. Even though USPTR is normally defined as "const
777        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
778        reported that: "This difference in casting confuses some C++ compilers, for
779        example, SunCC recognizes above declarations as different functions and
780        generates broken code for hbpcre." I have changed the declarations to use
781        USPTR.
782    
783    16. GNU libtool is named differently on some systems. The autogen.sh script now
784        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
785        (FreeBSD).
786    
787    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
788        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
789        comment: "Figure out how to create a longlong from a string: strtoll and
790        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
791        instance, but it only takes 2 args instead of 3!"
792    
793    18. A subtle bug concerned with back references has been fixed by a change of
794        specification, with a corresponding code fix. A pattern such as
795        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
796        refers, was giving matches when it shouldn't. For example, xa=xaaa would
797        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
798        same bug. Such groups have to be quantified to be useful, or contained
799        inside another quantified group. (If there's no repetition, the reference
800        can never match.) The problem arises because, having left the group and
801        moved on to the rest of the pattern, a later failure that backtracks into
802        the group uses the captured value from the final iteration of the group
803        rather than the correct earlier one. I have fixed this in PCRE by forcing
804        any group that contains a reference to itself to be an atomic group; that
805        is, there cannot be any backtracking into it once it has completed. This is
806        similar to recursive and subroutine calls.
807    
808    
809    Version 8.00 19-Oct-09
810    ----------------------
811    
812    1.  The table for translating pcre_compile() error codes into POSIX error codes
813        was out-of-date, and there was no check on the pcre_compile() error code
814        being within the table. This could lead to an OK return being given in
815        error.
816    
817    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
818        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
819        in a Windows environment.
820    
821    3.  The pcregrep --count option prints the count for each file even when it is
822        zero, as does GNU grep. However, pcregrep was also printing all files when
823        --files-with-matches was added. Now, when both options are given, it prints
824        counts only for those files that have at least one match. (GNU grep just
825        prints the file name in this circumstance, but including the count seems
826        more useful - otherwise, why use --count?) Also ensured that the
827        combination -clh just lists non-zero counts, with no names.
828    
829    4.  The long form of the pcregrep -F option was incorrectly implemented as
830        --fixed_strings instead of --fixed-strings. This is an incompatible change,
831        but it seems right to fix it, and I didn't think it was worth preserving
832        the old behaviour.
833    
834    5.  The command line items --regex=pattern and --regexp=pattern were not
835        recognized by pcregrep, which required --regex pattern or --regexp pattern
836        (with a space rather than an '='). The man page documented the '=' forms,
837        which are compatible with GNU grep; these now work.
838    
839    6.  No libpcreposix.pc file was created for pkg-config; there was just
840        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
841    
842    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
843        when UCP support is not needed, by modifying the Python script that
844        generates it from Unicode data files. This should not matter if the module
845        is correctly used as a library, but I received one complaint about 50K of
846        unwanted data. My guess is that the person linked everything into his
847        program rather than using a library. Anyway, it does no harm.
848    
849    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
850        was a minimum greater than 1 for a wide character in a possessive
851        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
852        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
853        character. Chaos in the form of incorrect output or a compiling loop could
854        result.
855    
856    9.  The restrictions on what a pattern can contain when partial matching is
857        requested for pcre_exec() have been removed. All patterns can now be
858        partially matched by this function. In addition, if there are at least two
859        slots in the offset vector, the offset of the earliest inspected character
860        for the match and the offset of the end of the subject are set in them when
861        PCRE_ERROR_PARTIAL is returned.
862    
863    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
864        synonymous with PCRE_PARTIAL, for backwards compatibility, and
865        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
866        and may be more useful for multi-segment matching.
867    
868    11. Partial matching with pcre_exec() is now more intuitive. A partial match
869        used to be given if ever the end of the subject was reached; now it is
870        given only if matching could not proceed because another character was
871        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
872        string "Z", which now yields "no match" instead of "partial match". In the
873        case of pcre_dfa_exec(), "no match" is given if every matching path for the
874        final character ended with (*FAIL).
875    
876    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
877        if the pattern had a "must contain" character that was already found in the
878        earlier partial match, unless partial matching was again requested. For
879        example, with the pattern /dog.(body)?/, the "must contain" character is
880        "g". If the first part-match was for the string "dog", restarting with
881        "sbody" failed. This bug has been fixed.
882    
883    13. The string returned by pcre_dfa_exec() after a partial match has been
884        changed so that it starts at the first inspected character rather than the
885        first character of the match. This makes a difference only if the pattern
886        starts with a lookbehind assertion or \b or \B (\K is not supported by
887        pcre_dfa_exec()). It's an incompatible change, but it makes the two
888        matching functions compatible, and I think it's the right thing to do.
889    
890    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
891        so that the demonstration program is easily available in environments where
892        PCRE has not been installed from source.
893    
894    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
895        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
896        library.
897    
898    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
899        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
900        is not the first non-POSIX option to be added. Clearly some people find
901        these options useful.
902    
903    17. If a caller to the POSIX matching function regexec() passes a non-zero
904        value for nmatch with a NULL value for pmatch, the value of
905        nmatch is forced to zero.
906    
907    18. RunGrepTest did not have a test for the availability of the -u option of
908        the diff command, as RunTest does. It now checks in the same way as
909        RunTest, and also checks for the -b option.
910    
911    19. If an odd number of negated classes containing just a single character
912        interposed, within parentheses, between a forward reference to a named
913        subpattern and the definition of the subpattern, compilation crashed with
914        an internal error, complaining that it could not find the referenced
915        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
916        [The bug was that it was starting one character too far in when skipping
917        over the character class, thus treating the ] as data rather than
918        terminating the class. This meant it could skip too much.]
919    
920    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
921        /g option in pcretest when the pattern contains \K, which makes it possible
922        to have an empty string match not at the start, even when the pattern is
923        anchored. Updated pcretest and pcredemo to use this option.
924    
925    21. If the maximum number of capturing subpatterns in a recursion was greater
926        than the maximum at the outer level, the higher number was returned, but
927        with unset values at the outer level. The correct (outer level) value is
928        now given.
929    
930    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
931        PCRE did not set those parentheses (unlike Perl). I have now found a way to
932        make it do so. The string so far is captured, making this feature
933        compatible with Perl.
934    
935    23. The tests have been re-organized, adding tests 11 and 12, to make it
936        possible to check the Perl 5.10 features against Perl 5.10.
937    
938    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
939        pattern matches a fixed length string. PCRE did not allow this; now it
940        does. Neither allows recursion.
941    
942    25. I finally figured out how to implement a request to provide the minimum
943        length of subject string that was needed in order to match a given pattern.
944        (It was back references and recursion that I had previously got hung up
945        on.) This code has now been added to pcre_study(); it finds a lower bound
946        to the length of subject needed. It is not necessarily the greatest lower
947        bound, but using it to avoid searching strings that are too short does give
948        some useful speed-ups. The value is available to calling programs via
949        pcre_fullinfo().
950    
951    26. While implementing 25, I discovered to my embarrassment that pcretest had
952        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
953        study optimizations had never been tested with that matching function.
954        Oops. What is worse, even when it was passed study data, there was a bug in
955        pcre_dfa_exec() that meant it never actually used it. Double oops. There
956        were also very few tests of studied patterns with pcre_dfa_exec().
957    
958    27. If (?| is used to create subpatterns with duplicate numbers, they are now
959        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
960        on the other side of the coin, they are no longer allowed to have different
961        names, because these cannot be distinguished in PCRE, and this has caused
962        confusion. (This is a difference from Perl.)
963    
964    28. When duplicate subpattern names are present (necessarily with different
965        numbers, as required by 27 above), and a test is made by name in a
966        conditional pattern, either for a subpattern having been matched, or for
967        recursion in such a pattern, all the associated numbered subpatterns are
968        tested, and the overall condition is true if the condition is true for any
969        one of them. This is the way Perl works, and is also more like the way
970        testing by number works.
971    
972    
973    Version 7.9 11-Apr-09
974    ---------------------
975    
976    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
977        (pcretest), all targets were linked against these libraries. This included
978        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
979        libraries. This caused unwanted dependencies to be created. This problem
980        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
981        pcretest is linked with readline.
982    
983    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
984        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
985        moved outside it again, because FALSE and TRUE are already defined in AIX,
986        but BOOL is not.
987    
988    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
989        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
990    
991    4.  The pcregrep documentation said spaces were inserted as well as colons (or
992        hyphens) following file names and line numbers when outputting matching
993        lines. This is not true; no spaces are inserted. I have also clarified the
994        wording for the --colour (or --color) option.
995    
996    5.  In pcregrep, when --colour was used with -o, the list of matching strings
997        was not coloured; this is different to GNU grep, so I have changed it to be
998        the same.
999    
1000    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1001        each matching line was coloured. Now it goes on to look for further matches
1002        of any of the test patterns, which is the same behaviour as GNU grep.
1003    
1004    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1005        doesn't make sense to accept an empty string match in pcregrep, so I have
1006        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1007        seems to be how GNU grep behaves.
1008    
1009    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1010        start or after a newline", because the conditional assertion was not being
1011        correctly handled. The rule now is that both the assertion and what follows
1012        in the first alternative must satisfy the test.
1013    
1014    9.  If auto-callout was enabled in a pattern with a conditional group whose
1015        condition was an assertion, PCRE could crash during matching, both with
1016        pcre_exec() and pcre_dfa_exec().
1017    
1018    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1019        used for matching.
1020    
1021    11. Unicode property support in character classes was not working for
1022        characters (bytes) greater than 127 when not in UTF-8 mode.
1023    
1024    12. Added the -M command line option to pcretest.
1025    
1026    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1027    
1028    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1029    
1030    16. Added comments and documentation about mis-use of no_arg in the C++
1031        wrapper.
1032    
1033    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1034        from Martin Jerabek that uses macro names for all relevant character and
1035        string constants.
1036    
1037    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1038        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1039        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1040        these, but not everybody uses configure.
1041    
1042    19. A conditional group that had only one branch was not being correctly
1043        recognized as an item that could match an empty string. This meant that an
1044        enclosing group might also not be so recognized, causing infinite looping
1045        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1046        with the subject "ab", where knowledge that the repeated group can match
1047        nothing is needed in order to break the loop.
1048    
1049    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1050        exec(), but without supplying a callout function, matching went wrong.
1051    
1052    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1053        leak if the size of the offset vector was greater than 30. When the vector
1054        is smaller, the saved offsets during recursion go onto a local stack
1055        vector, but for larger vectors malloc() is used. It was failing to free
1056        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1057        error, in fact).
1058    
1059    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1060        heapframe that is used only when UTF-8 support is enabled. This caused no
1061        problem, but was untidy.
1062    
1063    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1064        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1065        included within another project.
1066    
1067    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1068        slightly modified by me:
1069    
1070          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1071              not building pcregrep.
1072    
1073          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1074              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1075    
1076    25. Forward references, both numeric and by name, in patterns that made use of
1077        duplicate group numbers, could behave incorrectly or give incorrect errors,
1078        because when scanning forward to find the reference group, PCRE was not
1079        taking into account the duplicate group numbers. A pattern such as
1080        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1081    
1082    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1083        the feature of a custom pointer more persuasive (as requested by a user).
1084    
1085    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1086        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1087        pre-defined.
1088    
1089    28. Added support for (*UTF8) at the start of a pattern.
1090    
1091    29. Arrange for flags added by the "release type" setting in CMake to be shown
1092        in the configuration summary.
1093    
1094    
1095    Version 7.8 05-Sep-08
1096    ---------------------
1097    
1098    1.  Replaced UCP searching code with optimized version as implemented for Ad
1099        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1100        stage table and inline lookup instead of a function, giving speed ups of 2
1101        to 5 times on some simple patterns that I tested. Permission was given to
1102        distribute the MultiStage2.py script that generates the tables (it's not in
1103        the tarball, but is in the Subversion repository).
1104    
1105    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1106        scripts.
1107    
1108    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1109        a group with a zero qualifier. The result of the study could be incorrect,
1110        or the function might crash, depending on the pattern.
1111    
1112    4.  Caseless matching was not working for non-ASCII characters in back
1113        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1114        It now works when Unicode Property Support is available.
1115    
1116    5.  In pcretest, an escape such as \x{de} in the data was always generating
1117        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1118        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1119        truncation.
