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

Diff of /code/trunk/ChangeLog

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

revision 21 by nigel, Sat Feb 24 21:38:37 2007 UTC revision 740 by zherczeg, Mon Oct 31 06:10:14 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.21
5    ------------
6    
7  Version 1.09 28 Apr-98  1.  Updating the JIT compiler.
8    
9    
10    Version 8.20 21-Oct-2011
11    ------------------------
12    
13    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
14        a POSIX class. After further experiments with Perl, which convinced me that
15        Perl has bugs and confusions, a closing square bracket is no longer allowed
16        in a POSIX name. This bug also affected patterns with classes that started
17        with full stops.
18    
19    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
20        captured substring, but while checking the failing first alternative,
21        substring 1 is temporarily captured. If the output vector supplied to
22        pcre_exec() was not big enough for this capture, the yield of the function
23        was still zero ("insufficient space for captured substrings"). This cannot
24        be totally fixed without adding another stack variable, which seems a lot
25        of expense for a edge case. However, I have improved the situation in cases
26        such as /(a)(b)x|abc/ matched against "abc", where the return code
27        indicates that fewer than the maximum number of slots in the ovector have
28        been set.
29    
30    3.  Related to (2) above: when there are more back references in a pattern than
31        slots in the output vector, pcre_exec() uses temporary memory during
32        matching, and copies in the captures as far as possible afterwards. It was
33        using the entire output vector, but this conflicts with the specification
34        that only 2/3 is used for passing back captured substrings. Now it uses
35        only the first 2/3, for compatibility. This is, of course, another edge
36        case.
37    
38    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
39        main code base, and can be used by building with --enable-jit. When this is
40        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
41        runtime --no-jit option is given.
42    
43    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
44        ovector, the return from the function was zero, implying that there were
45        other matches that did not fit. The correct "exactly full" value is now
46        returned.
47    
48    6.  If a subpattern that was called recursively or as a subroutine contained
49        (*PRUNE) or any other control that caused it to give a non-standard return,
50        invalid errors such as "Error -26 (nested recursion at the same subject
51        position)" or even infinite loops could occur.
52    
53    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
54        computing the minimum length on reaching *ACCEPT, and so ended up with the
55        wrong value of 1 rather than 0. Further investigation indicates that
56        computing a minimum subject length in the presence of *ACCEPT is difficult
57        (think back references, subroutine calls), and so I have changed the code
58        so that no minimum is registered for a pattern that contains *ACCEPT.
59    
60    8.  If (*THEN) was present in the first (true) branch of a conditional group,
61        it was not handled as intended. [But see 16 below.]
62    
63    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
64        Sheri Pierce.
65    
66    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
67        the first byte in a match must be "a".
68    
69    11. Change 17 for 8.13 increased the recursion depth for patterns like
70        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
71        pattern contains any instances of (*THEN). If it does not, the old
72        optimizations are restored. It would be nice to do this on a per-group
73        basis, but at the moment that is not feasible.
74    
75    12. In some environments, the output of pcretest -C is CRLF terminated. This
76        broke RunTest's code that checks for the link size. A single white space
77        character after the value is now allowed for.
78    
79    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
80        For "fr", it uses the Windows-specific input and output files.
81    
82    14. If (*THEN) appeared in a group that was called recursively or as a
83        subroutine, it did not work as intended. [But see next item.]
84    
85    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
86        pattern fragments (but not containing any | characters). If A and B are
87        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
88        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
89        D. In other words, Perl considers parentheses that do not contain any |
90        characters to be part of a surrounding alternative, whereas PCRE was
91        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
92        differently. PCRE now behaves in the same way as Perl, except in the case
93        of subroutine/recursion calls such as (?1) which have in any case always
94        been different (but PCRE had them first :-).
95    
96    16. Related to 15 above: Perl does not treat the | in a conditional group as
97        creating alternatives. Such a group is treated in the same way as an
98        ordinary group without any | characters when processing (*THEN). PCRE has
99        been changed to match Perl's behaviour.
100    
101    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
102        RunGrepTest script failed.
103    
104    18. Change 22 for version 13 caused atomic groups to use more stack. This is
105        inevitable for groups that contain captures, but it can lead to a lot of
106        stack use in large patterns. The old behaviour has been restored for atomic
107        groups that do not contain any capturing parentheses.
108    
109    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
110        suppress the check for a minimum subject length at run time. (If it was
111        given to pcre_exec() or pcre_dfa_exec() it did work.)
112    
113    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
114        fail to work when decoding hex characters in data strings in EBCDIC
115        environments.
116    
117    21. It appears that in at least one Mac OS environment, the isxdigit() function
118        is implemented as a macro that evaluates to its argument more than once,
119        contravening the C 90 Standard (I haven't checked a later standard). There
120        was an instance in pcretest which caused it to go wrong when processing
121        \x{...} escapes in subject strings. The has been rewritten to avoid using
122        things like p++ in the argument of isxdigit().
123    
124    
125    Version 8.13 16-Aug-2011
126    ------------------------
127    
128    1.  The Unicode data tables have been updated to Unicode 6.0.0.
129    
130    2.  Two minor typos in pcre_internal.h have been fixed.
131    
132    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
133        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
134        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
135    
136    4.  There were a number of related bugs in the code for matching backrefences
137        caselessly in UTF-8 mode when codes for the characters concerned were
138        different numbers of bytes. For example, U+023A and U+2C65 are an upper
139        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
140        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
141        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
142        2-byte code at the end of the subject (it thought there wasn't enough data
143        left).
144    
145    5.  Comprehensive information about what went wrong is now returned by
146        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
147        as the output vector has at least 2 elements. The offset of the start of
148        the failing character and a reason code are placed in the vector.
149    
150    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
151        now returned is for the first byte of the failing character, instead of the
152        last byte inspected. This is an incompatible change, but I hope it is small
153        enough not to be a problem. It makes the returned offset consistent with
154        pcre_exec() and pcre_dfa_exec().
155    
156    7.  pcretest now gives a text phrase as well as the error number when
157        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
158        failure, the offset and reason code are output.
159    
160    8.  When \R was used with a maximizing quantifier it failed to skip backwards
161        over a \r\n pair if the subsequent match failed. Instead, it just skipped
162        back over a single character (\n). This seems wrong (because it treated the
163        two characters as a single entity when going forwards), conflicts with the
164        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
165        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
166        has been changed.
167    
168    9.  Some internal refactoring has changed the processing so that the handling
169        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
170        time (the PCRE_DOTALL option was changed this way some time ago: version
171        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
172        which was always a bit of a fudge. It also means that there is one less
173        argument for the match() function, which reduces its stack requirements
174        slightly. This change also fixes an incompatibility with Perl: the pattern
175        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
176    
177    10. More internal refactoring has drastically reduced the number of recursive
178        calls to match() for possessively repeated groups such as (abc)++ when
179        using pcre_exec().
180    
181    11. While implementing 10, a number of bugs in the handling of groups were
182        discovered and fixed:
183    
184        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
185        (a|)*(?1) gave a compile-time internal error.
186        ((a|)+)+  did not notice that the outer group could match an empty string.
187        (^a|^)+   was not marked as anchored.
188        (.*a|.*)+ was not marked as matching at start or after a newline.
189    
190    12. Yet more internal refactoring has removed another argument from the match()
191        function. Special calls to this function are now indicated by setting a
192        value in a variable in the "match data" data block.
193    
194    13. Be more explicit in pcre_study() instead of relying on "default" for
195        opcodes that mean there is no starting character; this means that when new
196        ones are added and accidentally left out of pcre_study(), testing should
197        pick them up.
198    
199    14. The -s option of pcretest has been documented for ages as being an old
200        synonym of -m (show memory usage). I have changed it to mean "force study
201        for every regex", that is, assume /S for every regex. This is similar to -i
202        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
203        using it. It makes it easier to run collections of tests with and without
204        study enabled, and thereby test pcre_study() more easily. All the standard
205        tests are now run with and without -s (but some patterns can be marked as
206        "never study" - see 20 below).
207    
208    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
209        restoration of the capturing data to the outer values was not happening
210        correctly.
211    
212    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
213        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
214        pattern had matched an empty string, and so incorrectly returned a no
215        match.
216    
217    17. There was optimizing code for the last branch of non-capturing parentheses,
218        and also for the obeyed branch of a conditional subexpression, which used
219        tail recursion to cut down on stack usage. Unfortunately, now that there is
220        the possibility of (*THEN) occurring in these branches, tail recursion is
221        no longer possible because the return has to be checked for (*THEN). These
222        two optimizations have therefore been removed. [But see 8.20/11 above.]
223    
224    18. If a pattern containing \R was studied, it was assumed that \R always
225        matched two bytes, thus causing the minimum subject length to be
226        incorrectly computed because \R can also match just one byte.
227    
228    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
229        was incorrectly computed.
230    
231    20. If /S is present twice on a test pattern in pcretest input, it now
232        *disables* studying, thereby overriding the use of -s on the command line
233        (see 14 above). This is necessary for one or two tests to keep the output
234        identical in both cases.
235    
236    21. When (*ACCEPT) was used in an assertion that matched an empty string and
237        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
238    
239    22. When an atomic group that contained a capturing parenthesis was
240        successfully matched, but the branch in which it appeared failed, the
241        capturing was not being forgotten if a higher numbered group was later
242        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
243        group 1 to "a", when in fact it should be unset. This applied to multi-
244        branched capturing and non-capturing groups, repeated or not, and also to
245        positive assertions (capturing in negative assertions does not happen
246        in PCRE) and also to nested atomic groups.
247    
248    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
249        subject after a captured substring, to make it easier to tell which of a
250        number of identical substrings has been captured.