1120    
1121    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1122    
1123    7.  Added two (int) casts to pcregrep when printing the difference of two
1124        pointers, in case they are 64-bit values.
1125    
1126    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1127        test 2 if it fails.
1128    
1129    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1130        and a #define of that name to empty if it is not externally set. This is to
1131        allow users of MSVC to set it if necessary.
1132    
1133    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1134        the convenience functions in the pcre_get.c source file.
1135    
1136    11. An option change at the start of a pattern that had top-level alternatives
1137        could cause overwriting and/or a crash. This command provoked a crash in
1138        some environments:
1139    
1140          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1141    
1142        This potential security problem was recorded as CVE-2008-2371.
1143    
1144    12. For a pattern where the match had to start at the beginning or immediately
1145        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1146        pcre_dfa_exec() could read past the end of the passed subject if there was
1147        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1148        pcretest so that it places the subject at the end of its malloc-ed buffer.
1149    
1150    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1151        exec() might read past the end of the data buffer in UTF-8 mode.
1152    
1153    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1154        the data contained the byte 0x85 as part of a UTF-8 character within its
1155        first line. This applied both to normal and DFA matching.
1156    
1157    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1158        /^[^d]*?$/8 failed to match "abc".
1159    
1160    16. Added a missing copyright notice to pcrecpp_internal.h.
1161    
1162    17. Make it more clear in the documentation that values returned from
1163        pcre_exec() in ovector are byte offsets, not character counts.
1164    
1165    18. Tidied a few places to stop certain compilers from issuing warnings.
1166    
1167    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1168        supplied by Stefan Weber. I made a further small update for 7.8 because
1169        there is a change of source arrangements: the pcre_searchfuncs.c module is
1170        replaced by pcre_ucd.c.
1171    
1172    
1173    Version 7.7 07-May-08
1174    ---------------------
1175    
1176    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1177        a string to a long long, pretend we don't even have a long long." This is
1178        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1179    
1180    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1181        pre-7.6 versions, which defined a global no_arg variable instead of putting
1182        it in the RE class. (See also #8 below.)
1183    
1184    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1185        Lopes.
1186    
1187    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1188    
1189        (1) The include/exclude patterns were being applied to the whole pathnames
1190            of files, instead of just to the final components.
1191    
1192        (2) If there was more than one level of directory, the subdirectories were
1193            skipped unless they satisfied the include/exclude conditions. This is
1194            inconsistent with GNU grep (and could even be seen as contrary to the
1195            pcregrep specification - which I improved to make it absolutely clear).
1196            The action now is always to scan all levels of directory, and just
1197            apply the include/exclude patterns to regular files.
1198    
1199    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1200        --exclude_dir in the tests to avoid scanning .svn directories.
1201    
1202    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1203        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1204        doesn't support NULs in patterns.
1205    
1206    7.  Added some missing "const"s to declarations of static tables in
1207        pcre_compile.c and pcre_dfa_exec.c.
1208    
1209    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1210        caused by fix #2  above. (Subsequently also a second patch to fix the
1211        first patch. And a third patch - this was a messy problem.)
1212    
1213    9.  Applied Craig's patch to remove the use of push_back().
1214    
1215    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1216        matching function regexec().
1217    
1218    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1219        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1220        references. PCRE supports relative numbers with this syntax (I don't think
1221        Oniguruma does).
1222    
1223    12. Previously, a group with a zero repeat such as (...){0} was completely
1224        omitted from the compiled regex. However, this means that if the group
1225        was called as a subroutine from elsewhere in the pattern, things went wrong
1226        (an internal error was given). Such groups are now left in the compiled
1227        pattern, with a new opcode that causes them to be skipped at execution
1228        time.
1229    
1230    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1231        to the way PCRE behaves:
1232    
1233        (a) A lone ] character is dis-allowed (Perl treats it as data).
1234    
1235        (b) A back reference to an unmatched subpattern matches an empty string
1236            (Perl fails the current match path).
1237    
1238        (c) A data ] in a character class must be notated as \] because if the
1239            first data character in a class is ], it defines an empty class. (In
1240            Perl it is not possible to have an empty class.) The empty class []
1241            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1242            The negative empty class [^] matches any one character, independently
1243            of the DOTALL setting.
1244    
1245    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1246        non-existent subpattern following a character class starting with ']' and
1247        containing () gave an internal compiling error instead of "reference to
1248        non-existent subpattern". Fortunately, when the pattern did exist, the
1249        compiled code was correct. (When scanning forwards to check for the
1250        existencd of the subpattern, it was treating the data ']' as terminating
1251        the class, so got the count wrong. When actually compiling, the reference
1252        was subsequently set up correctly.)
1253    
1254    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1255        it was being rejected as not supported by pcre_dfa_exec(), even though
1256        other assertions are supported. I have made pcre_dfa_exec() support
1257        (*FAIL).
1258    
1259    16. The implementation of 13c above involved the invention of a new opcode,
1260        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1261        cannot be changed at match time, I realized I could make a small
1262        improvement to matching performance by compiling OP_ALLANY instead of
1263        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1264        on the OP_ANY path.
1265    
1266    17. Compiling pcretest on Windows with readline support failed without the
1267        following two fixes: (1) Make the unistd.h include conditional on
1268        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1269    
1270    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1271        ncurses library to be included for pcretest when ReadLine support is
1272        requested, but also to allow for it to be overridden. This patch came from
1273        Daniel Bergström.
1274    
1275    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1276        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1277        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1278        spotting this.
1279    
1280    
1281    Version 7.6 28-Jan-08
1282    ---------------------
1283    
1284    1.  A character class containing a very large number of characters with
1285        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1286        overflow.
1287    
1288    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1289        HAVE_LONG_LONG is not defined.
1290    
1291    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1292        bring them up to date and include new features. This patch includes:
1293    
1294        - Fixed PH's badly added libz and libbz2 support.
1295        - Fixed a problem with static linking.
1296        - Added pcredemo. [But later removed - see 7 below.]
1297        - Fixed dftables problem and added an option.
1298        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1299            HAVE_LONG_LONG.
1300        - Added readline support for pcretest.
1301        - Added an listing of the option settings after cmake has run.
1302    
1303    4.  A user submitted a patch to Makefile that makes it easy to create
1304        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1305        Makefile.am that cause it to include this special target, without
1306        affecting anything else. Note that the same mingw target plus all
1307        the other distribution libraries and programs are now supported
1308        when configuring with CMake (see 6 below) instead of with
1309        Configure/Make.
1310    
1311    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1312        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1313        exported in the Windows port". It has not yet been confirmed that the patch
1314        solves the problem, but it does no harm.
1315    
1316    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1317        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1318        with CMake, and also correct the comment about stack recursion.
1319    
1320    7.  Remove the automatic building of pcredemo from the ./configure system and
1321        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1322        of a program that users should build themselves after PCRE is installed, so
1323        building it automatically is not really right. What is more, it gave
1324        trouble in some build environments.
1325    
1326    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1327    
1328    
1329    Version 7.5 10-Jan-08
1330    ---------------------
1331    
1332    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1333        values in parens when parsing an RE using the C++ wrapper."
1334    
1335    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1336        Characters greater than 255 were excluded from the class instead of being
1337        included.
1338    
1339    3.  The same bug as (2) above applied to negated POSIX classes such as
1340        [:^space:].
1341    
1342    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1343        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1344        I have changed it.
1345    
1346    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1347        first named subpattern) and a construct such as (?&a) would reference the
1348        first named subpattern whose name started with "a" (in other words, the
1349        length check was missing). Both these problems are fixed. "Subpattern name
1350        expected" is now given for (?&) (a zero-length name), and this patch also
1351        makes it give the same error for \k'' (previously it complained that that
1352        was a reference to a non-existent subpattern).
1353    
1354    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1355        this is right because (?- can be followed by option settings as well as by
1356        digits. I have, however, made the messages clearer.
1357    
1358    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1359        than the number used in the conditional) now cause a compile-time error.
1360        This is actually not compatible with Perl, which accepts such patterns, but
1361        treats the conditional as always being FALSE (as PCRE used to), but it
1362        seems to me that giving a diagnostic is better.
1363    
1364    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1365        and messages.
1366    
1367    9.  Fix two occurrences of "backslash" in comments that should have been
1368        "backspace".
1369    
1370    10. Remove two redundant lines of code that can never be obeyed (their function
1371        was moved elsewhere).
1372    
1373    11. The program that makes PCRE's Unicode character property table had a bug
1374        which caused it to generate incorrect table entries for sequences of
1375        characters that have the same character type, but are in different scripts.
1376        It amalgamated them into a single range, with the script of the first of
1377        them. In other words, some characters were in the wrong script. There were
1378        thirteen such cases, affecting characters in the following ranges:
1379    
1380          U+002b0 - U+002c1
1381          U+0060c - U+0060d
1382          U+0061e - U+00612
1383          U+0064b - U+0065e
1384          U+0074d - U+0076d
1385          U+01800 - U+01805
1386          U+01d00 - U+01d77
1387          U+01d9b - U+01dbf
1388          U+0200b - U+0200f
1389          U+030fc - U+030fe
1390          U+03260 - U+0327f
1391          U+0fb46 - U+0fbb1
1392          U+10450 - U+1049d
1393    
1394    12. The -o option (show only the matching part of a line) for pcregrep was not
1395        compatible with GNU grep in that, if there was more than one match in a
1396        line, it showed only the first of them. It now behaves in the same way as
1397        GNU grep.
1398    
1399    13. If the -o and -v options were combined for pcregrep, it printed a blank
1400        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1401        does the same. The return code can be used to tell if there were any
1402        non-matching lines.
1403    
1404    14. Added --file-offsets and --line-offsets to pcregrep.
1405    
1406    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1407        infinitely looping recursion. The bug was that positive lookaheads were not
1408        being skipped when checking for a possible empty match (negative lookaheads
1409        and both kinds of lookbehind were skipped).
1410    
1411    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1412        inclusion of <windows.h> to before rather than after the definition of
1413        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1414    
1415    17. Specifying a possessive quantifier with a specific limit for a Unicode
1416        character property caused pcre_compile() to compile bad code, which led at
1417        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1418        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1419        caused the error; without that there was no problem.
1420    
1421    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1422    
1423    19. Added --enable-pcretest-libreadline.
1424    
1425    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1426        RE::GlobalReplace(). As a result, the number of replacements returned was
1427        double what it should be. I removed one of the increments, but Craig sent a
1428        later patch that removed the other one (the right fix) and added unit tests
1429        that check the return values (which was not done before).
1430    
1431    21. Several CMake things:
1432    
1433        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1434            the names libpcre and libpcreposix, not just pcre and pcreposix.
1435    
1436        (2) The above change means that pcretest and pcregrep are now correctly
1437            linked with the newly-built libraries, not previously installed ones.
1438    
1439        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1440    
1441    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1442        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1443        UTF-8 newline character). The key issue is that the pattern starts .*;
1444        this means that the match must be either at the beginning, or after a
1445        newline. The bug was in the code for advancing after a failed match and
1446        checking that the new position followed a newline. It was not taking
1447        account of UTF-8 characters correctly.
1448    
1449    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1450        character classes. PCRE was not treating the sequence [:...:] as a
1451        character class unless the ... were all letters. Perl, however, seems to
1452        allow any characters between [: and :], though of course it rejects as
1453        unknown any "names" that contain non-letters, because all the known class
1454        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1455        for example, whereas PCRE did not - it did not recognize a POSIX character
1456        class. This seemed a bit dangerous, so the code has been changed to be
1457        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1458        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1459        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1460        Perl does, and where it didn't before.
1461    
1462    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1463        Windows environments %n is disabled by default.
1464    
1465    
1466    Version 7.4 21-Sep-07
1467    ---------------------
1468    
1469    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1470        means that a class such as [\s] counted as "explicit reference to CR or
1471        LF". That isn't really right - the whole point of the change was to try to
1472        help when there was an actual mention of one of the two characters. So now
1473        the change happens only if \r or \n (or a literal CR or LF) character is
1474        encountered.
1475    
1476    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1477        of both had grown to the point where there were only 3 bits left.
1478        Fortunately, there was spare space in the data structure, and so I have
1479        moved the internal flags into a new 16-bit field to free up more option
1480        bits.
1481    
1482    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1483        but did not set the internal JCHANGED flag - either of these is enough to
1484        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1485        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1486        start sets both bits.