251    
252    24. The way atomic groups are processed by pcre_exec() has been changed so that
253        if they are repeated, backtracking one repetition now resets captured
254        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
255        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
256        "aaa". Previously, it would have been "a". As part of this code
257        refactoring, the way recursive calls are handled has also been changed.
258    
259    25. If an assertion condition captured any substrings, they were not passed
260        back unless some other capturing happened later. For example, if
261        (?(?=(a))a) was matched against "a", no capturing was returned.
262    
263    26. When studying a pattern that contained subroutine calls or assertions,
264        the code for finding the minimum length of a possible match was handling
265        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
266        group 1 called group 2 while simultaneously a separate group 2 called group
267        1). A stack overflow occurred in this case. I have fixed this by limiting
268        the recursion depth to 10.
269    
270    27. Updated RunTest.bat in the distribution to the version supplied by Tom
271        Fortmann. This supports explicit test numbers on the command line, and has
272        argument validation and error reporting.
273    
274    28. An instance of \X with an unlimited repeat could fail if at any point the
275        first character it looked at was a mark character.
276    
277    29. Some minor code refactoring concerning Unicode properties and scripts
278        should reduce the stack requirement of match() slightly.
279    
280    30. Added the '=' option to pcretest to check the setting of unused capturing
281        slots at the end of the pattern, which are documented as being -1, but are
282        not included in the return count.
283    
284    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
285        compiled something random. Now it gives a compile-time error (as does
286        Perl).
287    
288    32. A *MARK encountered during the processing of a positive assertion is now
289        recorded and passed back (compatible with Perl).
290    
291    33. If --only-matching or --colour was set on a pcregrep call whose pattern
292        had alternative anchored branches, the search for a second match in a line
293        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
294        matched the line "0102" twice. The same bug affected patterns that started
295        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
296        twice.
297    
298    34. Previously, PCRE did not allow quantification of assertions. However, Perl
299        does, and because of capturing effects, quantifying parenthesized
300        assertions may at times be useful. Quantifiers are now allowed for
301        parenthesized assertions.
302    
303    35. A minor code tidy in pcre_compile() when checking options for \R usage.
304    
305    36. \g was being checked for fancy things in a character class, when it should
306        just be a literal "g".
307    
308    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
309        appearance of a nested POSIX class supersedes an apparent external class.
310        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
311        unescaped square brackets may also appear as part of class names. For
312        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
313        more like Perl. (But see 8.20/1 above.)
314    
315    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
316        was because it thought it was \N{name}, which is not supported).
317    
318    39. Add minix to OS list not supporting the -S option in pcretest.
319    
320    40. PCRE tries to detect cases of infinite recursion at compile time, but it
321        cannot analyze patterns in sufficient detail to catch mutual recursions
322        such as ((?1))((?2)). There is now a runtime test that gives an error if a
323        subgroup is called recursively as a subpattern for a second time at the
324        same position in the subject string. In previous releases this might have
325        been caught by the recursion limit, or it might have run out of stack.
326    
327    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
328        happen only once. PCRE was, however incorrectly giving a compile time error
329        "recursive call could loop indefinitely" because it cannot analyze the
330        pattern in sufficient detail. The compile time test no longer happens when
331        PCRE is compiling a conditional subpattern, but actual runaway loops are
332        now caught at runtime (see 40 above).
333    
334    42. It seems that Perl allows any characters other than a closing parenthesis
335        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
336        has been changed to be the same.
337    
338    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
339        as not to get warnings when autogen.sh is called. Also changed
340        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
341    
342    44. To help people who use pcregrep to scan files containing exceedingly long
343        lines, the following changes have been made:
344    
345        (a) The default value of the buffer size parameter has been increased from
346            8K to 20K. (The actual buffer used is three times this size.)
347    
348        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
349            PCRE is built.
350    
351        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
352            to be set at run time.
353    
354        (d) Numerical values in pcregrep options can be followed by K or M, for
355            example --buffer-size=50K.
356    
357        (e) If a line being scanned overflows pcregrep's buffer, an error is now
358            given and the return code is set to 2.
359    
360    45. Add a pointer to the latest mark to the callout data block.
361    
362    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
363        partial match of an empty string instead of no match. This was specific to
364        the use of ".".
365    
366    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
367        complete match instead of a partial match. This bug was dependent on both
368        the PCRE_UTF8 and PCRE_DOTALL options being set.
369    
370    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
371        starting byte set, because \b was not being ignored.
372    
373    
374    Version 8.12 15-Jan-2011
375    ------------------------
376    
377    1.  Fixed some typos in the markup of the man pages, and wrote a script that
378        checks for such things as part of the documentation building process.
379    
380    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
381        --match-limit and --recursion-limit options (added for 8.11). In
382        particular, this made one of the standard tests fail. (The integer value
383        went into the wrong half of a long int.)
384    
385    3.  If the --colour option was given to pcregrep with -v (invert match), it
386        did strange things, either producing crazy output, or crashing. It should,
387        of course, ignore a request for colour when reporting lines that do not
388        match.
389    
390    4.  Another pcregrep bug caused similar problems if --colour was specified with
391        -M (multiline) and the pattern match finished with a line ending.
392    
393    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
394        matched in multiline mode, the following line was shown as part of the
395        match. This seems wrong, so I have changed it.
396    
397    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
398        the check for further matches in the same line (so they could be coloured)
399        to overrun the end of the current line. If another match was found, it was
400        incorrectly shown (and then shown again when found in the next line).
401    
402    7.  If pcregrep was compiled under Windows, there was a reference to the
403        function pcregrep_exit() before it was defined. I am assuming this was
404        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
405        reported by a user. I've moved the definition above the reference.
406    
407    
408    Version 8.11 10-Dec-2010
409    ------------------------
410    
411    1.  (*THEN) was not working properly if there were untried alternatives prior
412        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
413        backtracked to try for "b" instead of moving to the next alternative branch
414        at the same level (in this case, to look for "c"). The Perl documentation
415        is clear that when (*THEN) is backtracked onto, it goes to the "next
416        alternative in the innermost enclosing group".
417    
418    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
419        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
420        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
421        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
422        (*THEN).
423    
424    3.  If \s appeared in a character class, it removed the VT character from
425        the class, even if it had been included by some previous item, for example
426        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
427        of \s, but is part of the POSIX "space" class.)
428    
429    4.  A partial match never returns an empty string (because you can always
430        match an empty string at the end of the subject); however the checking for
431        an empty string was starting at the "start of match" point. This has been
432        changed to the "earliest inspected character" point, because the returned
433        data for a partial match starts at this character. This means that, for
434        example, /(?<=abc)def/ gives a partial match for the subject "abc"
435        (previously it gave "no match").
436    
437    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
438        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
439        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
440        has an implication that the given string is incomplete (because a partial
441        match is preferred over a full match). For this reason, these items now
442        give a partial match in this situation. [Aside: previously, the one case
443        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
444        match rather than a full match, which was wrong by the old rules, but is
445        now correct.]
446    
447    6.  There was a bug in the handling of #-introduced comments, recognized when
448        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
449        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
450        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
451        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
452        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
453        places in pcre_compile().
454    
455    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
456        comments when looking ahead for named forward references to subpatterns,
457        the only newline sequence it recognized was NL. It now handles newlines
458        according to the set newline convention.
459    
460    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
461        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
462        cater for a lack of strerror(). These oversights have been fixed.
463    
464    9.  Added --match-limit and --recursion-limit to pcregrep.
465    
466    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
467    
468    11. When the -o option was used, pcregrep was setting a return code of 1, even
469        when matches were found, and --line-buffered was not being honoured.
470    
471    12. Added an optional parentheses number to the -o and --only-matching options
472        of pcregrep.
473    
474    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
475        can match an empty string. The code to do it in pcretest and pcredemo
476        needed fixing:
477    
478        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
479            only one byte after an empty string match just before CRLF (this case
480            just got forgotten; "any" and "anycrlf" were OK).
481    
482        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
483            mode when an empty string match preceded an ASCII character followed by
484            a non-ASCII character. (The code for advancing by one character rather
485            than one byte was nonsense.)
486    
487        (c) The pcredemo.c sample program did not have any code at all to handle
488            the cases when CRLF is a valid newline sequence.
489    
490    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
491        as a starting offset was within the subject string. There is now a new
492        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
493        negative or greater than the length of the string. In order to test this,
494        pcretest is extended to allow the setting of negative starting offsets.
495    
496    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
497        starting offset points to the beginning of a UTF-8 character was
498        unnecessarily clumsy. I tidied it up.
499    
500    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
501        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
502    
503    17. Nobody had reported that the --include_dir option, which was added in
504        release 7.7 should have been called --include-dir (hyphen, not underscore)
505        for compatibility with GNU grep. I have changed it to --include-dir, but
506        left --include_dir as an undocumented synonym, and the same for
507        --exclude-dir, though that is not available in GNU grep, at least as of
508        release 2.5.4.
509    
510    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
511        characters from a string of bytes) have been redefined so as not to use
512        loops, in order to improve performance in some environments. At the same
513        time, I abstracted some of the common code into auxiliary macros to save
514        repetition (this should not affect the compiled code).
515    
516    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
517        compile-time error is now given if \c is not followed by an ASCII
518        character, that is, a byte less than 128. (In EBCDIC mode, the code is
519        different, and any byte value is allowed.)
520    
521    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
522        START_OPTIMIZE option, which is now allowed at compile time - but just
523        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
524        to pcregrep and other applications that have no direct access to PCRE
525        options. The new /Y option in pcretest sets this option when calling
526        pcre_compile().