1487    
1488    4.  Added options (at build time, compile time, exec time) to change \R from
1489        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1490    
1491    5.  doc/pcresyntax.html was missing from the distribution.
1492    
1493    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1494        compatibility, even though it is no longer used.
1495    
1496    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1497        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1498        windows.h file is present (where different names are used). [This was
1499        reversed later after testing - see 16 below.]
1500    
1501    8.  Changed all #include <config.h> to #include "config.h". There were also
1502        some further <pcre.h> cases that I changed to "pcre.h".
1503    
1504    9.  When pcregrep was used with the --colour option, it missed the line ending
1505        sequence off the lines that it output.
1506    
1507    10. It was pointed out to me that arrays of string pointers cause lots of
1508        relocations when a shared library is dynamically loaded. A technique of
1509        using a single long string with a table of offsets can drastically reduce
1510        these. I have refactored PCRE in four places to do this. The result is
1511        dramatic:
1512    
1513          Originally:                          290
1514          After changing UCP table:            187
1515          After changing error message table:   43
1516          After changing table of "verbs"       36
1517          After changing table of Posix names   22
1518    
1519        Thanks to the folks working on Gregex for glib for this insight.
1520    
1521    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1522        unicode-properties was also set.
1523    
1524    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1525    
1526    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1527        checked only for CRLF.
1528    
1529    14. Added casts to pcretest.c to avoid compiler warnings.
1530    
1531    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1532    
1533    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1534        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1535        entirely. This removes changes made in 7 above.
1536    
1537    17. The CMake files have been updated, and there is now more information about
1538        building with CMake in the NON-UNIX-USE document.
1539    
1540    
1541    Version 7.3 28-Aug-07
1542    ---------------------
1543    
1544     1. In the rejigging of the build system that eventually resulted in 7.1, the
1545        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1546        brackets there is not right, since it causes compilers to look for an
1547        installed pcre.h, not the version that is in the source that is being
1548        compiled (which of course may be different). I have changed it back to:
1549    
1550          #include "pcre.h"
1551    
1552        I have a vague recollection that the change was concerned with compiling in
1553        different directories, but in the new build system, that is taken care of
1554        by the VPATH setting the Makefile.
1555    
1556     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1557        when the subject happened to end in the byte 0x85 (e.g. if the last
1558        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1559        characters but of course it shouldn't be taken as a newline when it is part
1560        of another character. The bug was that, for an unlimited repeat of . in
1561        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1562        characters when looking for a newline.
1563    
1564     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1565    
1566     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1567        in debug output.
1568    
1569     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1570        long printing in the pcrecpp unittest when running under MinGW.
1571    
1572     6. ESC_K was left out of the EBCDIC table.
1573    
1574     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1575        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1576        limit also applies to "virtual nesting" when a pattern is recursive, and in
1577        this case 1000 isn't so big. I have been able to remove this limit at the
1578        expense of backing off one optimization in certain circumstances. Normally,
1579        when pcre_exec() would call its internal match() function recursively and
1580        immediately return the result unconditionally, it uses a "tail recursion"
1581        feature to save stack. However, when a subpattern that can match an empty
1582        string has an unlimited repetition quantifier, it no longer makes this
1583        optimization. That gives it a stack frame in which to save the data for
1584        checking that an empty string has been matched. Previously this was taken
1585        from the 1000-entry workspace that had been reserved. So now there is no
1586        explicit limit, but more stack is used.
1587    
1588     8. Applied Daniel's patches to solve problems with the import/export magic
1589        syntax that is required for Windows, and which was going wrong for the
1590        pcreposix and pcrecpp parts of the library. These were overlooked when this
1591        problem was solved for the main library.
1592    
1593     9. There were some crude static tests to avoid integer overflow when computing
1594        the size of patterns that contain repeated groups with explicit upper
1595        limits. As the maximum quantifier is 65535, the maximum group length was
1596        set at 30,000 so that the product of these two numbers did not overflow a
1597        32-bit integer. However, it turns out that people want to use groups that
1598        are longer than 30,000 bytes (though not repeat them that many times).
1599        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1600        made it possible to implement the integer overflow checks in a much more
1601        dynamic way, which I have now done. The artificial limitation on group
1602        length has been removed - we now have only the limit on the total length of
1603        the compiled pattern, which depends on the LINK_SIZE setting.
1604    
1605    10. Fixed a bug in the documentation for get/copy named substring when
1606        duplicate names are permitted. If none of the named substrings are set, the
1607        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1608        empty string.
1609    
1610    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1611        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1612        because the ] is interpreted as the first data character and the
1613        terminating ] is not found. PCRE has been made compatible with Perl in this
1614        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1615        cause memory overwriting.
1616    
1617    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1618        string has been matched (to stop an infinite loop). It was not recognizing
1619        a conditional subpattern that could match an empty string if that
1620        subpattern was within another subpattern. For example, it looped when
1621        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1622        condition was not nested. This bug has been fixed.
1623    
1624    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1625        past the start of the subject in the presence of bytes with the top bit
1626        set, for example "\x8aBCD".
1627    
1628    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1629        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1630    
1631    14. Optimized (?!) to (*FAIL).
1632    
1633    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1634        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1635        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1636        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1637        does: it's just the validity check that is more restrictive.
1638    
1639    16. Inserted checks for integer overflows during escape sequence (backslash)
1640        processing, and also fixed erroneous offset values for syntax errors during
1641        backslash processing.
1642    
1643    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1644        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1645    
1646    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1647        caused an overrun.
1648    
1649    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1650        something other than just ASCII characters) inside a group that had an
1651        unlimited repeat caused a loop at compile time (while checking to see
1652        whether the group could match an empty string).
1653    
1654    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1655        [\P{Any}] did so. (Error in the code for printing property names.)
1656    
1657    21. An orphan \E inside a character class could cause a crash.
1658    
1659    22. A repeated capturing bracket such as (A)? could cause a wild memory
1660        reference during compilation.
1661    
1662    23. There are several functions in pcre_compile() that scan along a compiled
1663        expression for various reasons (e.g. to see if it's fixed length for look
1664        behind). There were bugs in these functions when a repeated \p or \P was
1665        present in the pattern. These operators have additional parameters compared
1666        with \d, etc, and these were not being taken into account when moving along
1667        the compiled data. Specifically:
1668    
1669        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1670            length.
1671    
1672        (b) An item such as \pL+ within a repeated group could cause crashes or
1673            loops.
1674    
1675        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1676            "reference to non-existent subpattern" error.
1677    
1678        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1679    
1680    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1681        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1682    
1683    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1684    
1685    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1686        character were causing crashes (broken optimization).
1687    
1688    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1689        \p or \P) caused a compile-time loop.
1690    
1691    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1692        break. For example, the unstudied pattern [\r\n]A does not match the string
1693        "\r\nA" because change 7.0/46 below moves the current point on by two
1694        characters after failing to match at the start. However, the pattern \nA
1695        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1696        the same is true. There doesn't seem any very clean way out of this, but
1697        what I have chosen to do makes the common cases work: PCRE now takes note
1698        of whether there can be an explicit match for \r or \n anywhere in the
1699        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1700        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1701        pattern has explicit CR or LF references.
1702    
1703    29. Added (*CR) etc for changing newline setting at start of pattern.
1704    
1705    
1706    Version 7.2 19-Jun-07
1707    ---------------------
1708    
1709     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1710        which is apparently normally available under Windows.
1711    
1712     2. Re-jig the pcregrep tests with different newline settings in an attempt
1713        to make them independent of the local environment's newline setting.
1714    
1715     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1716    
1717     4. Some of the "internals" tests were previously cut out when the link size
1718        was not 2, because the output contained actual offsets. The recent new
1719        "Z" feature of pcretest means that these can be cut out, making the tests
1720        usable with all link sizes.
1721    
1722     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1723        stack recursion. This gives a massive performance boost under BSD, but just
1724        a small improvement under Linux. However, it saves one field in the frame
1725        in all cases.
1726    
1727     6. Added more features from the forthcoming Perl 5.10:
1728    
1729        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1730            recursion call. It refers to the nth most recently opened parentheses.
1731    
1732        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1733            to be opened parentheses.
1734    
1735        (c) Conditions that refer to capturing parentheses can be specified
1736            relatively, for example, (?(-2)... or (?(+3)...
1737    
1738        (d) \K resets the start of the current match so that everything before
1739            is not part of it.
1740    
1741        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1742    
1743        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1744            reference syntax.
1745    
1746        (g) (?| introduces a group in which the numbering of parentheses in each
1747            alternative starts with the same number.
1748    
1749        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1750    
1751     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1752        PCRE_INFO_JCHANGED.
1753    
1754     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1755        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1756        for detecting groups that can match an empty string.
1757    
1758     9. A pattern with a very large number of alternatives (more than several
1759        hundred) was running out of internal workspace during the pre-compile
1760        phase, where pcre_compile() figures out how much memory will be needed. A
1761        bit of new cunning has reduced the workspace needed for groups with
1762        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1763        workspace instead of running out of the 4096 that are available.
1764    
1765    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1766    
1767    11. Applied patch from Google to remove an optimization that didn't quite work.
1768        The report of the bug said:
1769    
1770          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1771          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1772          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1773    
1774    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1775        it matched the wrong number of bytes.
1776    
1777    
1778    Version 7.1 24-Apr-07
1779    ---------------------
1780    
1781     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1782        that is more "standard", making use of automake and other Autotools. There
1783        is some re-arrangement of the files and adjustment of comments consequent
1784        on this.
1785    
1786     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1787        for recursive directory scanning broke on some systems because the files
1788        are not scanned in any specific order and on different systems the order
1789        was different. A call to "sort" has been inserted into RunGrepTest for the
1790        approprate test as a short-term fix. In the longer term there may be an
1791        alternative.
1792    
1793     3. I had an email from Eric Raymond about problems translating some of PCRE's
1794        man pages to HTML (despite the fact that I distribute HTML pages, some
1795        people do their own conversions for various reasons). The problems
1796        concerned the use of low-level troff macros .br and .in. I have therefore
1797        removed all such uses from the man pages (some were redundant, some could
1798        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1799        HTML has been updated to handle .nf/.fi and to complain if it encounters
1800        .br or .in.
1801    
1802     4. Updated comments in configure.ac that get placed in config.h.in and also
1803        arranged for config.h to be included in the distribution, with the name
1804        config.h.generic, for the benefit of those who have to compile without
1805        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1806    
1807     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1808        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1809        makevp.bat for the current PCRE, using the additional files
1810        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1811    
1812     6. A Windows user reported a minor discrepancy with test 2, which turned out
1813        to be caused by a trailing space on an input line that had got lost in his
1814        copy. The trailing space was an accident, so I've just removed it.
1815    
1816     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1817        that is needed.
1818    
1819     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1820        as "const" (a) because they are and (b) because it helps the PHP
1821        maintainers who have recently made a script to detect big data structures
1822        in the php code that should be moved to the .rodata section. I remembered
1823        to update Builducptable as well, so it won't revert if ucptable.h is ever
1824        re-created.
1825    
1826     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1827        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1828        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1829        support is not required. This saves 1.5-2K of code, which is important in
1830        some applications.
1831    
1832        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1833        so as not to refer to the tables, even though these functions will never be
1834        called when UTF-8 support is disabled. Otherwise there are problems with a
1835        shared library.
1836    
1837    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1838    
1839        (a) It was defining its arguments as char * instead of void *.
1840    
1841        (b) It was assuming that all moves were upwards in memory; this was true
1842            a long time ago when I wrote it, but is no longer the case.
1843    
1844        The emulated memove() is provided for those environments that have neither
1845        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1846        is clearly not the case, as these two bugs were recently reported.
1847    
1848    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1849        and Detrail to create the HTML documentation, the .txt form of the man
1850        pages, and it removes trailing spaces from listed files. It also creates
1851        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1852        case, it wraps all the #defines with #ifndefs. This script should be run
1853        before "make dist".
1854    
1855    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1856        with Unicode property support.
1857    
1858        (a) For a maximizing quantifier, if the two different cases of the
1859            character were of different lengths in their UTF-8 codings (there are
1860            some cases like this - I found 11), and the matching function had to
1861            back up over a mixture of the two cases, it incorrectly assumed they
1862            were both the same length.