527    
528    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
529        back references. Groups containing recursive back references were forced to
530        be atomic by that change, but in the case of named groups, the amount of
531        memory required was incorrectly computed, leading to "Failed: internal
532        error: code overflow". This has been fixed.
533    
534    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
535        pcretest.c, to avoid build problems in some Borland environments.
536    
537    
538    Version 8.10 25-Jun-2010
539    ------------------------
540    
541    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
542        THEN.
543    
544    2.  (*ACCEPT) was not working when inside an atomic group.
545    
546    3.  Inside a character class, \B is treated as a literal by default, but
547        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
548        causes the error). The code is unchanged, but I tidied the documentation.
549    
550    4.  Inside a character class, PCRE always treated \R and \X as literals,
551        whereas Perl faults them if its -w option is set. I have changed PCRE so
552        that it faults them when PCRE_EXTRA is set.
553    
554    5.  Added support for \N, which always matches any character other than
555        newline. (It is the same as "." when PCRE_DOTALL is not set.)
556    
557    6.  When compiling pcregrep with newer versions of gcc which may have
558        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
559        declared with attribute warn_unused_result" were given. Just casting the
560        result to (void) does not stop the warnings; a more elaborate fudge is
561        needed. I've used a macro to implement this.
562    
563    7.  Minor change to pcretest.c to avoid a compiler warning.
564    
565    8.  Added four artifical Unicode properties to help with an option to make
566        \s etc use properties (see next item). The new properties are: Xan
567        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
568    
569    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
570        use Unicode properties. (*UCP) at the start of a pattern can be used to set
571        this option. Modified pcretest to add /W to test this facility. Added
572        REG_UCP to make it available via the POSIX interface.
573    
574    10. Added --line-buffered to pcregrep.
575    
576    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
577        studied, and the match started with a letter with a code point greater than
578        127 whose first byte was different to the first byte of the other case of
579        the letter, the other case of this starting letter was not recognized
580        (#976).
581    
582    12. If a pattern that was studied started with a repeated Unicode property
583        test, for example, \p{Nd}+, there was the theoretical possibility of
584        setting up an incorrect bitmap of starting bytes, but fortunately it could
585        not have actually happened in practice until change 8 above was made (it
586        added property types that matched character-matching opcodes).
587    
588    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
589        possible starting bytes for non-anchored patterns.
590    
591    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
592        \R, and also a number of cases that involve Unicode properties, both
593        explicit and implicit when PCRE_UCP is set.
594    
595    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
596        input, it could crash or give wrong results if characters with values
597        greater than 0xc0 were present in the subject string. (Detail: it assumed
598        UTF-8 input when processing these items.)
599    
600    16. Added a lot of (int) casts to avoid compiler warnings in systems where
601        size_t is 64-bit (#991).
602    
603    17. Added a check for running out of memory when PCRE is compiled with
604        --disable-stack-for-recursion (#990).
605    
606    18. If the last data line in a file for pcretest does not have a newline on
607        the end, a newline was missing in the output.
608    
609    19. The default pcre_chartables.c file recognizes only ASCII characters (values
610        less than 128) in its various bitmaps. However, there is a facility for
611        generating tables according to the current locale when PCRE is compiled. It
612        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
613        characters, are recognized by isspace() and therefore were getting set in
614        these tables, and indeed these tables seem to approximate to ISO 8859. This
615        caused a problem in UTF-8 mode when pcre_study() was used to create a list
616        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
617        which of course cannot start UTF-8 characters. I have changed the code so
618        that only real ASCII characters (less than 128) and the correct starting
619        bytes for UTF-8 encodings are set for characters greater than 127 when in
620        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
621        altogether.)
622    
623    20. Added the /T option to pcretest so as to be able to run tests with non-
624        standard character tables, thus making it possible to include the tests
625        used for 19 above in the standard set of tests.
626    
627    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
628        reference to a subpattern the other side of a comment that contains an
629        opening parenthesis caused either an internal compiling error, or a
630        reference to the wrong subpattern.
631    
632    
633    Version 8.02 19-Mar-2010
634    ------------------------
635    
636    1.  The Unicode data tables have been updated to Unicode 5.2.0.
637    
638    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
639        configured.
640    
641    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
642        original author of that file, following a query about its status.
643    
644    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
645        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
646    
647    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
648        quantifier applied to a forward-referencing subroutine call, could compile
649        incorrect code or give the error "internal error: previously-checked
650        referenced subpattern not found".
651    
652    6.  Both MS Visual Studio and Symbian OS have problems with initializing
653        variables to point to external functions. For these systems, therefore,
654        pcre_malloc etc. are now initialized to local functions that call the
655        relevant global functions.
656    
657    7.  There were two entries missing in the vectors called coptable and poptable
658        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
659        I've fixed the data, and added a kludgy way of testing at compile time that
660        the lengths are correct (equal to the number of opcodes).
661    
662    8.  Following on from 7, I added a similar kludge to check the length of the
663        eint vector in pcreposix.c.
664    
665    9.  Error texts for pcre_compile() are held as one long string to avoid too
666        much relocation at load time. To find a text, the string is searched,
667        counting zeros. There was no check for running off the end of the string,
668        which could happen if a new error number was added without updating the
669        string.
670    
671    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
672    
673    11. \K was not working if it appeared in an atomic group or in a group that
674        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
675        \K is "not well defined" if used in an assertion. PCRE now accepts it if
676        the assertion is positive, but not if it is negative.
677    
678    12. Change 11 fortuitously reduced the size of the stack frame used in the
679        "match()" function of pcre_exec.c by one pointer. Forthcoming
680        implementation of support for (*MARK) will need an extra pointer on the
681        stack; I have reserved it now, so that the stack frame size does not
682        decrease.
683    
684    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
685        item in branch that calls a recursion is a subroutine call - as in the
686        second branch in the above example - was incorrectly given the compile-
687        time error "recursive call could loop indefinitely" because pcre_compile()
688        was not correctly checking the subroutine for matching a non-empty string.
689    
690    14. The checks for overrunning compiling workspace could trigger after an
691        overrun had occurred. This is a "should never occur" error, but it can be
692        triggered by pathological patterns such as hundreds of nested parentheses.
693        The checks now trigger 100 bytes before the end of the workspace.
694    
695    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
696    
697    
698    Version 8.01 19-Jan-2010
699    ------------------------
700    
701    1.  If a pattern contained a conditional subpattern with only one branch (in
702        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
703        computed the wrong minimum data length (which is of course zero for such
704        subpatterns). This could cause incorrect "no match" results.
705    
706    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
707        the pattern is reset in the first branch, pcre_compile() failed with
708        "internal error: code overflow at offset...". This happened only when
709        the reset was to the original external option setting. (An optimization
710        abstracts leading options settings into an external setting, which was the
711        cause of this.)
712    
713    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
714        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
715        assertion pattern did not match (meaning that the assertion was true), it
716        was incorrectly treated as false if the SKIP had been reached during the
717        matching. This also applied to assertions used as conditions.
718    
719    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
720        assertion subpattern, including such a pattern used as a condition,
721        unpredictable results occurred, instead of the error return
722        PCRE_ERROR_DFA_UITEM.
723    
724    5.  The C++ GlobalReplace function was not working like Perl for the special
725        situation when an empty string is matched. It now does the fancy magic
726        stuff that is necessary.
727    
728    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
729        removed. (These were left over from very, very early versions of PCRE.)
730    
731    7.  Some cosmetic changes to the code to make life easier when compiling it
732        as part of something else:
733    
734        (a) Change DEBUG to PCRE_DEBUG.
735    
736        (b) In pcre_compile(), rename the member of the "branch_chain" structure
737            called "current" as "current_branch", to prevent a collision with the
738            Linux macro when compiled as a kernel module.
739    
740        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
741            prevent a collision with the Linux macro when compiled as a kernel
742            module.
743    
744    8.  In pcre_compile() there are some checks for integer overflows that used to
745        cast potentially large values to (double). This has been changed to that
746        when building, a check for int64_t is made, and if it is found, it is used
747        instead, thus avoiding the use of floating point arithmetic. (There is no
748        other use of FP in PCRE.) If int64_t is not found, the fallback is to
749        double.
750    
751    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
752        2005 (difference between two addresses compared to an unsigned value).
753    
754    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
755        custom one, because of the following reported problem in Windows:
756    
757          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
758              under Win32.
759          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
760              therefore missing the function definition.
761          - The compiler thus generates a "C" signature for the test function.
762          - The linker fails to find the "C" function.
763          - PCRE fails to configure if asked to do so against libbz2.
764    
765    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
766        messages were output:
767    
768          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
769          rerunning libtoolize, to keep the correct libtool macros in-tree.
770          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
771    
772        I have done both of these things.
773    
774    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
775        most of the time, it *can* run out if it is given a pattern that contains a
776        runaway infinite recursion. I updated the discussion in the pcrestack man
777        page.
778    
779    13. Now that we have gone to the x.xx style of version numbers, the minor
780        version may start with zero. Using 08 or 09 is a bad idea because users
781        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
782        interpreted as invalid octal numbers. I've updated the previous comment in
783        configure.ac, and also added a check that gives an error if 08 or 09 are
784        used.
785    
786    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
787        causing partial matching to fail when the end of the subject matched \W
788        in a UTF-8 pattern where \W was quantified with a minimum of 3.
789    
790    15. There were some discrepancies between the declarations in pcre_internal.h
791        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
792        their definitions. The declarations used "const uschar *" and the
793        definitions used USPTR. Even though USPTR is normally defined as "const
794        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
795        reported that: "This difference in casting confuses some C++ compilers, for
796        example, SunCC recognizes above declarations as different functions and
797        generates broken code for hbpcre." I have changed the declarations to use
798        USPTR.
799    
800    16. GNU libtool is named differently on some systems. The autogen.sh script now
801        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
802        (FreeBSD).
803    
804    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
805        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
806        comment: "Figure out how to create a longlong from a string: strtoll and
807        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
808        instance, but it only takes 2 args instead of 3!"
809    
810    18. A subtle bug concerned with back references has been fixed by a change of
811        specification, with a corresponding code fix. A pattern such as
812        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
813        refers, was giving matches when it shouldn't. For example, xa=xaaa would
814        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
815        same bug. Such groups have to be quantified to be useful, or contained
816        inside another quantified group. (If there's no repetition, the reference
817        can never match.) The problem arises because, having left the group and
818        moved on to the rest of the pattern, a later failure that backtracks into
819        the group uses the captured value from the final iteration of the group
820        rather than the correct earlier one. I have fixed this in PCRE by forcing
821        any group that contains a reference to itself to be an atomic group; that
822        is, there cannot be any backtracking into it once it has completed. This is
823        similar to recursive and subroutine calls.
824    
825    
826    Version 8.00 19-Oct-09
827    ----------------------
828    
829    1.  The table for translating pcre_compile() error codes into POSIX error codes
830        was out-of-date, and there was no check on the pcre_compile() error code
831        being within the table. This could lead to an OK return being given in
832        error.
833    
834    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
835        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
836        in a Windows environment.
837    
838    3.  The pcregrep --count option prints the count for each file even when it is
839        zero, as does GNU grep. However, pcregrep was also printing all files when
840        --files-with-matches was added. Now, when both options are given, it prints
841        counts only for those files that have at least one match. (GNU grep just
842        prints the file name in this circumstance, but including the count seems
843        more useful - otherwise, why use --count?) Also ensured that the
844        combination -clh just lists non-zero counts, with no names.
845    
846    4.  The long form of the pcregrep -F option was incorrectly implemented as
847        --fixed_strings instead of --fixed-strings. This is an incompatible change,
848        but it seems right to fix it, and I didn't think it was worth preserving
849        the old behaviour.
850    
851    5.  The command line items --regex=pattern and --regexp=pattern were not
852        recognized by pcregrep, which required --regex pattern or --regexp pattern
853        (with a space rather than an '='). The man page documented the '=' forms,
854        which are compatible with GNU grep; these now work.
855    
856    6.  No libpcreposix.pc file was created for pkg-config; there was just
857        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
858    
859    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
860        when UCP support is not needed, by modifying the Python script that
861        generates it from Unicode data files. This should not matter if the module
862        is correctly used as a library, but I received one complaint about 50K of
863        unwanted data. My guess is that the person linked everything into his
864        program rather than using a library. Anyway, it does no harm.
865    
866    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
867        was a minimum greater than 1 for a wide character in a possessive
868        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
869        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
870        character. Chaos in the form of incorrect output or a compiling loop could
871        result.
872    
873    9.  The restrictions on what a pattern can contain when partial matching is
874        requested for pcre_exec() have been removed. All patterns can now be
875        partially matched by this function. In addition, if there are at least two
876        slots in the offset vector, the offset of the earliest inspected character
877        for the match and the offset of the end of the subject are set in them when
878        PCRE_ERROR_PARTIAL is returned.
879    
880    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
881        synonymous with PCRE_PARTIAL, for backwards compatibility, and
882        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
883        and may be more useful for multi-segment matching.
884    
885    11. Partial matching with pcre_exec() is now more intuitive. A partial match
886        used to be given if ever the end of the subject was reached; now it is
887        given only if matching could not proceed because another character was
888        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
889        string "Z", which now yields "no match" instead of "partial match". In the
890        case of pcre_dfa_exec(), "no match" is given if every matching path for the
891        final character ended with (*FAIL).
892    
893    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
894        if the pattern had a "must contain" character that was already found in the
895        earlier partial match, unless partial matching was again requested. For
896        example, with the pattern /dog.(body)?/, the "must contain" character is
897        "g". If the first part-match was for the string "dog", restarting with
898        "sbody" failed. This bug has been fixed.
899    
900    13. The string returned by pcre_dfa_exec() after a partial match has been
901        changed so that it starts at the first inspected character rather than the
902        first character of the match. This makes a difference only if the pattern
903        starts with a lookbehind assertion or \b or \B (\K is not supported by
904        pcre_dfa_exec()). It's an incompatible change, but it makes the two
905        matching functions compatible, and I think it's the right thing to do.
906    
907    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
908        so that the demonstration program is easily available in environments where
909        PCRE has not been installed from source.
910    
911    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
912        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
913        library.
914    
915    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
916        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
917        is not the first non-POSIX option to be added. Clearly some people find
918        these options useful.
919    
920    17. If a caller to the POSIX matching function regexec() passes a non-zero
921        value for nmatch with a NULL value for pmatch, the value of
922        nmatch is forced to zero.
923    
924    18. RunGrepTest did not have a test for the availability of the -u option of
925        the diff command, as RunTest does. It now checks in the same way as
926        RunTest, and also checks for the -b option.
927    
928    19. If an odd number of negated classes containing just a single character
929        interposed, within parentheses, between a forward reference to a named
930        subpattern and the definition of the subpattern, compilation crashed with
931        an internal error, complaining that it could not find the referenced
932        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
933        [The bug was that it was starting one character too far in when skipping
934        over the character class, thus treating the ] as data rather than
935        terminating the class. This meant it could skip too much.]
936    
937    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
938        /g option in pcretest when the pattern contains \K, which makes it possible
939        to have an empty string match not at the start, even when the pattern is
940        anchored. Updated pcretest and pcredemo to use this option.
941    
942    21. If the maximum number of capturing subpatterns in a recursion was greater
943        than the maximum at the outer level, the higher number was returned, but
944        with unset values at the outer level. The correct (outer level) value is
945        now given.
946    
947    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
948        PCRE did not set those parentheses (unlike Perl). I have now found a way to
949        make it do so. The string so far is captured, making this feature
950        compatible with Perl.
951    
952    23. The tests have been re-organized, adding tests 11 and 12, to make it
953        possible to check the Perl 5.10 features against Perl 5.10.
954    
955    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
956        pattern matches a fixed length string. PCRE did not allow this; now it
957        does. Neither allows recursion.
958    
959    25. I finally figured out how to implement a request to provide the minimum
960        length of subject string that was needed in order to match a given pattern.
961        (It was back references and recursion that I had previously got hung up
962        on.) This code has now been added to pcre_study(); it finds a lower bound
963        to the length of subject needed. It is not necessarily the greatest lower
964        bound, but using it to avoid searching strings that are too short does give
965        some useful speed-ups. The value is available to calling programs via
966        pcre_fullinfo().
967    
968    26. While implementing 25, I discovered to my embarrassment that pcretest had
969        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
970        study optimizations had never been tested with that matching function.
971        Oops. What is worse, even when it was passed study data, there was a bug in
972        pcre_dfa_exec() that meant it never actually used it. Double oops. There
973        were also very few tests of studied patterns with pcre_dfa_exec().
974    
975    27. If (?| is used to create subpatterns with duplicate numbers, they are now
976        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
977        on the other side of the coin, they are no longer allowed to have different
978        names, because these cannot be distinguished in PCRE, and this has caused
979        confusion. (This is a difference from Perl.)
980    
981    28. When duplicate subpattern names are present (necessarily with different
982        numbers, as required by 27 above), and a test is made by name in a
983        conditional pattern, either for a subpattern having been matched, or for
984        recursion in such a pattern, all the associated numbered subpatterns are
985        tested, and the overall condition is true if the condition is true for any
986        one of them. This is the way Perl works, and is also more like the way
987        testing by number works.
988    
989    
990    Version 7.9 11-Apr-09
991    ---------------------
992    
993    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
994        (pcretest), all targets were linked against these libraries. This included
995        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
996        libraries. This caused unwanted dependencies to be created. This problem
997        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
998        pcretest is linked with readline.
999    
1000    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1001        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1002        moved outside it again, because FALSE and TRUE are already defined in AIX,
1003        but BOOL is not.
1004    
1005    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1006        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1007    
1008    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1009        hyphens) following file names and line numbers when outputting matching
1010        lines. This is not true; no spaces are inserted. I have also clarified the
1011        wording for the --colour (or --color) option.
1012    
1013    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1014        was not coloured; this is different to GNU grep, so I have changed it to be
1015        the same.
1016    
1017    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1018        each matching line was coloured. Now it goes on to look for further matches
1019        of any of the test patterns, which is the same behaviour as GNU grep.
1020    
1021    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1022        doesn't make sense to accept an empty string match in pcregrep, so I have
1023        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1024        seems to be how GNU grep behaves.
1025    
1026    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1027        start or after a newline", because the conditional assertion was not being
1028        correctly handled. The rule now is that both the assertion and what follows
1029        in the first alternative must satisfy the test.
1030    
1031    9.  If auto-callout was enabled in a pattern with a conditional group whose
1032        condition was an assertion, PCRE could crash during matching, both with
1033        pcre_exec() and pcre_dfa_exec().
1034    
1035    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1036        used for matching.
1037    
1038    11. Unicode property support in character classes was not working for
1039        characters (bytes) greater than 127 when not in UTF-8 mode.
1040    
1041    12. Added the -M command line option to pcretest.
1042    
1043    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1044    
1045    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1046    
1047    16. Added comments and documentation about mis-use of no_arg in the C++
1048        wrapper.
1049    
1050    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1051        from Martin Jerabek that uses macro names for all relevant character and
1052        string constants.