1863    
1864        (b) When PCRE was configured to use the heap rather than the stack for
1865            recursion during matching, it was not correctly preserving the data for
1866            the other case of a UTF-8 character when checking ahead for a match
1867            while processing a minimizing repeat. If the check also involved
1868            matching a wide character, but failed, corruption could cause an
1869            erroneous result when trying to check for a repeat of the original
1870            character.
1871    
1872    13. Some tidying changes to the testing mechanism:
1873    
1874        (a) The RunTest script now detects the internal link size and whether there
1875            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1876            values substituted by "configure". (The RunGrepTest script already did
1877            this for UTF-8.) The configure.ac script no longer substitutes the
1878            relevant variables.
1879    
1880        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1881            with length and offset values. This means that the output is different
1882            for different internal link sizes. Test 2 is skipped for link sizes
1883            other than 2 because of this, bypassing the problem. Unfortunately,
1884            there was also a test in test 3 (the locale tests) that used /B and
1885            failed for link sizes other than 2. Rather than cut the whole test out,
1886            I have added a new /Z option to pcretest that replaces the length and
1887            offset values with spaces. This is now used to make test 3 independent
1888            of link size. (Test 2 will be tidied up later.)
1889    
1890    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1891        segmentation fault instead of returning the appropriate error message.
1892    
1893    15. In multiline mode when the newline sequence was set to "any", the pattern
1894        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1895        This doesn't seem right; it now treats the CRLF combination as the line
1896        ending, and so does not match in that case. It's only a pattern such as ^$
1897        that would hit this one: something like ^ABC$ would have failed after \r
1898        and then tried again after \r\n.
1899    
1900    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1901        in an attempt to make files that differ only in their line terminators
1902        compare equal. This works on Linux.
1903    
1904    17. Under certain error circumstances pcregrep might try to free random memory
1905        as it exited. This is now fixed, thanks to valgrind.
1906    
1907    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1908        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1909        was because its rules for how to advance for /g after matching an empty
1910        string at the end of a line did not allow for this case. They now check for
1911        it specially.
1912    
1913    20. pcretest is supposed to handle patterns and data of any length, by
1914        extending its buffers when necessary. It was getting this wrong when the
1915        buffer for a data line had to be extended.
1916    
1917    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1918        CRLF as a newline sequence.
1919    
1920    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1921        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1922        I have nevertheless tidied it up.
1923    
1924    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1925    
1926    24. Added a man page for pcre-config.
1927    
1928    
1929    Version 7.0 19-Dec-06
1930    ---------------------
1931    
1932     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1933        moving to gcc 4.1.1.
1934    
1935     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1936        sys/time.h, which is documented as needed for this function. It doesn't
1937        seem to matter on Linux, but it showed up on some releases of OS X.
1938    
1939     3. It seems that there are systems where bytes whose values are greater than
1940        127 match isprint() in the "C" locale. The "C" locale should be the
1941        default when a C program starts up. In most systems, only ASCII printing
1942        characters match isprint(). This difference caused the output from pcretest
1943        to vary, making some of the tests fail. I have changed pcretest so that:
1944    
1945        (a) When it is outputting text in the compiled version of a pattern, bytes
1946            other than 32-126 are always shown as hex escapes.
1947    
1948        (b) When it is outputting text that is a matched part of a subject string,
1949            it does the same, unless a different locale has been set for the match
1950            (using the /L modifier). In this case, it uses isprint() to decide.
1951    
1952     4. Fixed a major bug that caused incorrect computation of the amount of memory
1953        required for a compiled pattern when options that changed within the
1954        pattern affected the logic of the preliminary scan that determines the
1955        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1956        that the computed length was too small. The symptoms of this bug were
1957        either the PCRE error "internal error: code overflow" from pcre_compile(),
1958        or a glibc crash with a message such as "pcretest: free(): invalid next
1959        size (fast)". Examples of patterns that provoked this bug (shown in
1960        pcretest format) are:
1961    
1962          /(?-x: )/x
1963          /(?x)(?-x: \s*#\s*)/
1964          /((?i)[\x{c0}])/8
1965          /(?i:[\x{c0}])/8
1966    
1967        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1968        is now done differently.
1969    
1970     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1971        wrapper classes; (b) implement a new function in the C++ scanner that is
1972        more efficient than the old way of doing things because it avoids levels of
1973        recursion in the regex matching; (c) add a paragraph to the documentation
1974        for the FullMatch() function.
1975    
1976     6. The escape sequence \n was being treated as whatever was defined as
1977        "newline". Not only was this contrary to the documentation, which states
1978        that \n is character 10 (hex 0A), but it also went horribly wrong when
1979        "newline" was defined as CRLF. This has been fixed.
1980    
1981     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1982        was being set to -1 for the "end of line" case (supposedly a value that no
1983        character can have). Though this value is never used (the check for end of
1984        line is "zero bytes in current character"), it caused compiler complaints.
1985        I've changed it to 0xffffffff.
1986    
1987     8. In pcre_version.c, the version string was being built by a sequence of
1988        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1989        string (as it is for production releases) called a macro with an empty
1990        argument. The C standard says the result of this is undefined. The gcc
1991        compiler treats it as an empty string (which was what was wanted) but it is
1992        reported that Visual C gives an error. The source has been hacked around to
1993        avoid this problem.
1994    
1995     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1996        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1997        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1998        of them did).
1999    
2000    10. Originally, pcretest opened its input and output without "b"; then I was
2001        told that "b" was needed in some environments, so it was added for release
2002        5.0 to both the input and output. (It makes no difference on Unix-like
2003        systems.) Later I was told that it is wrong for the input on Windows. I've
2004        now abstracted the modes into two macros, to make it easier to fiddle with
2005        them, and removed "b" from the input mode under Windows.
2006    
2007    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2008    
2009    12. Added -help and --help to pcretest as an official way of being reminded
2010        of the options.
2011    
2012    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2013        and pcrecpp.cc because they annoy compilers at high warning levels.
2014    
2015    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2016    
2017    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2018        scripts are not C programs :-) and which was not noticed because it works
2019        on Linux.
2020    
2021    16. pcretest is supposed to handle any length of pattern and data line (as one
2022        line or as a continued sequence of lines) by extending its input buffer if
2023        necessary. This feature was broken for very long pattern lines, leading to
2024        a string of junk being passed to pcre_compile() if the pattern was longer
2025        than about 50K.
2026    
2027    17. I have done a major re-factoring of the way pcre_compile() computes the
2028        amount of memory needed for a compiled pattern. Previously, there was code
2029        that made a preliminary scan of the pattern in order to do this. That was
2030        OK when PCRE was new, but as the facilities have expanded, it has become
2031        harder and harder to keep it in step with the real compile phase, and there
2032        have been a number of bugs (see for example, 4 above). I have now found a
2033        cunning way of running the real compile function in a "fake" mode that
2034        enables it to compute how much memory it would need, while actually only
2035        ever using a few hundred bytes of working memory and without too many
2036        tests of the mode. This should make future maintenance and development
2037        easier. A side effect of this work is that the limit of 200 on the nesting
2038        depth of parentheses has been removed (though this was never a serious
2039        limitation, I suspect). However, there is a downside: pcre_compile() now
2040        runs more slowly than before (30% or more, depending on the pattern). I
2041        hope this isn't a big issue. There is no effect on runtime performance.
2042    
2043    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2044        newline (only possible for the last line of a file) and it was a
2045        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2046    
2047    19. Added additional timing features to pcretest. (1) The -tm option now times
2048        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2049        separate command line item, by a number that specifies the number of
2050        repeats to use when timing. The default is 50000; this gives better
2051        precision, but takes uncomfortably long for very large patterns.
2052    
2053    20. Extended pcre_study() to be more clever in cases where a branch of a
2054        subpattern has no definite first character. For example, (a*|b*)[cd] would
2055        previously give no result from pcre_study(). Now it recognizes that the
2056        first character must be a, b, c, or d.
2057    
2058    21. There was an incorrect error "recursive call could loop indefinitely" if
2059        a subpattern (or the entire pattern) that was being tested for matching an
2060        empty string contained only one non-empty item after a nested subpattern.
2061        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2062        incorrectly, because the \d was being skipped in the check.
2063    
2064    22. The pcretest program now has a new pattern option /B and a command line
2065        option -b, which is equivalent to adding /B to every pattern. This causes
2066        it to show the compiled bytecode, without the additional information that
2067        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2068        is the same as /B/I).
2069    
2070    23. A new optimization is now able automatically to treat some sequences such
2071        as a*b as a*+b. More specifically, if something simple (such as a character
2072        or a simple class like \d) has an unlimited quantifier, and is followed by
2073        something that cannot possibly match the quantified thing, the quantifier
2074        is automatically "possessified".
2075    
2076    24. A recursive reference to a subpattern whose number was greater than 39
2077        went wrong under certain circumstances in UTF-8 mode. This bug could also
2078        have affected the operation of pcre_study().
2079    
2080    25. Realized that a little bit of performance could be had by replacing
2081        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2082    
2083    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2084    
2085    27. Possessive quantifiers such as a++ were previously implemented by turning
2086        them into atomic groups such as ($>a+). Now they have their own opcodes,
2087        which improves performance. This includes the automatically created ones
2088        from 23 above.
2089    
2090    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2091        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2092        the first matched character to be a colon. This applied both to named and
2093        numbered groups.
2094    
2095    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2096    
2097    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2098        building PCRE on Windows easier, so I have included it in the distribution.
2099    
2100    31. There is now a check in pcretest against a ridiculously large number being
2101        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2102        loop, the loop is abandoned.
2103    
2104    32. Forward references to subpatterns in conditions such as (?(2)...) where
2105        subpattern 2 is defined later cause pcre_compile() to search forwards in
2106        the pattern for the relevant set of parentheses. This search went wrong
2107        when there were unescaped parentheses in a character class, parentheses
2108        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2109    
2110    33. "Subroutine" calls and backreferences were previously restricted to
2111        referencing subpatterns earlier in the regex. This restriction has now
2112        been removed.
2113    
2114    34. Added a number of extra features that are going to be in Perl 5.10. On the
2115        whole, these are just syntactic alternatives for features that PCRE had
2116        previously implemented using the Python syntax or my own invention. The
2117        other formats are all retained for compatibility.
2118    
2119        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2120            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2121            also .NET compatible.
2122    
2123        (b) A recursion or subroutine call to a named group can now be defined as
2124            (?&name) as well as (?P>name).
2125    
2126        (c) A backreference to a named group can now be defined as \k<name> or
2127            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2128            5.10, are also .NET compatible.
2129    
2130        (d) A conditional reference to a named group can now use the syntax
2131            (?(<name>) or (?('name') as well as (?(name).
2132    
2133        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2134            groups (named and numbered) that are never evaluated inline, but can be
2135            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2136            is always false. There may be only one alternative in such a group.
2137    
2138        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2139            as the simple (?(R). The condition is true only if the most recent
2140            recursion is that of the given number or name. It does not search out
2141            through the entire recursion stack.
2142    
2143        (g) The escape \gN or \g{N} has been added, where N is a positive or
2144            negative number, specifying an absolute or relative reference.
2145    
2146    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2147        some "unreachable code" warnings.
2148    
2149    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2150        things, this adds five new scripts.
2151    
2152    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2153        There were also incompatibilities regarding the handling of \Q..\E inside
2154        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2155        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2156    
2157    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2158        matches an empty string, and forcibly breaks the loop. There were bugs in
2159        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2160        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2161        separate and independent bugs (that affected different cases) have been
2162        fixed.
2163    
2164    39. Refactored the code to abolish the use of different opcodes for small
2165        capturing bracket numbers. This is a tidy that I avoided doing when I
2166        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2167        The new approach is not only tidier, it makes it possible to reduce the
2168        memory needed to fix the previous bug (38).
2169    
2170    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2171        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2172        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2173        mode.
2174    
2175    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2176        report.
2177    
2178    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2179        copy construction and assignment in the C++ wrapper.
2180    
2181    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2182        couple of bugs that could have given wrong results in the "--newline=crlf"
2183        case.
2184    
2185    44. Added a number of casts and did some reorganization of signed/unsigned int
2186        variables following suggestions from Dair Grant. Also renamed the variable
2187        "this" as "item" because it is a C++ keyword.