1053    
1054    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1055        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1056        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1057        these, but not everybody uses configure.
1058    
1059    19. A conditional group that had only one branch was not being correctly
1060        recognized as an item that could match an empty string. This meant that an
1061        enclosing group might also not be so recognized, causing infinite looping
1062        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1063        with the subject "ab", where knowledge that the repeated group can match
1064        nothing is needed in order to break the loop.
1065    
1066    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1067        exec(), but without supplying a callout function, matching went wrong.
1068    
1069    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1070        leak if the size of the offset vector was greater than 30. When the vector
1071        is smaller, the saved offsets during recursion go onto a local stack
1072        vector, but for larger vectors malloc() is used. It was failing to free
1073        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1074        error, in fact).
1075    
1076    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1077        heapframe that is used only when UTF-8 support is enabled. This caused no
1078        problem, but was untidy.
1079    
1080    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1081        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1082        included within another project.
1083    
1084    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1085        slightly modified by me:
1086    
1087          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1088              not building pcregrep.
1089    
1090          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1091              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1092    
1093    25. Forward references, both numeric and by name, in patterns that made use of
1094        duplicate group numbers, could behave incorrectly or give incorrect errors,
1095        because when scanning forward to find the reference group, PCRE was not
1096        taking into account the duplicate group numbers. A pattern such as
1097        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1098    
1099    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1100        the feature of a custom pointer more persuasive (as requested by a user).
1101    
1102    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1103        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1104        pre-defined.
1105    
1106    28. Added support for (*UTF8) at the start of a pattern.
1107    
1108    29. Arrange for flags added by the "release type" setting in CMake to be shown
1109        in the configuration summary.
1110    
1111    
1112    Version 7.8 05-Sep-08
1113    ---------------------
1114    
1115    1.  Replaced UCP searching code with optimized version as implemented for Ad
1116        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1117        stage table and inline lookup instead of a function, giving speed ups of 2
1118        to 5 times on some simple patterns that I tested. Permission was given to
1119        distribute the MultiStage2.py script that generates the tables (it's not in
1120        the tarball, but is in the Subversion repository).
1121    
1122    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1123        scripts.
1124    
1125    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1126        a group with a zero qualifier. The result of the study could be incorrect,
1127        or the function might crash, depending on the pattern.
1128    
1129    4.  Caseless matching was not working for non-ASCII characters in back
1130        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1131        It now works when Unicode Property Support is available.
1132    
1133    5.  In pcretest, an escape such as \x{de} in the data was always generating
1134        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1135        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1136        truncation.
1137    
1138    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1139    
1140    7.  Added two (int) casts to pcregrep when printing the difference of two
1141        pointers, in case they are 64-bit values.
1142    
1143    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1144        test 2 if it fails.
1145    
1146    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1147        and a #define of that name to empty if it is not externally set. This is to
1148        allow users of MSVC to set it if necessary.
1149    
1150    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1151        the convenience functions in the pcre_get.c source file.
1152    
1153    11. An option change at the start of a pattern that had top-level alternatives
1154        could cause overwriting and/or a crash. This command provoked a crash in
1155        some environments:
1156    
1157          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1158    
1159        This potential security problem was recorded as CVE-2008-2371.
1160    
1161    12. For a pattern where the match had to start at the beginning or immediately
1162        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1163        pcre_dfa_exec() could read past the end of the passed subject if there was
1164        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1165        pcretest so that it places the subject at the end of its malloc-ed buffer.
1166    
1167    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1168        exec() might read past the end of the data buffer in UTF-8 mode.
1169    
1170    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1171        the data contained the byte 0x85 as part of a UTF-8 character within its
1172        first line. This applied both to normal and DFA matching.
1173    
1174    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1175        /^[^d]*?$/8 failed to match "abc".
1176    
1177    16. Added a missing copyright notice to pcrecpp_internal.h.
1178    
1179    17. Make it more clear in the documentation that values returned from
1180        pcre_exec() in ovector are byte offsets, not character counts.
1181    
1182    18. Tidied a few places to stop certain compilers from issuing warnings.
1183    
1184    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1185        supplied by Stefan Weber. I made a further small update for 7.8 because
1186        there is a change of source arrangements: the pcre_searchfuncs.c module is
1187        replaced by pcre_ucd.c.
1188    
1189    
1190    Version 7.7 07-May-08
1191    ---------------------
1192    
1193    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1194        a string to a long long, pretend we don't even have a long long." This is
1195        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1196    
1197    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1198        pre-7.6 versions, which defined a global no_arg variable instead of putting
1199        it in the RE class. (See also #8 below.)
1200    
1201    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1202        Lopes.
1203    
1204    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1205    
1206        (1) The include/exclude patterns were being applied to the whole pathnames
1207            of files, instead of just to the final components.
1208    
1209        (2) If there was more than one level of directory, the subdirectories were
1210            skipped unless they satisfied the include/exclude conditions. This is
1211            inconsistent with GNU grep (and could even be seen as contrary to the
1212            pcregrep specification - which I improved to make it absolutely clear).
1213            The action now is always to scan all levels of directory, and just
1214            apply the include/exclude patterns to regular files.
1215    
1216    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1217        --exclude_dir in the tests to avoid scanning .svn directories.
1218    
1219    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1220        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1221        doesn't support NULs in patterns.
1222    
1223    7.  Added some missing "const"s to declarations of static tables in
1224        pcre_compile.c and pcre_dfa_exec.c.
1225    
1226    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1227        caused by fix #2  above. (Subsequently also a second patch to fix the
1228        first patch. And a third patch - this was a messy problem.)
1229    
1230    9.  Applied Craig's patch to remove the use of push_back().
1231    
1232    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1233        matching function regexec().
1234    
1235    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1236        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1237        references. PCRE supports relative numbers with this syntax (I don't think
1238        Oniguruma does).
1239    
1240    12. Previously, a group with a zero repeat such as (...){0} was completely
1241        omitted from the compiled regex. However, this means that if the group
1242        was called as a subroutine from elsewhere in the pattern, things went wrong
1243        (an internal error was given). Such groups are now left in the compiled
1244        pattern, with a new opcode that causes them to be skipped at execution
1245        time.
1246    
1247    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1248        to the way PCRE behaves:
1249    
1250        (a) A lone ] character is dis-allowed (Perl treats it as data).
1251    
1252        (b) A back reference to an unmatched subpattern matches an empty string
1253            (Perl fails the current match path).
1254    
1255        (c) A data ] in a character class must be notated as \] because if the
1256            first data character in a class is ], it defines an empty class. (In
1257            Perl it is not possible to have an empty class.) The empty class []
1258            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1259            The negative empty class [^] matches any one character, independently
1260            of the DOTALL setting.
1261    
1262    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1263        non-existent subpattern following a character class starting with ']' and
1264        containing () gave an internal compiling error instead of "reference to
1265        non-existent subpattern". Fortunately, when the pattern did exist, the
1266        compiled code was correct. (When scanning forwards to check for the
1267        existencd of the subpattern, it was treating the data ']' as terminating
1268        the class, so got the count wrong. When actually compiling, the reference
1269        was subsequently set up correctly.)
1270    
1271    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1272        it was being rejected as not supported by pcre_dfa_exec(), even though
1273        other assertions are supported. I have made pcre_dfa_exec() support
1274        (*FAIL).
1275    
1276    16. The implementation of 13c above involved the invention of a new opcode,
1277        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1278        cannot be changed at match time, I realized I could make a small
1279        improvement to matching performance by compiling OP_ALLANY instead of
1280        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1281        on the OP_ANY path.
1282    
1283    17. Compiling pcretest on Windows with readline support failed without the
1284        following two fixes: (1) Make the unistd.h include conditional on
1285        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1286    
1287    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1288        ncurses library to be included for pcretest when ReadLine support is
1289        requested, but also to allow for it to be overridden. This patch came from
1290        Daniel Bergström.
1291    
1292    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1293        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1294        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1295        spotting this.
1296    
1297    
1298    Version 7.6 28-Jan-08
1299    ---------------------
1300    
1301    1.  A character class containing a very large number of characters with
1302        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1303        overflow.
1304    
1305    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1306        HAVE_LONG_LONG is not defined.
1307    
1308    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1309        bring them up to date and include new features. This patch includes:
1310    
1311        - Fixed PH's badly added libz and libbz2 support.
1312        - Fixed a problem with static linking.
1313        - Added pcredemo. [But later removed - see 7 below.]
1314        - Fixed dftables problem and added an option.
1315        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1316            HAVE_LONG_LONG.
1317        - Added readline support for pcretest.
1318        - Added an listing of the option settings after cmake has run.
1319    
1320    4.  A user submitted a patch to Makefile that makes it easy to create
1321        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1322        Makefile.am that cause it to include this special target, without
1323        affecting anything else. Note that the same mingw target plus all
1324        the other distribution libraries and programs are now supported
1325        when configuring with CMake (see 6 below) instead of with
1326        Configure/Make.
1327    
1328    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1329        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1330        exported in the Windows port". It has not yet been confirmed that the patch
1331        solves the problem, but it does no harm.
1332    
1333    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1334        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1335        with CMake, and also correct the comment about stack recursion.
1336    
1337    7.  Remove the automatic building of pcredemo from the ./configure system and
1338        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1339        of a program that users should build themselves after PCRE is installed, so
1340        building it automatically is not really right. What is more, it gave
1341        trouble in some build environments.
1342    
1343    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1344    
1345    
1346    Version 7.5 10-Jan-08
1347    ---------------------
1348    
1349    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1350        values in parens when parsing an RE using the C++ wrapper."
1351    
1352    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1353        Characters greater than 255 were excluded from the class instead of being
1354        included.
1355    
1356    3.  The same bug as (2) above applied to negated POSIX classes such as
1357        [:^space:].
1358    
1359    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1360        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1361        I have changed it.
1362    
1363    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1364        first named subpattern) and a construct such as (?&a) would reference the
1365        first named subpattern whose name started with "a" (in other words, the
1366        length check was missing). Both these problems are fixed. "Subpattern name
1367        expected" is now given for (?&) (a zero-length name), and this patch also
1368        makes it give the same error for \k'' (previously it complained that that
1369        was a reference to a non-existent subpattern).
1370    
1371    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1372        this is right because (?- can be followed by option settings as well as by
1373        digits. I have, however, made the messages clearer.
1374    
1375    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1376        than the number used in the conditional) now cause a compile-time error.
1377        This is actually not compatible with Perl, which accepts such patterns, but
1378        treats the conditional as always being FALSE (as PCRE used to), but it
1379        seems to me that giving a diagnostic is better.
1380    
1381    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1382        and messages.
1383    
1384    9.  Fix two occurrences of "backslash" in comments that should have been
1385        "backspace".
1386    
1387    10. Remove two redundant lines of code that can never be obeyed (their function
1388        was moved elsewhere).
1389    
1390    11. The program that makes PCRE's Unicode character property table had a bug
1391        which caused it to generate incorrect table entries for sequences of
1392        characters that have the same character type, but are in different scripts.
1393        It amalgamated them into a single range, with the script of the first of
1394        them. In other words, some characters were in the wrong script. There were
1395        thirteen such cases, affecting characters in the following ranges:
1396    
1397          U+002b0 - U+002c1
1398          U+0060c - U+0060d
1399          U+0061e - U+00612
1400          U+0064b - U+0065e
1401          U+0074d - U+0076d
1402          U+01800 - U+01805
1403          U+01d00 - U+01d77
1404          U+01d9b - U+01dbf
1405          U+0200b - U+0200f
1406          U+030fc - U+030fe
1407          U+03260 - U+0327f
1408          U+0fb46 - U+0fbb1
1409          U+10450 - U+1049d
1410    
1411    12. The -o option (show only the matching part of a line) for pcregrep was not
1412        compatible with GNU grep in that, if there was more than one match in a
1413        line, it showed only the first of them. It now behaves in the same way as
1414        GNU grep.
1415    
1416    13. If the -o and -v options were combined for pcregrep, it printed a blank
1417        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1418        does the same. The return code can be used to tell if there were any
1419        non-matching lines.
1420    
1421    14. Added --file-offsets and --line-offsets to pcregrep.
1422    
1423    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1424        infinitely looping recursion. The bug was that positive lookaheads were not
1425        being skipped when checking for a possible empty match (negative lookaheads
1426        and both kinds of lookbehind were skipped).
1427    
1428    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1429        inclusion of <windows.h> to before rather than after the definition of
1430        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1431    
1432    17. Specifying a possessive quantifier with a specific limit for a Unicode
1433        character property caused pcre_compile() to compile bad code, which led at
1434        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1435        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1436        caused the error; without that there was no problem.
1437    
1438    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1439    
1440    19. Added --enable-pcretest-libreadline.
1441    
1442    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1443        RE::GlobalReplace(). As a result, the number of replacements returned was
1444        double what it should be. I removed one of the increments, but Craig sent a
1445        later patch that removed the other one (the right fix) and added unit tests
1446        that check the return values (which was not done before).
1447    
1448    21. Several CMake things:
1449    
1450        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1451            the names libpcre and libpcreposix, not just pcre and pcreposix.
1452    
1453        (2) The above change means that pcretest and pcregrep are now correctly
1454            linked with the newly-built libraries, not previously installed ones.
1455    
1456        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1457    
1458    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1459        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1460        UTF-8 newline character). The key issue is that the pattern starts .*;
1461        this means that the match must be either at the beginning, or after a
1462        newline. The bug was in the code for advancing after a failed match and
1463        checking that the new position followed a newline. It was not taking
1464        account of UTF-8 characters correctly.
1465    
1466    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1467        character classes. PCRE was not treating the sequence [:...:] as a
1468        character class unless the ... were all letters. Perl, however, seems to
1469        allow any characters between [: and :], though of course it rejects as
1470        unknown any "names" that contain non-letters, because all the known class
1471        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1472        for example, whereas PCRE did not - it did not recognize a POSIX character
1473        class. This seemed a bit dangerous, so the code has been changed to be
1474        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1475        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1476        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1477        Perl does, and where it didn't before.
1478    
1479    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1480        Windows environments %n is disabled by default.
1481    
1482    
1483    Version 7.4 21-Sep-07
1484    ---------------------
1485    
1486    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1487        means that a class such as [\s] counted as "explicit reference to CR or
1488        LF". That isn't really right - the whole point of the change was to try to
1489        help when there was an actual mention of one of the two characters. So now
1490        the change happens only if \r or \n (or a literal CR or LF) character is
1491        encountered.
1492    
1493    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1494        of both had grown to the point where there were only 3 bits left.
1495        Fortunately, there was spare space in the data structure, and so I have
1496        moved the internal flags into a new 16-bit field to free up more option
1497        bits.
1498    
1499    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1500        but did not set the internal JCHANGED flag - either of these is enough to
1501        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1502        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1503        start sets both bits.
1504    
1505    4.  Added options (at build time, compile time, exec time) to change \R from
1506        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1507    
1508    5.  doc/pcresyntax.html was missing from the distribution.
1509    
1510    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1511        compatibility, even though it is no longer used.
1512    
1513    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1514        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1515        windows.h file is present (where different names are used). [This was
1516        reversed later after testing - see 16 below.]
1517    
1518    8.  Changed all #include <config.h> to #include "config.h". There were also
1519        some further <pcre.h> cases that I changed to "pcre.h".
1520    
1521    9.  When pcregrep was used with the --colour option, it missed the line ending
1522        sequence off the lines that it output.
1523    
1524    10. It was pointed out to me that arrays of string pointers cause lots of
1525        relocations when a shared library is dynamically loaded. A technique of
1526        using a single long string with a table of offsets can drastically reduce
1527        these. I have refactored PCRE in four places to do this. The result is
1528        dramatic:
1529    
1530          Originally:                          290
1531          After changing UCP table:            187
1532          After changing error message table:   43
1533          After changing table of "verbs"       36
1534          After changing table of Posix names   22
1535    
1536        Thanks to the folks working on Gregex for glib for this insight.
1537    
1538    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1539        unicode-properties was also set.
1540    
1541    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1542    
1543    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1544        checked only for CRLF.
1545    
1546    14. Added casts to pcretest.c to avoid compiler warnings.
1547    
1548    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1549    
1550    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1551        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1552        entirely. This removes changes made in 7 above.
1553    
1554    17. The CMake files have been updated, and there is now more information about
1555        building with CMake in the NON-UNIX-USE document.
1556    
1557    
1558    Version 7.3 28-Aug-07
1559    ---------------------
1560    
1561     1. In the rejigging of the build system that eventually resulted in 7.1, the
1562        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1563        brackets there is not right, since it causes compilers to look for an
1564        installed pcre.h, not the version that is in the source that is being
1565        compiled (which of course may be different). I have changed it back to:
1566    
1567          #include "pcre.h"
1568    
1569        I have a vague recollection that the change was concerned with compiling in
1570        different directories, but in the new build system, that is taken care of
1571        by the VPATH setting the Makefile.
1572    
1573     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1574        when the subject happened to end in the byte 0x85 (e.g. if the last
1575        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1576        characters but of course it shouldn't be taken as a newline when it is part
1577        of another character. The bug was that, for an unlimited repeat of . in
1578        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1579        characters when looking for a newline.
1580    
1581     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1582    
1583     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1584        in debug output.
1585    
1586     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1587        long printing in the pcrecpp unittest when running under MinGW.
1588    
1589     6. ESC_K was left out of the EBCDIC table.
1590    
1591     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1592        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1593        limit also applies to "virtual nesting" when a pattern is recursive, and in
1594        this case 1000 isn't so big. I have been able to remove this limit at the
1595        expense of backing off one optimization in certain circumstances. Normally,
1596        when pcre_exec() would call its internal match() function recursively and
1597        immediately return the result unconditionally, it uses a "tail recursion"
1598        feature to save stack. However, when a subpattern that can match an empty
1599        string has an unlimited repetition quantifier, it no longer makes this
1600        optimization. That gives it a stack frame in which to save the data for
1601        checking that an empty string has been matched. Previously this was taken
1602        from the 1000-entry workspace that had been reserved. So now there is no
1603        explicit limit, but more stack is used.
1604    
1605     8. Applied Daniel's patches to solve problems with the import/export magic
1606        syntax that is required for Windows, and which was going wrong for the
1607        pcreposix and pcrecpp parts of the library. These were overlooked when this
1608        problem was solved for the main library.
1609    
1610     9. There were some crude static tests to avoid integer overflow when computing
1611        the size of patterns that contain repeated groups with explicit upper
1612        limits. As the maximum quantifier is 65535, the maximum group length was
1613        set at 30,000 so that the product of these two numbers did not overflow a
1614        32-bit integer. However, it turns out that people want to use groups that
1615        are longer than 30,000 bytes (though not repeat them that many times).
1616        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1617        made it possible to implement the integer overflow checks in a much more
1618        dynamic way, which I have now done. The artificial limitation on group
1619        length has been removed - we now have only the limit on the total length of
1620        the compiled pattern, which depends on the LINK_SIZE setting.