2188    
2189    45. Arranged for dftables to add
2190    
2191          #include "pcre_internal.h"
2192    
2193        to pcre_chartables.c because without it, gcc 4.x may remove the array
2194        definition from the final binary if PCRE is built into a static library and
2195        dead code stripping is activated.
2196    
2197    46. For an unanchored pattern, if a match attempt fails at the start of a
2198        newline sequence, and the newline setting is CRLF or ANY, and the next two
2199        characters are CRLF, advance by two characters instead of one.
2200    
2201    
2202    Version 6.7 04-Jul-06
2203    ---------------------
2204    
2205     1. In order to handle tests when input lines are enormously long, pcretest has
2206        been re-factored so that it automatically extends its buffers when
2207        necessary. The code is crude, but this _is_ just a test program. The
2208        default size has been increased from 32K to 50K.
2209    
2210     2. The code in pcre_study() was using the value of the re argument before
2211        testing it for NULL. (Of course, in any sensible call of the function, it
2212        won't be NULL.)
2213    
2214     3. The memmove() emulation function in pcre_internal.h, which is used on
2215        systems that lack both memmove() and bcopy() - that is, hardly ever -
2216        was missing a "static" storage class specifier.
2217    
2218     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2219        containing an extended class (one that cannot be represented by a bitmap
2220        because it contains high-valued characters or Unicode property items, e.g.
2221        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2222        pattern, but PCRE should not loop if you do not (it no longer does).
2223        [Detail: two cases were found: (a) a repeated subpattern containing an
2224        extended class; (b) a recursive reference to a subpattern that followed a
2225        previous extended class. It wasn't skipping over the extended class
2226        correctly when UTF-8 mode was not set.]
2227    
2228     5. A negated single-character class was not being recognized as fixed-length
2229        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2230        compile error "lookbehind assertion is not fixed length".
2231    
2232     6. The RunPerlTest auxiliary script was showing an unexpected difference
2233        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2234        write a Perl script that can interpret lines of an input file either as
2235        byte characters or as UTF-8, which is what "perltest" was being required to
2236        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2237        can't do is switch easily at run time between having the "use utf8;" pragma
2238        or not. In the end, I fudged it by using the RunPerlTest script to insert
2239        "use utf8;" explicitly for the UTF-8 tests.
2240    
2241     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2242        the end of the subject string, contrary to the documentation and to what
2243        Perl does. This was true of both matching functions. Now it matches only at
2244        the start of the subject and immediately after *internal* newlines.
2245    
2246     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2247        a pointer to an int instead of a pointer to an unsigned long int. This
2248        caused problems on 64-bit systems.
2249    
2250     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2251        instance of the 'standard' template library not being so standard".
2252    
2253    10. There was no check on the number of named subpatterns nor the maximum
2254        length of a subpattern name. The product of these values is used to compute
2255        the size of the memory block for a compiled pattern. By supplying a very
2256        long subpattern name and a large number of named subpatterns, the size
2257        computation could be caused to overflow. This is now prevented by limiting
2258        the length of names to 32 characters, and the number of named subpatterns
2259        to 10,000.
2260    
2261    11. Subpatterns that are repeated with specific counts have to be replicated in
2262        the compiled pattern. The size of memory for this was computed from the
2263        length of the subpattern and the repeat count. The latter is limited to
2264        65535, but there was no limit on the former, meaning that integer overflow
2265        could in principle occur. The compiled length of a repeated subpattern is
2266        now limited to 30,000 bytes in order to prevent this.
2267    
2268    12. Added the optional facility to have named substrings with the same name.
2269    
2270    13. Added the ability to use a named substring as a condition, using the
2271        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2272        are numbers (not recommended). Forward references are permitted.
2273    
2274    14. Added forward references in named backreferences (if you see what I mean).
2275    
2276    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2277        pattern could run off the end of the subject. For example, the pattern
2278        "(?s)(.{1,5})"8 did this with the subject "ab".
2279    
2280    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2281        PCRE_CASELESS was set when matching characters that were quantified with ?
2282        or *.
2283    
2284    17. A character class other than a single negated character that had a minimum
2285        but no maximum quantifier - for example [ab]{6,} - was not handled
2286        correctly by pce_dfa_exec(). It would match only one character.
2287    
2288    18. A valid (though odd) pattern that looked like a POSIX character
2289        class but used an invalid character after [ (for example [[,abc,]]) caused
2290        pcre_compile() to give the error "Failed: internal error: code overflow" or
2291        in some cases to crash with a glibc free() error. This could even happen if
2292        the pattern terminated after [[ but there just happened to be a sequence of
2293        letters, a binary zero, and a closing ] in the memory that followed.
2294    
2295    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2296        over the years. Originally (before any Unicode support), just the bottom 8
2297        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2298        output from "man perlunicode" includes this:
2299    
2300          The regular expression compiler produces polymorphic opcodes.  That
2301          is, the pattern adapts to the data and automatically switches to
2302          the Unicode character scheme when presented with Unicode data--or
2303          instead uses a traditional byte scheme when presented with byte
2304          data.
2305    
2306        Sadly, a wide octal escape does not cause a switch, and in a string with
2307        no other multibyte characters, these octal escapes are treated as before.
2308        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2309        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2310        Unicode string.
2311    
2312        I have not perpetrated such confusion in PCRE. Up till now, it took just
2313        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2314        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2315        translate to the appropriate multibyte character.
2316    
2317    29. Applied some refactoring to reduce the number of warnings from Microsoft
2318        and Borland compilers. This has included removing the fudge introduced
2319        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2320        a warning about an unused variable.
2321    
2322    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2323        characters since release 4.0, because Perl (from release 5.004) does not.
2324        [Or at least, is documented not to: some releases seem to be in conflict
2325        with the documentation.] However, when a pattern was studied with
2326        pcre_study() and all its branches started with \s, PCRE still included VT
2327        as a possible starting character. Of course, this did no harm; it just
2328        caused an unnecessary match attempt.
2329    
2330    22. Removed a now-redundant internal flag bit that recorded the fact that case
2331        dependency changed within the pattern. This was once needed for "required
2332        byte" processing, but is no longer used. This recovers a now-scarce options
2333        bit. Also moved the least significant internal flag bit to the most-
2334        significant bit of the word, which was not previously used (hangover from
2335        the days when it was an int rather than a uint) to free up another bit for
2336        the future.
2337    
2338    23. Added support for CRLF line endings as well as CR and LF. As well as the
2339        default being selectable at build time, it can now be changed at runtime
2340        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2341        specify that it is scanning data with non-default line endings.
2342    
2343    24. Changed the definition of CXXLINK to make it agree with the definition of
2344        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2345    
2346    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2347        recursions. This makes a big different to stack usage for some patterns.
2348    
2349    26. If a subpattern containing a named recursion or subroutine reference such
2350        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2351        the space required for the compiled pattern went wrong and gave too small a
2352        value. Depending on the environment, this could lead to "Failed: internal
2353        error: code overflow at offset 49" or "glibc detected double free or
2354        corruption" errors.
2355    
2356    27. Applied patches from Google (a) to support the new newline modes and (b) to
2357        advance over multibyte UTF-8 characters in GlobalReplace.
2358    
2359    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2360        difference for some implementation of PCRE in some Windows version.
2361    
2362    29. Added some extra testing facilities to pcretest:
2363    
2364        \q<number>   in a data line sets the "match limit" value
2365        \Q<number>   in a data line sets the "match recursion limt" value
2366        -S <number>  sets the stack size, where <number> is in megabytes
2367    
2368        The -S option isn't available for Windows.
2369    
2370    
2371    Version 6.6 06-Feb-06
2372    ---------------------
2373    
2374     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2375        in pcreposix.h. I have copied the definition from pcre.h.
2376    
2377     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2378        because pcre.h is no longer a built file.
2379    
2380     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2381        not normally included in the compiled code.
2382    
2383    
2384    Version 6.5 01-Feb-06
2385    ---------------------
2386    
2387     1. When using the partial match feature with pcre_dfa_exec(), it was not
2388        anchoring the second and subsequent partial matches at the new starting
2389        point. This could lead to incorrect results. For example, with the pattern
2390        /1234/, partially matching against "123" and then "a4" gave a match.
2391    
2392     2. Changes to pcregrep:
2393    
2394        (a) All non-match returns from pcre_exec() were being treated as failures
2395            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2396            error message is output. Some extra information is given for the
2397            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2398            probably the only errors that are likely to be caused by users (by
2399            specifying a regex that has nested indefinite repeats, for instance).
2400            If there are more than 20 of these errors, pcregrep is abandoned.
2401    
2402        (b) A binary zero was treated as data while matching, but terminated the
2403            output line if it was written out. This has been fixed: binary zeroes
2404            are now no different to any other data bytes.
2405    
2406        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2407            used to set a locale for matching. The --locale=xxxx long option has
2408            been added (no short equivalent) to specify a locale explicitly on the
2409            pcregrep command, overriding the environment variables.
2410    
2411        (d) When -B was used with -n, some line numbers in the output were one less
2412            than they should have been.
2413    
2414        (e) Added the -o (--only-matching) option.
2415    
2416        (f) If -A or -C was used with -c (count only), some lines of context were
2417            accidentally printed for the final match.
2418    
2419        (g) Added the -H (--with-filename) option.
2420    
2421        (h) The combination of options -rh failed to suppress file names for files
2422            that were found from directory arguments.
2423    
2424        (i) Added the -D (--devices) and -d (--directories) options.
2425    
2426        (j) Added the -F (--fixed-strings) option.
2427    
2428        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2429    
2430        (l) Added the --colo(u)r option.
2431    
2432        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2433            is not present by default.
2434    
2435     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2436        items such as (?R) or (?1), when the recursion could match a number of
2437        alternatives. If it matched one of the alternatives, but subsequently,
2438        outside the recursion, there was a failure, the code tried to back up into
2439        the recursion. However, because of the way PCRE is implemented, this is not
2440        possible, and the result was an incorrect result from the match.
2441    
2442        In order to prevent this happening, the specification of recursion has
2443        been changed so that all such subpatterns are automatically treated as
2444        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2445    
2446     4. I had overlooked the fact that, in some locales, there are characters for
2447        which isalpha() is true but neither isupper() nor islower() are true. In
2448        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2449        and ordfeminine) are like this. This affected the treatment of \w and \W
2450        when they appeared in character classes, but not when they appeared outside
2451        a character class. The bit map for "word" characters is now created
2452        separately from the results of isalnum() instead of just taking it from the
2453        upper, lower, and digit maps. (Plus the underscore character, of course.)
2454    
2455     5. The above bug also affected the handling of POSIX character classes such as
2456        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2457        permanent tables. Instead, the bit maps for such a class were previously
2458        created as the appropriate unions of the upper, lower, and digit bitmaps.
2459        Now they are created by subtraction from the [[:word:]] class, which has
2460        its own bitmap.
2461    
2462     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2463        It is created by subtracting the vertical space characters (\x09, \x0a,
2464        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2465        subtraction was done in the overall bitmap for a character class, meaning
2466        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2467        be recognized. This bug has been fixed.
2468    
2469     7. Patches from the folks at Google:
2470    
2471          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2472          real life, but is still worth protecting against".
2473    
2474          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2475          regular expressions".
2476    
2477          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2478          have it.
2479    
2480          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2481          "configure" and the latter not, in order to fix a problem somebody had
2482          with compiling the Arg class on HP-UX.
2483    
2484          (e) Improve the error-handling of the C++ wrapper a little bit.
2485    
2486          (f) New tests for checking recursion limiting.
2487    
2488     8. The pcre_memmove() function, which is used only if the environment does not
2489        have a standard memmove() function (and is therefore rarely compiled),
2490        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2491        returning a result (though PCRE never actually uses the result).
2492    
2493     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2494        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2495        returned instead of calling malloc() with an overflowing number that would
2496        most likely cause subsequent chaos.
2497    
2498    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2499    
2500    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2501        with this option is matched, the nmatch and pmatch options of regexec() are
2502        ignored.
2503    
2504    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2505        provided in case anyone wants to the the POSIX interface with UTF-8
2506        strings.
2507    
2508    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2509        C++ linking (needed for some HP-UX environments).
2510    
2511    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2512        (unused parameter) and in the pcre_printint() function (omitted "default"
2513        switch label when the default is to do nothing).