1621    
1622    10. Fixed a bug in the documentation for get/copy named substring when
1623        duplicate names are permitted. If none of the named substrings are set, the
1624        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1625        empty string.
1626    
1627    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1628        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1629        because the ] is interpreted as the first data character and the
1630        terminating ] is not found. PCRE has been made compatible with Perl in this
1631        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1632        cause memory overwriting.
1633    
1634    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1635        string has been matched (to stop an infinite loop). It was not recognizing
1636        a conditional subpattern that could match an empty string if that
1637        subpattern was within another subpattern. For example, it looped when
1638        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1639        condition was not nested. This bug has been fixed.
1640    
1641    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1642        past the start of the subject in the presence of bytes with the top bit
1643        set, for example "\x8aBCD".
1644    
1645    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1646        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1647    
1648    14. Optimized (?!) to (*FAIL).
1649    
1650    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1651        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1652        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1653        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1654        does: it's just the validity check that is more restrictive.
1655    
1656    16. Inserted checks for integer overflows during escape sequence (backslash)
1657        processing, and also fixed erroneous offset values for syntax errors during
1658        backslash processing.
1659    
1660    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1661        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1662    
1663    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1664        caused an overrun.
1665    
1666    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1667        something other than just ASCII characters) inside a group that had an
1668        unlimited repeat caused a loop at compile time (while checking to see
1669        whether the group could match an empty string).
1670    
1671    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1672        [\P{Any}] did so. (Error in the code for printing property names.)
1673    
1674    21. An orphan \E inside a character class could cause a crash.
1675    
1676    22. A repeated capturing bracket such as (A)? could cause a wild memory
1677        reference during compilation.
1678    
1679    23. There are several functions in pcre_compile() that scan along a compiled
1680        expression for various reasons (e.g. to see if it's fixed length for look
1681        behind). There were bugs in these functions when a repeated \p or \P was
1682        present in the pattern. These operators have additional parameters compared
1683        with \d, etc, and these were not being taken into account when moving along
1684        the compiled data. Specifically:
1685    
1686        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1687            length.
1688    
1689        (b) An item such as \pL+ within a repeated group could cause crashes or
1690            loops.
1691    
1692        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1693            "reference to non-existent subpattern" error.
1694    
1695        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1696    
1697    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1698        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1699    
1700    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1701    
1702    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1703        character were causing crashes (broken optimization).
1704    
1705    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1706        \p or \P) caused a compile-time loop.
1707    
1708    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1709        break. For example, the unstudied pattern [\r\n]A does not match the string
1710        "\r\nA" because change 7.0/46 below moves the current point on by two
1711        characters after failing to match at the start. However, the pattern \nA
1712        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1713        the same is true. There doesn't seem any very clean way out of this, but
1714        what I have chosen to do makes the common cases work: PCRE now takes note
1715        of whether there can be an explicit match for \r or \n anywhere in the
1716        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1717        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1718        pattern has explicit CR or LF references.
1719    
1720    29. Added (*CR) etc for changing newline setting at start of pattern.
1721    
1722    
1723    Version 7.2 19-Jun-07
1724    ---------------------
1725    
1726     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1727        which is apparently normally available under Windows.
1728    
1729     2. Re-jig the pcregrep tests with different newline settings in an attempt
1730        to make them independent of the local environment's newline setting.
1731    
1732     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1733    
1734     4. Some of the "internals" tests were previously cut out when the link size
1735        was not 2, because the output contained actual offsets. The recent new
1736        "Z" feature of pcretest means that these can be cut out, making the tests
1737        usable with all link sizes.
1738    
1739     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1740        stack recursion. This gives a massive performance boost under BSD, but just
1741        a small improvement under Linux. However, it saves one field in the frame
1742        in all cases.
1743    
1744     6. Added more features from the forthcoming Perl 5.10:
1745    
1746        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1747            recursion call. It refers to the nth most recently opened parentheses.
1748    
1749        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1750            to be opened parentheses.
1751    
1752        (c) Conditions that refer to capturing parentheses can be specified
1753            relatively, for example, (?(-2)... or (?(+3)...
1754    
1755        (d) \K resets the start of the current match so that everything before
1756            is not part of it.
1757    
1758        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1759    
1760        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1761            reference syntax.
1762    
1763        (g) (?| introduces a group in which the numbering of parentheses in each
1764            alternative starts with the same number.
1765    
1766        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1767    
1768     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1769        PCRE_INFO_JCHANGED.
1770    
1771     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1772        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1773        for detecting groups that can match an empty string.
1774    
1775     9. A pattern with a very large number of alternatives (more than several
1776        hundred) was running out of internal workspace during the pre-compile
1777        phase, where pcre_compile() figures out how much memory will be needed. A
1778        bit of new cunning has reduced the workspace needed for groups with
1779        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1780        workspace instead of running out of the 4096 that are available.
1781    
1782    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1783    
1784    11. Applied patch from Google to remove an optimization that didn't quite work.
1785        The report of the bug said:
1786    
1787          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1788          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1789          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1790    
1791    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1792        it matched the wrong number of bytes.
1793    
1794    
1795    Version 7.1 24-Apr-07
1796    ---------------------
1797    
1798     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1799        that is more "standard", making use of automake and other Autotools. There
1800        is some re-arrangement of the files and adjustment of comments consequent
1801        on this.
1802    
1803     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1804        for recursive directory scanning broke on some systems because the files
1805        are not scanned in any specific order and on different systems the order
1806        was different. A call to "sort" has been inserted into RunGrepTest for the
1807        approprate test as a short-term fix. In the longer term there may be an
1808        alternative.
1809    
1810     3. I had an email from Eric Raymond about problems translating some of PCRE's
1811        man pages to HTML (despite the fact that I distribute HTML pages, some
1812        people do their own conversions for various reasons). The problems
1813        concerned the use of low-level troff macros .br and .in. I have therefore
1814        removed all such uses from the man pages (some were redundant, some could
1815        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1816        HTML has been updated to handle .nf/.fi and to complain if it encounters
1817        .br or .in.
1818    
1819     4. Updated comments in configure.ac that get placed in config.h.in and also
1820        arranged for config.h to be included in the distribution, with the name
1821        config.h.generic, for the benefit of those who have to compile without
1822        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1823    
1824     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1825        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1826        makevp.bat for the current PCRE, using the additional files
1827        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1828    
1829     6. A Windows user reported a minor discrepancy with test 2, which turned out
1830        to be caused by a trailing space on an input line that had got lost in his
1831        copy. The trailing space was an accident, so I've just removed it.
1832    
1833     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1834        that is needed.
1835    
1836     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1837        as "const" (a) because they are and (b) because it helps the PHP
1838        maintainers who have recently made a script to detect big data structures
1839        in the php code that should be moved to the .rodata section. I remembered
1840        to update Builducptable as well, so it won't revert if ucptable.h is ever
1841        re-created.
1842    
1843     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1844        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1845        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1846        support is not required. This saves 1.5-2K of code, which is important in
1847        some applications.
1848    
1849        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1850        so as not to refer to the tables, even though these functions will never be
1851        called when UTF-8 support is disabled. Otherwise there are problems with a
1852        shared library.
1853    
1854    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1855    
1856        (a) It was defining its arguments as char * instead of void *.
1857    
1858        (b) It was assuming that all moves were upwards in memory; this was true
1859            a long time ago when I wrote it, but is no longer the case.
1860    
1861        The emulated memove() is provided for those environments that have neither
1862        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1863        is clearly not the case, as these two bugs were recently reported.
1864    
1865    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1866        and Detrail to create the HTML documentation, the .txt form of the man
1867        pages, and it removes trailing spaces from listed files. It also creates
1868        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1869        case, it wraps all the #defines with #ifndefs. This script should be run
1870        before "make dist".
1871    
1872    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1873        with Unicode property support.
1874    
1875        (a) For a maximizing quantifier, if the two different cases of the
1876            character were of different lengths in their UTF-8 codings (there are
1877            some cases like this - I found 11), and the matching function had to
1878            back up over a mixture of the two cases, it incorrectly assumed they
1879            were both the same length.
1880    
1881        (b) When PCRE was configured to use the heap rather than the stack for
1882            recursion during matching, it was not correctly preserving the data for
1883            the other case of a UTF-8 character when checking ahead for a match
1884            while processing a minimizing repeat. If the check also involved
1885            matching a wide character, but failed, corruption could cause an
1886            erroneous result when trying to check for a repeat of the original
1887            character.
1888    
1889    13. Some tidying changes to the testing mechanism:
1890    
1891        (a) The RunTest script now detects the internal link size and whether there
1892            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1893            values substituted by "configure". (The RunGrepTest script already did
1894            this for UTF-8.) The configure.ac script no longer substitutes the
1895            relevant variables.
1896    
1897        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1898            with length and offset values. This means that the output is different
1899            for different internal link sizes. Test 2 is skipped for link sizes
1900            other than 2 because of this, bypassing the problem. Unfortunately,
1901            there was also a test in test 3 (the locale tests) that used /B and
1902            failed for link sizes other than 2. Rather than cut the whole test out,
1903            I have added a new /Z option to pcretest that replaces the length and
1904            offset values with spaces. This is now used to make test 3 independent
1905            of link size. (Test 2 will be tidied up later.)
1906    
1907    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1908        segmentation fault instead of returning the appropriate error message.
1909    
1910    15. In multiline mode when the newline sequence was set to "any", the pattern
1911        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1912        This doesn't seem right; it now treats the CRLF combination as the line
1913        ending, and so does not match in that case. It's only a pattern such as ^$
1914        that would hit this one: something like ^ABC$ would have failed after \r
1915        and then tried again after \r\n.
1916    
1917    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1918        in an attempt to make files that differ only in their line terminators
1919        compare equal. This works on Linux.