2514    
2515    15. Added some code to make it possible, when PCRE is compiled as a C++
2516        library, to replace subject pointers for pcre_exec() with a smart pointer
2517        class, thus making it possible to process discontinuous strings.
2518    
2519    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2520        much the same function. They were added by different people who were trying
2521        to make PCRE easy to compile on non-Unix systems. It has been suggested
2522        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2523        for compiling on Windows systems. I have therefore replaced it with
2524        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2525        defaults to "extern" for C or "extern C" for C++, which works fine on
2526        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2527        SCOPE with something explicit in config.h. In addition:
2528    
2529        (a) pcreposix.h still had just "extern" instead of either of these macros;
2530            I have replaced it with PCRE_DATA_SCOPE.
2531    
2532        (b) Functions such as _pcre_xclass(), which are internal to the library,
2533            but external in the C sense, all had PCRE_EXPORT in their definitions.
2534            This is apparently wrong for the Windows case, so I have removed it.
2535            (It makes no difference on Unix-like systems.)
2536    
2537    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2538        of recursive calls to match(). This is different to MATCH_LIMIT because
2539        that limits the total number of calls to match(), not all of which increase
2540        the depth of recursion. Limiting the recursion depth limits the amount of
2541        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2542        when PCRE is compiled, and changed at run time. A patch from Google adds
2543        this functionality to the C++ interface.
2544    
2545    18. Changes to the handling of Unicode character properties:
2546    
2547        (a) Updated the table to Unicode 4.1.0.
2548    
2549        (b) Recognize characters that are not in the table as "Cn" (undefined).
2550    
2551        (c) I revised the way the table is implemented to a much improved format
2552            which includes recognition of ranges. It now supports the ranges that
2553            are defined in UnicodeData.txt, and it also amalgamates other
2554            characters into ranges. This has reduced the number of entries in the
2555            table from around 16,000 to around 3,000, thus reducing its size
2556            considerably. I realized I did not need to use a tree structure after
2557            all - a binary chop search is just as efficient. Having reduced the
2558            number of entries, I extended their size from 6 bytes to 8 bytes to
2559            allow for more data.
2560    
2561        (d) Added support for Unicode script names via properties such as \p{Han}.
2562    
2563    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2564        matching that character.
2565    
2566    20. When matching a repeated Unicode property with a minimum greater than zero,
2567        (for example \pL{2,}), PCRE could look past the end of the subject if it
2568        reached it while seeking the minimum number of characters. This could
2569        happen only if some of the characters were more than one byte long, because
2570        there is a check for at least the minimum number of bytes.
2571    
2572    21. Refactored the implementation of \p and \P so as to be more general, to
2573        allow for more different types of property in future. This has changed the
2574        compiled form incompatibly. Anybody with saved compiled patterns that use
2575        \p or \P will have to recompile them.
2576    
2577    22. Added "Any" and "L&" to the supported property types.
2578    
2579    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2580        but give a compile time error if the value is greater than 0xff.
2581    
2582    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2583        accidentally not being installed or uninstalled.
2584    
2585    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2586        made were to insert the current release number. This seemed silly, because
2587        it made things harder for people building PCRE on systems that don't run
2588        "configure". I have turned pcre.h into a distributed file, no longer built
2589        by "configure", with the version identification directly included. There is
2590        no longer a pcre.h.in file.
2591    
2592        However, this change necessitated a change to the pcre-config script as
2593        well. It is built from pcre-config.in, and one of the substitutions was the
2594        release number. I have updated configure.ac so that ./configure now finds
2595        the release number by grepping pcre.h.
2596    
2597    26. Added the ability to run the tests under valgrind.
2598    
2599    
2600    Version 6.4 05-Sep-05
2601    ---------------------
2602    
2603     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2604        "--" to be printed when multiple files were scanned, even when none of the
2605        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2606        consider it to be a bug, and have restored the previous behaviour.
2607    
2608     2. A couple of code tidies to get rid of compiler warnings.
2609    
2610     3. The pcretest program used to cheat by referring to symbols in the library
2611        whose names begin with _pcre_. These are internal symbols that are not
2612        really supposed to be visible externally, and in some environments it is
2613        possible to suppress them. The cheating is now confined to including
2614        certain files from the library's source, which is a bit cleaner.
2615    
2616     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2617        file's purpose clearer.
2618    
2619     5. Reorganized pcre_ucp_findchar().
2620    
2621    
2622    Version 6.3 15-Aug-05
2623    ---------------------
2624    
2625     1. The file libpcre.pc.in did not have general read permission in the tarball.
2626    
2627     2. There were some problems when building without C++ support:
2628    
2629        (a) If C++ support was not built, "make install" and "make test" still
2630            tried to test it.
2631    
2632        (b) There were problems when the value of CXX was explicitly set. Some
2633            changes have been made to try to fix these, and ...
2634    
2635        (c) --disable-cpp can now be used to explicitly disable C++ support.
2636    
2637        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2638            backslash in a target when C++ was disabled. This confuses some
2639            versions of "make", apparently. Using an intermediate variable solves
2640            this. (Same for CPP_LOBJ.)
2641    
2642     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2643        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2644        necessary on certain architectures.
2645    
2646     4. Added a setting of -export-symbols-regex to the link command to remove
2647        those symbols that are exported in the C sense, but actually are local
2648        within the library, and not documented. Their names all begin with
2649        "_pcre_". This is not a perfect job, because (a) we have to except some
2650        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2651        available (and never for static libraries). I have made a note to try to
2652        find a way round (a) in the future.
2653    
2654    
2655    Version 6.2 01-Aug-05
2656    ---------------------
2657    
2658     1. There was no test for integer overflow of quantifier values. A construction
2659        such as {1111111111111111} would give undefined results. What is worse, if
2660        a minimum quantifier for a parenthesized subpattern overflowed and became
2661        negative, the calculation of the memory size went wrong. This could have
2662        led to memory overwriting.
2663    
2664     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2665    
2666     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2667        operating environments where this matters.
2668    
2669     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2670        PCRE options from within the C++ wrapper.
2671    
2672     5. Named capturing subpatterns were not being correctly counted when a pattern
2673        was compiled. This caused two problems: (a) If there were more than 100
2674        such subpatterns, the calculation of the memory needed for the whole
2675        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2676        back references of the form \12, where the number was greater than 9, were
2677        not recognized as back references, even though there were sufficient
2678        previous subpatterns.
2679    
2680     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2681        versions of gcc, e.g. 2.95.4.
2682    
2683    
2684    Version 6.1 21-Jun-05
2685    ---------------------
2686    
2687     1. There was one reference to the variable "posix" in pcretest.c that was not
2688        surrounded by "#if !defined NOPOSIX".
2689    
2690     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2691        the cross-check on the old pcre_info() function, for the benefit of the
2692        cut-down version of PCRE that is currently imported into Exim.
2693    
2694     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2695        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2696        patterns that start (?i) but I don't think that matters. The use of (?i) is
2697        just an example; this all applies to the other options as well.
2698    
2699     4. Since libtool seems to echo the compile commands it is issuing, the output
2700        from "make" can be reduced a bit by putting "@" in front of each libtool
2701        compile command.
2702    
2703     5. Patch from the folks at Google for configure.in to be a bit more thorough
2704        in checking for a suitable C++ installation before trying to compile the
2705        C++ stuff. This should fix a reported problem when a compiler was present,
2706        but no suitable headers.
2707    
2708     6. The man pages all had just "PCRE" as their title. I have changed them to
2709        be the relevant file name. I have also arranged that these names are
2710        retained in the file doc/pcre.txt, which is a concatenation in text format
2711        of all the man pages except the little individual ones for each function.
2712    
2713     7. The NON-UNIX-USE file had not been updated for the different set of source
2714        files that come with release 6. I also added a few comments about the C++
2715        wrapper.
2716    
2717    
2718    Version 6.0 07-Jun-05
2719    ---------------------
2720    
2721     1. Some minor internal re-organization to help with my DFA experiments.
2722    
2723     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2724        didn't matter for the library itself when fully configured, but did matter
2725        when compiling without UCP support, or within Exim, where the ucp files are
2726        not imported.
2727    
2728     3. Refactoring of the library code to split up the various functions into
2729        different source modules. The addition of the new DFA matching code (see
2730        below) to a single monolithic source would have made it really too
2731        unwieldy, quite apart from causing all the code to be include in a
2732        statically linked application, when only some functions are used. This is
2733        relevant even without the DFA addition now that patterns can be compiled in
2734        one application and matched in another.
2735    
2736        The downside of splitting up is that there have to be some external
2737        functions and data tables that are used internally in different modules of
2738        the library but which are not part of the API. These have all had their
2739        names changed to start with "_pcre_" so that they are unlikely to clash
2740        with other external names.
2741    
2742     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2743        a different (DFA) algorithm. Although it is slower than the original
2744        function, it does have some advantages for certain types of matching
2745        problem.
2746    
2747     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2748        including restarting after a partial match.
2749    
2750     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2751        defined when compiling for Windows was sent to me. I have put it into the
2752        code, though I have no means of testing or verifying it.
2753    
2754     7. Added the pcre_refcount() auxiliary function.
2755    
2756     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2757        match before or at the first newline in the subject string. In pcretest,
2758        the /f option on a pattern can be used to set this.
2759    
2760     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2761        would behave wrongly. This has been present in PCRE since release 4.0.
2762    
2763    10. A number of changes to the pcregrep command:
2764    
2765        (a) Refactored how -x works; insert ^(...)$ instead of setting
2766            PCRE_ANCHORED and checking the length, in preparation for adding
2767            something similar for -w.
2768    
2769        (b) Added the -w (match as a word) option.
2770    
2771        (c) Refactored the way lines are read and buffered so as to have more
2772            than one at a time available.
2773    
2774        (d) Implemented a pcregrep test script.
2775    
2776        (e) Added the -M (multiline match) option. This allows patterns to match
2777            over several lines of the subject. The buffering ensures that at least
2778            8K, or the rest of the document (whichever is the shorter) is available
2779            for matching (and similarly the previous 8K for lookbehind assertions).
2780    
2781        (f) Changed the --help output so that it now says
2782    
2783              -w, --word-regex(p)
2784    
2785            instead of two lines, one with "regex" and the other with "regexp"
2786            because that confused at least one person since the short forms are the
2787            same. (This required a bit of code, as the output is generated
2788            automatically from a table. It wasn't just a text change.)
2789    
2790        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2791            option but starts with a hyphen. Could be a pattern or a path name
2792            starting with a hyphen, for instance.
2793    
2794        (h) "-" can be given as a file name to represent stdin.
2795    
2796        (i) When file names are being printed, "(standard input)" is used for
2797            the standard input, for compatibility with GNU grep. Previously
2798            "<stdin>" was used.
2799    
2800        (j) The option --label=xxx can be used to supply a name to be used for
2801            stdin when file names are being printed. There is no short form.
2802    
2803        (k) Re-factored the options decoding logic because we are going to add
2804            two more options that take data. Such options can now be given in four
2805            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2806    
2807        (l) Added the -A, -B, and -C options for requesting that lines of context
2808            around matches be printed.
2809    
2810        (m) Added the -L option to print the names of files that do not contain
2811            any matching lines, that is, the complement of -l.
2812    
2813        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2814            continue to scan other files.
2815    
2816        (o) The -s option was incorrectly implemented. For compatibility with other
2817            greps, it now suppresses the error message for a non-existent or non-
2818            accessible file (but not the return code). There is a new option called
2819            -q that suppresses the output of matching lines, which was what -s was
2820            previously doing.
2821    
2822        (p) Added --include and --exclude options to specify files for inclusion
2823            and exclusion when recursing.
2824    
2825    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2826        Hopefully, it now does.
2827    
2828    12. Missing cast in pcre_study().
2829    
2830    13. Added an "uninstall" target to the makefile.
2831    
2832    14. Replaced "extern" in the function prototypes in Makefile.in with
2833        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2834        world, but is set differently for Windows.
2835    
2836    15. Added a second compiling function called pcre_compile2(). The only
2837        difference is that it has an extra argument, which is a pointer to an
2838        integer error code. When there is a compile-time failure, this is set
2839        non-zero, in addition to the error test pointer being set to point to an
2840        error message. The new argument may be NULL if no error number is required
2841        (but then you may as well call pcre_compile(), which is now just a
2842        wrapper). This facility is provided because some applications need a
2843        numeric error indication, but it has also enabled me to tidy up the way
2844        compile-time errors are handled in the POSIX wrapper.
2845    
2846    16. Added VPATH=.libs to the makefile; this should help when building with one
2847        prefix path and installing with another. (Or so I'm told by someone who
2848        knows more about this stuff than I do.)
2849    
2850    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2851        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2852        match everything, including newlines. This is not POSIX-compatible, but
2853        somebody wanted the feature. From pcretest it can be activated by using
2854        both the P and the s flags.
2855    
2856    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2857    
2858    19. libpcre.pc was being incorrectly installed as executable.
2859    
2860    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2861        it now also looks for '\r' so that it will work unmodified on Windows.
2862    
2863    21. Added Google's contributed C++ wrapper to the distribution.
2864    
2865    22. Added some untidy missing memory free() calls in pcretest, to keep
2866        Electric Fence happy when testing.
2867    
2868    
2869    
2870    Version 5.0 13-Sep-04
2871    ---------------------
2872    
2873     1. Internal change: literal characters are no longer packed up into items
2874        containing multiple characters in a single byte-string. Each character
2875        is now matched using a separate opcode. However, there may be more than one
2876        byte in the character in UTF-8 mode.
2877    
2878     2. The pcre_callout_block structure has two new fields: pattern_position and
2879        next_item_length. These contain the offset in the pattern to the next match
2880        item, and its length, respectively.
2881    
2882     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2883        insertion of callouts before each pattern item. Added the /C option to
2884        pcretest to make use of this.
2885    
2886     4. On the advice of a Windows user, the lines
2887    
2888          #if defined(_WIN32) || defined(WIN32)
2889          _setmode( _fileno( stdout ), 0x8000 );
2890          #endif  /* defined(_WIN32) || defined(WIN32) */
2891    
2892        have been added to the source of pcretest. This apparently does useful
2893        magic in relation to line terminators.
2894    
2895     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2896        for the benefit of those environments where the "b" makes a difference.
2897    
2898     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2899        to know about it. I have put a hack into configure.in that adds in code
2900        to set GCC=yes if CC=icc. This seems to end up at a point in the
2901        generated configure script that is early enough to affect the setting of
2902        compiler options, which is what is needed, but I have no means of testing
2903        whether it really works. (The user who reported this had patched the
2904        generated configure script, which of course I cannot do.)
2905    
2906        LATER: After change 22 below (new libtool files), the configure script
2907        seems to know about icc (and also ecc). Therefore, I have commented out
2908        this hack in configure.in.
2909    
2910     7. Added support for pkg-config (2 patches were sent in).
2911    
2912     8. Negated POSIX character classes that used a combination of internal tables
2913        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2914        [[:^ascii]]. Typically, they would match almost any characters. The other
2915        POSIX classes were not broken in this way.
2916    
2917     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2918        to find the match, as PCRE was deluded into thinking that the match had to
2919        start at the start point or following a newline. The same bug applied to
2920        patterns with negative forward assertions or any backward assertions
2921        preceding ".*" at the start, unless the pattern required a fixed first
2922        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2923    
2924    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2925        starting at the last subject character, bytes beyond the end of the subject
2926        string were read.
2927    
2928    11. Renamed the variable "class" as "classbits" to make life easier for C++
2929        users. (Previously there was a macro definition, but it apparently wasn't
2930        enough.)
2931    
2932    12. Added the new field "tables" to the extra data so that tables can be passed
2933        in at exec time, or the internal tables can be re-selected. This allows
2934        a compiled regex to be saved and re-used at a later time by a different
2935        program that might have everything at different addresses.
2936    
2937    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2938        -R library as well as a -L library.
2939    
2940    14. The debugging options of pcretest (-d on the command line or D on a
2941        pattern) showed incorrect output for anything following an extended class
2942        that contained multibyte characters and which was followed by a quantifier.
2943    
2944    15. Added optional support for general category Unicode character properties
2945        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2946        support. It adds about 90K to the size of the library. The meanings of the
2947        inbuilt class escapes such as \d and \s have NOT been changed.
2948    
2949    16. Updated pcredemo.c to include calls to free() to release the memory for the
2950        compiled pattern.
2951    
2952    17. The generated file chartables.c was being created in the source directory
2953        instead of in the building directory. This caused the build to fail if the
2954        source directory was different from the building directory, and was
2955        read-only.
2956    
2957    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2958        file. No doubt somebody will tell me if they don't make sense... Also added
2959        Dan Mooney's comments about building on OpenVMS.
2960    
2961    19. Added support for partial matching via the PCRE_PARTIAL option for
2962        pcre_exec() and the \P data escape in pcretest.
2963    
2964    20. Extended pcretest with 3 new pattern features:
2965    
2966        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2967              write the compiled pattern to the file whose name is "rest-of-line".
2968              This is a straight binary dump of the data, with the saved pointer to
2969              the character tables forced to be NULL. The study data, if any, is
2970              written too. After writing, pcretest reads a new pattern.
2971    
2972        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2973              compiled pattern from the given file. There must not be any
2974              occurrences of "<" in the file name (pretty unlikely); if there are,
2975              pcretest will instead treat the initial "<" as a pattern delimiter.
2976              After reading in the pattern, pcretest goes on to read data lines as
2977              usual.
2978    
2979        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2980              and 16-bit fields in a compiled pattern, to simulate a pattern that
2981              was compiled on a host of opposite endianness.
2982    
2983    21. The pcre-exec() function can now cope with patterns that were compiled on
2984        hosts of opposite endianness, with this restriction:
2985    
2986          As for any compiled expression that is saved and used later, the tables
2987          pointer field cannot be preserved; the extra_data field in the arguments
2988          to pcre_exec() should be used to pass in a tables address if a value
2989          other than the default internal tables were used at compile time.
2990    
2991    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2992        now diagnosed as an error. Previously, most of the time, a negative number
2993        would have been treated as zero, but if in addition "ovector" was passed as
2994        NULL, a crash could occur.
2995    
2996    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2997        new versions from the libtool 1.5 distribution (the last one is a copy of
2998        a file called libtool.m4). This seems to have fixed the need to patch
2999        "configure" to support Darwin 1.3 (which I used to do). However, I still
3000        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3001        workstation).
3002    
3003    24. Changed the PCRE licence to be the more standard "BSD" licence.
3004    
3005    
3006    Version 4.5 01-Dec-03
3007    ---------------------
3008    
3009     1. There has been some re-arrangement of the code for the match() function so
3010        that it can be compiled in a version that does not call itself recursively.
3011        Instead, it keeps those local variables that need separate instances for
3012        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3013        needs to "recurse". Keeping track of where control must go is done by means
3014        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3015        hide most of the details from the main code, and operates only if
3016        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3017        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3018        operating.
3019    
3020        To make it easier for callers to provide specially tailored get/free
3021        functions for this usage, two new functions, pcre_stack_malloc, and
3022        pcre_stack_free, are used. They are always called in strict stacking order,
3023        and the size of block requested is always the same.
3024    
3025        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3026        PCRE has been compiled to use the stack or the heap for recursion. The
3027        -C option of pcretest uses this to show which version is compiled.
3028    
3029        A new data escape \S, is added to pcretest; it causes the amounts of store
3030        obtained and freed by both kinds of malloc/free at match time to be added
3031        to the output.
3032    
3033     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3034        what's available on my current Linux desktop machine.
3035    
3036     3. When matching a UTF-8 string, the test for a valid string at the start has
3037        been extended. If start_offset is not zero, PCRE now checks that it points
3038        to a byte that is the start of a UTF-8 character. If not, it returns
3039        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3040        this is necessary because there may be backward assertions in the pattern.
3041        When matching the same subject several times, it may save resources to use
3042        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3043    
3044     4. The code for checking the validity of UTF-8 strings has been tightened so
3045        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3046        containing "overlong sequences".
3047    
3048     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3049        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3050        should have been "&", but it just so happened that all the cases this let
3051        through by mistake were picked up later in the function.
3052    
3053     6. I had used a variable called "isblank" - this is a C99 function, causing
3054        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3055    
3056     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3057        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3058        so the version is shown; (c) stop "make test" echoing "./RunTest".
3059    
3060     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3061    
3062     9. The prototype for memmove() for systems that don't have it was using
3063        size_t, but the inclusion of the header that defines size_t was later. I've
3064        moved the #includes for the C headers earlier to avoid this.
3065    
3066    10. Added some adjustments to the code to make it easier to compiler on certain
3067        special systems:
3068    
3069          (a) Some "const" qualifiers were missing.
3070          (b) Added the macro EXPORT before all exported functions; by default this
3071              is defined to be empty.
3072          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3073              that it reads its output file name as an argument instead of writing
3074              to the standard output and assuming this can be redirected.
3075    
3076    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3077        class containing characters with values greater than 255, PCRE compilation
3078        went into a loop.
3079    
3080    12. A recursive reference to a subpattern that was within another subpattern
3081        that had a minimum quantifier of zero caused PCRE to crash. For example,
3082        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3083        recursion. If the recursively-called subpattern itself had a zero repeat,
3084        that was OK.
3085    
3086    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3087        buffer into which it was copied (for escape processing) was still set at
3088        1024, so long lines caused crashes.
3089    
3090    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3091        "internal error: code overflow...". This applied to any character class
3092        that was followed by a possessive quantifier.
3093    
3094    15. Modified the Makefile to add libpcre.la as a prerequisite for
3095        libpcreposix.la because I was told this is needed for a parallel build to
3096        work.
3097    
3098    16. If a pattern that contained .* following optional items at the start was
3099        studied, the wrong optimizing data was generated, leading to matching
3100        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3101        matching string must start with a or b or c. The correct conclusion for
3102        this pattern is that a match can start with any character.
3103    
3104    
3105    Version 4.4 13-Aug-03
3106    ---------------------
3107    
3108     1. In UTF-8 mode, a character class containing characters with values between
3109        127 and 255 was not handled correctly if the compiled pattern was studied.
3110        In fixing this, I have also improved the studying algorithm for such
3111        classes (slightly).
3112    
3113     2. Three internal functions had redundant arguments passed to them. Removal
3114        might give a very teeny performance improvement.
3115    
3116     3. Documentation bug: the value of the capture_top field in a callout is *one
3117        more than* the number of the hightest numbered captured substring.
3118    
3119     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3120        in incorrectly linking with a previously installed version. They now link
3121        explicitly with libpcre.la.
3122    
3123     5. configure.in no longer needs to recognize Cygwin specially.
3124    
3125     6. A problem in pcre.in for Windows platforms is fixed.
3126    
3127     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3128        pcretest, it used to include the size of the study block as part of its
3129        output. Unfortunately, the structure contains a field that has a different
3130        size on different hardware architectures. This meant that the tests that
3131        showed this size failed. As the block is currently always of a fixed size,
3132        this information isn't actually particularly useful in pcretest output, so
3133        I have just removed it.
3134    
3135     8. Three pre-processor statements accidentally did not start in column 1.
3136        Sadly, there are *still* compilers around that complain, even though
3137        standard C has not required this for well over a decade. Sigh.
3138    
3139     9. In pcretest, the code for checking callouts passed small integers in the
3140        callout_data field, which is a void * field. However, some picky compilers
3141        complained about the casts involved for this on 64-bit systems. Now
3142        pcretest passes the address of the small integer instead, which should get
3143        rid of the warnings.
3144    
3145    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3146        both compile and run time, and gives an error if an invalid UTF-8 sequence
3147        is found. There is a option for disabling this check in cases where the
3148        string is known to be correct and/or the maximum performance is wanted.
3149    
3150    11. In response to a bug report, I changed one line in Makefile.in from
3151    
3152            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3153        to
3154            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3155    
3156        to look similar to other lines, but I have no way of telling whether this
3157        is the right thing to do, as I do not use Windows. No doubt I'll get told
3158        if it's wrong...
3159    
3160    
3161    Version 4.3 21-May-03
3162    ---------------------
3163    
3164    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3165       Makefile.
3166    
3167    2. Some refactoring to improve the quality of the code:
3168    
3169       (i)   The utf8_table... variables are now declared "const".
3170    
3171       (ii)  The code for \cx, which used the "case flipping" table to upper case
3172             lower case letters, now just substracts 32. This is ASCII-specific,
3173             but the whole concept of \cx is ASCII-specific, so it seems
3174             reasonable.
3175    
3176       (iii) PCRE was using its character types table to recognize decimal and
3177             hexadecimal digits in the pattern. This is silly, because it handles
3178             only 0-9, a-f, and A-F, but the character types table is locale-
3179             specific, which means strange things might happen. A private
3180             table is now used for this - though it costs 256 bytes, a table is
3181             much faster than multiple explicit tests. Of course, the standard
3182             character types table is still used for matching digits in subject
3183             strings against \d.
3184    
3185       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3186             ending in _t are). So I've renamed it as ESC_tee.
3187    
3188    3. The first argument for regexec() in the POSIX wrapper should have been
3189       defined as "const".
3190    
3191    4. Changed pcretest to use malloc() for its buffers so that they can be
3192       Electric Fenced for debugging.
3193    
3194    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3195       to read one or more bytes before the start of the subject string. Often this
3196       had no effect on PCRE's behaviour, but in some circumstances it could
3197       provoke a segmentation fault.
3198    
3199    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3200       to try to read one or more bytes before the start of the subject string.
3201    
3202    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3203       UTF-8 support could misbehave in various ways if the subject string
3204       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3205       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3206       back over UTF-8 characters.)
3207    
3208    
3209    Version 4.2 14-Apr-03
3210    ---------------------
3211    
3212    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3213    
3214    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3215         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3216         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3217         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3218         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3219           and BUILD_EXEEXT
3220         Note: automatic setting of the BUILD variables is not yet working
3221         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3222           compile-time but not at link-time
3223         [LINK]: use for linking executables only
3224         make different versions for Windows and non-Windows
3225         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3226           libraries
3227         [LINK_FOR_BUILD]: new variable
3228         [OBJEXT]: use throughout
3229         [EXEEXT]: use throughout
3230         <winshared>: new target
3231         <wininstall>: new target
3232         <dftables.o>: use native compiler
3233         <dftables>: use native linker
3234         <install>: handle Windows platform correctly
3235         <clean>: ditto
3236         <check>: ditto
3237         copy DLL to top builddir before testing
3238    
3239       As part of these changes, -no-undefined was removed again. This was reported
3240       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3241       in any case.
3242    
3243    3. Some tidies to get rid of compiler warnings:
3244    
3245       . In the match_data structure, match_limit was an unsigned long int, whereas
3246         match_call_count was an int. I've made them both unsigned long ints.
3247    
3248       . In pcretest the fact that a const uschar * doesn't automatically cast to
3249         a void * provoked a warning.
3250    
3251       . Turning on some more compiler warnings threw up some "shadow" variables
3252         and a few more missing casts.
3253    
3254    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3255       option, a class that contained a single character with a value between 128
3256       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3257    
3258    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3259       option, a class that contained several characters, but with at least one
3260       whose value was between 128 and 255 caused PCRE to crash.
3261    
3262    
3263    Version 4.1 12-Mar-03
3264    ---------------------
3265    
3266    1. Compiling with gcc -pedantic found a couple of places where casts were
3267    needed, and a string in dftables.c that was longer than standard compilers are
3268    required to support.
3269    
3270    2. Compiling with Sun's compiler found a few more places where the code could
3271    be tidied up in order to avoid warnings.
3272    
3273    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3274    first of these names is deprecated in the latest Autoconf in favour of the name
3275    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3276    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3277    analogy I have changed it to CFLAGS_FOR_BUILD.
3278    
3279    4. Added -no-undefined to the linking command in the Makefile, because this is
3280    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3281    linking step for the pcreposix library.
3282    
3283    5. PCRE was failing to diagnose the case of two named groups with the same
3284    name.
3285    
3286    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3287    literal character that is needed in the subject for a match, and scans along to
3288    ensure that it is present before embarking on the full matching process. This
3289    saves time in cases of nested unlimited repeats that are never going to match.
3290    Problem: the scan can take a lot of time if the subject is very long (e.g.
3291    megabytes), thus penalizing straightforward matches. It is now done only if the
3292    amount of subject to be scanned is less than 1000 bytes.
3293    
3294    7. A lesser problem with the same optimization is that it was recording the
3295    first character of an anchored pattern as "needed", thus provoking a search
3296    right along the subject, even when the first match of the pattern was going to
3297    fail. The "needed" character is now not set for anchored patterns, unless it
3298    follows something in the pattern that is of non-fixed length. Thus, it still
3299    fulfils its original purpose of finding quick non-matches in cases of nested
3300    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3301    
3302    
3303    Version 4.0 17-Feb-03
3304    ---------------------
3305    
3306    1. If a comment in an extended regex that started immediately after a meta-item
3307    extended to the end of string, PCRE compiled incorrect data. This could lead to
3308    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3309    
3310    2. Moved to autoconf 2.53 and libtool 1.4.2.
3311    
3312    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3313    the special perltest8 script is no longer needed - all the tests can be run
3314    from a single perltest script.
3315    
3316    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3317    by \s. It has now been removed in PCRE. This means it isn't recognized as
3318    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3319    class [:space:] *does* include VT, thereby creating a mess.
3320    
3321    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3322    space and tab.
3323    
3324    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3325    its new features into the main test script, reducing the number of scripts.
3326    
3327    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3328    were backward compatible, and made the (?i) apply to the whole pattern, as if
3329    /i were given. Now it behaves more logically, and applies the option setting
3330    only to what follows. PCRE has been changed to follow suit. However, if it
3331    finds options settings right at the start of the pattern, it extracts them into
3332    the global options, as before. Thus, they show up in the info data.
3333    
3334    8. Added support for the \Q...\E escape sequence. Characters in between are
3335    treated as literals. This is slightly different from Perl in that $ and @ are
3336    also handled as literals inside the quotes. In Perl, they will cause variable
3337    interpolation. Note the following examples:
3338    
3339        Pattern            PCRE matches      Perl matches
3340    
3341        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3342        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3343        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3344    
3345    For compatibility with Perl, \Q...\E sequences are recognized inside character
3346    classes as well as outside them.
3347    
3348    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3349    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3350    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3351    signed/unsigned warnings.
3352    
3353    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3354    option for pcretest, so I've replaced it by a simple function that does just
3355    that job.
3356    
3357    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3358    "pcregrep -".
3359    
3360    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3361    Java package. This provides some syntactic sugar for simple cases of what my
3362    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3363    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3364    item, you can use this simplified notation. Note that only makes sense with
3365    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3366    greediness, whatever the setting of the PCRE_UNGREEDY option.
3367    
3368    13. A change of greediness default within a pattern was not taking effect at
3369    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3370    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3371    was abstracted outside.
3372    
3373    14. PCRE now supports the \G assertion. It is true when the current matching
3374    position is at the start point of the match. This differs from \A when the
3375    starting offset is non-zero. Used with the /g option of pcretest (or similar
3376    code), it works in the same way as it does for Perl's /g option. If all
3377    alternatives of a regex begin with \G, the expression is anchored to the start
3378    match position, and the "anchored" flag is set in the compiled expression.
3379    
3380    15. Some bugs concerning the handling of certain option changes within patterns
3381    have been fixed. These applied to options other than (?ims). For example,
3382    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3383    been the other way round. Some of this was related to change 7 above.
3384    
3385    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3386    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3387    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3388    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3389    
3390    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3391    mode. Unlike ".", it always matches newline, whatever the setting of
3392    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3393    assertions. Perl allows it, but it doesn't (in general) work because it can't
3394    calculate the length of the lookbehind. At least, that's the case for Perl
3395    5.8.0 - I've been told they are going to document that it doesn't work in
3396    future.
3397    
3398    18. Added an error diagnosis for escapes that PCRE does not support: these are
3399    \L, \l, \N, \P, \p, \U, \u, and \X.
3400    
3401    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3402    reading past the end of the pattern in cases such as /[abcd/.
3403    
3404    20. PCRE was getting more memory than necessary for patterns with classes that
3405    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3406    
3407    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3408    compiling PCRE for use with Virtual Pascal.
3409    
3410    22. Small fix to the Makefile to make it work properly if the build is done
3411    outside the source tree.
3412    
3413    23. Added a new extension: a condition to go with recursion. If a conditional
3414    subpattern starts with (?(R) the "true" branch is used if recursion has
3415    happened, whereas the "false" branch is used only at the top level.
3416    
3417    24. When there was a very long string of literal characters (over 255 bytes
3418    without UTF support, over 250 bytes with UTF support), the computation of how
3419    much memory was required could be incorrect, leading to segfaults or other
3420    strange effects.
3421    
3422    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3423    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3424    there was a subsequent back reference to those brackets. This meant that, for
3425    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3426    possible to check for precisely this case. All we can do is abandon the
3427    optimization if .* occurs inside capturing brackets when there are any back
3428    references whatsoever. (See below for a better fix that came later.)
3429    
3430    26. The handling of the optimization for finding the first character of a
3431    non-anchored pattern, and for finding a character that is required later in the
3432    match were failing in some cases. This didn't break the matching; it just
3433    failed to optimize when it could. The way this is done has been re-implemented.
3434    
3435    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3436    
3437    28. Added a new feature that provides some of the functionality that Perl
3438    provides with (?{...}). The facility is termed a "callout". The way it is done
3439    in PCRE is for the caller to provide an optional function, by setting
3440    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3441    global variable. By default it is unset, which disables all calling out. To get
3442    the function called, the regex must include (?C) at appropriate points. This
3443    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3444    This provides a means of identifying different callout points. When PCRE
3445    reaches such a point in the regex, if pcre_callout has been set, the external
3446    function is called. It is provided with data in a structure called
3447    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3448    matching continues; if it returns a non-zero value, the match at the current
3449    point fails. However, backtracking will occur if possible. [This was changed
3450    later and other features added - see item 49 below.]
3451    
3452    29. pcretest is upgraded to test the callout functionality. It provides a
3453    callout function that displays information. By default, it shows the start of
3454    the match and the current position in the text. There are some new data escapes
3455    to vary what happens:
3456    
3457        \C+         in addition, show current contents of captured substrings
3458        \C-         do not supply a callout function
3459        \C!n        return 1 when callout number n is reached
3460        \C!n!m      return 1 when callout number n is reached for the mth time
3461    
3462    30. If pcregrep was called with the -l option and just a single file name, it
3463    output "<stdin>" if a match was found, instead of the file name.
3464    
3465    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3466    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3467    pcre_exec(). This saves a malloc/free per call. The default value of
3468    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3469    when configuring.
3470    
3471    32. The default maximum size of a compiled pattern is 64K. There have been a
3472    few cases of people hitting this limit. The code now uses macros to handle the
3473    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3474    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3475    configuring. Tests 2 and 5 work only with 2-byte links because they output
3476    debugging information about compiled patterns.
3477    
3478    33. Internal code re-arrangements:
3479    
3480    (a) Moved the debugging function for printing out a compiled regex into
3481        its own source file (printint.c) and used #include to pull it into
3482        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3483        separate copies.
3484    
3485    (b) Defined the list of op-code names for debugging as a macro in
3486        internal.h so that it is next to the definition of the opcodes.
3487    
3488    (c) Defined a table of op-code lengths for simpler skipping along compiled
3489        code. This is again a macro in internal.h so that it is next to the
3490        definition of the opcodes.
3491    
3492    34. Added support for recursive calls to individual subpatterns, along the
3493    lines of Robin Houston's patch (but implemented somewhat differently).
3494    
3495    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3496    allow it to read and process whole directories in Win32. This code was
3497    contributed by Lionel Fourquaux; it has not been tested by me.
3498    
3499    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3500    used to name a group. Names consist of alphanumerics and underscores, and must
3501    be unique. Back references use the syntax (?P=name) and recursive calls use
3502    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3503    numbers. The function pcre_fullinfo() can be used after compilation to extract
3504    a name/number map. There are three relevant calls:
3505    
3506      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3507      PCRE_INFO_NAMECOUNT            yields the number of entries
3508      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3509    
3510    The map is a vector of fixed-size entries. The size of each entry depends on
3511    the length of the longest name used. The first two bytes of each entry are the
3512    group number, most significant byte first. There follows the corresponding
3513    name, zero terminated. The names are in alphabetical order.
3514    
3515    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3516    case instead of 255. Making it the same both with and without UTF-8 support
3517    means that the same test output works with both.
3518    
3519    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3520    calling malloc() with a zero argument.