1920    
1921    17. Under certain error circumstances pcregrep might try to free random memory
1922        as it exited. This is now fixed, thanks to valgrind.
1923    
1924    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1925        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1926        was because its rules for how to advance for /g after matching an empty
1927        string at the end of a line did not allow for this case. They now check for
1928        it specially.
1929    
1930    20. pcretest is supposed to handle patterns and data of any length, by
1931        extending its buffers when necessary. It was getting this wrong when the
1932        buffer for a data line had to be extended.
1933    
1934    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1935        CRLF as a newline sequence.
1936    
1937    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1938        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1939        I have nevertheless tidied it up.
1940    
1941    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1942    
1943    24. Added a man page for pcre-config.
1944    
1945    
1946    Version 7.0 19-Dec-06
1947    ---------------------
1948    
1949     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1950        moving to gcc 4.1.1.
1951    
1952     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1953        sys/time.h, which is documented as needed for this function. It doesn't
1954        seem to matter on Linux, but it showed up on some releases of OS X.
1955    
1956     3. It seems that there are systems where bytes whose values are greater than
1957        127 match isprint() in the "C" locale. The "C" locale should be the
1958        default when a C program starts up. In most systems, only ASCII printing
1959        characters match isprint(). This difference caused the output from pcretest
1960        to vary, making some of the tests fail. I have changed pcretest so that:
1961    
1962        (a) When it is outputting text in the compiled version of a pattern, bytes
1963            other than 32-126 are always shown as hex escapes.
1964    
1965        (b) When it is outputting text that is a matched part of a subject string,
1966            it does the same, unless a different locale has been set for the match
1967            (using the /L modifier). In this case, it uses isprint() to decide.
1968    
1969     4. Fixed a major bug that caused incorrect computation of the amount of memory
1970        required for a compiled pattern when options that changed within the
1971        pattern affected the logic of the preliminary scan that determines the
1972        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1973        that the computed length was too small. The symptoms of this bug were
1974        either the PCRE error "internal error: code overflow" from pcre_compile(),
1975        or a glibc crash with a message such as "pcretest: free(): invalid next
1976        size (fast)". Examples of patterns that provoked this bug (shown in
1977        pcretest format) are:
1978    
1979          /(?-x: )/x
1980          /(?x)(?-x: \s*#\s*)/
1981          /((?i)[\x{c0}])/8
1982          /(?i:[\x{c0}])/8
1983    
1984        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1985        is now done differently.
1986    
1987     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1988        wrapper classes; (b) implement a new function in the C++ scanner that is
1989        more efficient than the old way of doing things because it avoids levels of
1990        recursion in the regex matching; (c) add a paragraph to the documentation
1991        for the FullMatch() function.
1992    
1993     6. The escape sequence \n was being treated as whatever was defined as
1994        "newline". Not only was this contrary to the documentation, which states
1995        that \n is character 10 (hex 0A), but it also went horribly wrong when
1996        "newline" was defined as CRLF. This has been fixed.
1997    
1998     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1999        was being set to -1 for the "end of line" case (supposedly a value that no
2000        character can have). Though this value is never used (the check for end of
2001        line is "zero bytes in current character"), it caused compiler complaints.
2002        I've changed it to 0xffffffff.
2003    
2004     8. In pcre_version.c, the version string was being built by a sequence of
2005        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2006        string (as it is for production releases) called a macro with an empty
2007        argument. The C standard says the result of this is undefined. The gcc
2008        compiler treats it as an empty string (which was what was wanted) but it is
2009        reported that Visual C gives an error. The source has been hacked around to
2010        avoid this problem.
2011    
2012     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2013        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2014        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2015        of them did).
2016    
2017    10. Originally, pcretest opened its input and output without "b"; then I was
2018        told that "b" was needed in some environments, so it was added for release
2019        5.0 to both the input and output. (It makes no difference on Unix-like
2020        systems.) Later I was told that it is wrong for the input on Windows. I've
2021        now abstracted the modes into two macros, to make it easier to fiddle with
2022        them, and removed "b" from the input mode under Windows.
2023    
2024    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2025    
2026    12. Added -help and --help to pcretest as an official way of being reminded
2027        of the options.
2028    
2029    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2030        and pcrecpp.cc because they annoy compilers at high warning levels.
2031    
2032    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2033    
2034    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2035        scripts are not C programs :-) and which was not noticed because it works
2036        on Linux.
2037    
2038    16. pcretest is supposed to handle any length of pattern and data line (as one
2039        line or as a continued sequence of lines) by extending its input buffer if
2040        necessary. This feature was broken for very long pattern lines, leading to
2041        a string of junk being passed to pcre_compile() if the pattern was longer
2042        than about 50K.
2043    
2044    17. I have done a major re-factoring of the way pcre_compile() computes the
2045        amount of memory needed for a compiled pattern. Previously, there was code
2046        that made a preliminary scan of the pattern in order to do this. That was
2047        OK when PCRE was new, but as the facilities have expanded, it has become
2048        harder and harder to keep it in step with the real compile phase, and there
2049        have been a number of bugs (see for example, 4 above). I have now found a
2050        cunning way of running the real compile function in a "fake" mode that
2051        enables it to compute how much memory it would need, while actually only
2052        ever using a few hundred bytes of working memory and without too many
2053        tests of the mode. This should make future maintenance and development
2054        easier. A side effect of this work is that the limit of 200 on the nesting
2055        depth of parentheses has been removed (though this was never a serious
2056        limitation, I suspect). However, there is a downside: pcre_compile() now
2057        runs more slowly than before (30% or more, depending on the pattern). I
2058        hope this isn't a big issue. There is no effect on runtime performance.
2059    
2060    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2061        newline (only possible for the last line of a file) and it was a
2062        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2063    
2064    19. Added additional timing features to pcretest. (1) The -tm option now times
2065        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2066        separate command line item, by a number that specifies the number of
2067        repeats to use when timing. The default is 50000; this gives better
2068        precision, but takes uncomfortably long for very large patterns.
2069    
2070    20. Extended pcre_study() to be more clever in cases where a branch of a
2071        subpattern has no definite first character. For example, (a*|b*)[cd] would
2072        previously give no result from pcre_study(). Now it recognizes that the
2073        first character must be a, b, c, or d.
2074    
2075    21. There was an incorrect error "recursive call could loop indefinitely" if
2076        a subpattern (or the entire pattern) that was being tested for matching an
2077        empty string contained only one non-empty item after a nested subpattern.
2078        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2079        incorrectly, because the \d was being skipped in the check.
2080    
2081    22. The pcretest program now has a new pattern option /B and a command line
2082        option -b, which is equivalent to adding /B to every pattern. This causes
2083        it to show the compiled bytecode, without the additional information that
2084        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2085        is the same as /B/I).
2086    
2087    23. A new optimization is now able automatically to treat some sequences such
2088        as a*b as a*+b. More specifically, if something simple (such as a character
2089        or a simple class like \d) has an unlimited quantifier, and is followed by
2090        something that cannot possibly match the quantified thing, the quantifier
2091        is automatically "possessified".
2092    
2093    24. A recursive reference to a subpattern whose number was greater than 39
2094        went wrong under certain circumstances in UTF-8 mode. This bug could also
2095        have affected the operation of pcre_study().
2096    
2097    25. Realized that a little bit of performance could be had by replacing
2098        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2099    
2100    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2101    
2102    27. Possessive quantifiers such as a++ were previously implemented by turning
2103        them into atomic groups such as ($>a+). Now they have their own opcodes,
2104        which improves performance. This includes the automatically created ones
2105        from 23 above.
2106    
2107    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2108        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2109        the first matched character to be a colon. This applied both to named and
2110        numbered groups.
2111    
2112    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2113    
2114    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2115        building PCRE on Windows easier, so I have included it in the distribution.
2116    
2117    31. There is now a check in pcretest against a ridiculously large number being
2118        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2119        loop, the loop is abandoned.
2120    
2121    32. Forward references to subpatterns in conditions such as (?(2)...) where
2122        subpattern 2 is defined later cause pcre_compile() to search forwards in
2123        the pattern for the relevant set of parentheses. This search went wrong
2124        when there were unescaped parentheses in a character class, parentheses
2125        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2126    
2127    33. "Subroutine" calls and backreferences were previously restricted to
2128        referencing subpatterns earlier in the regex. This restriction has now
2129        been removed.
2130    
2131    34. Added a number of extra features that are going to be in Perl 5.10. On the
2132        whole, these are just syntactic alternatives for features that PCRE had
2133        previously implemented using the Python syntax or my own invention. The
2134        other formats are all retained for compatibility.
2135    
2136        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2137            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2138            also .NET compatible.
2139    
2140        (b) A recursion or subroutine call to a named group can now be defined as
2141            (?&name) as well as (?P>name).
2142    
2143        (c) A backreference to a named group can now be defined as \k<name> or
2144            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2145            5.10, are also .NET compatible.
2146    
2147        (d) A conditional reference to a named group can now use the syntax
2148            (?(<name>) or (?('name') as well as (?(name).
2149    
2150        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2151            groups (named and numbered) that are never evaluated inline, but can be
2152            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2153            is always false. There may be only one alternative in such a group.
2154    
2155        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2156            as the simple (?(R). The condition is true only if the most recent
2157            recursion is that of the given number or name. It does not search out
2158            through the entire recursion stack.
2159    
2160        (g) The escape \gN or \g{N} has been added, where N is a positive or
2161            negative number, specifying an absolute or relative reference.
2162    
2163    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2164        some "unreachable code" warnings.
2165    
2166    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2167        things, this adds five new scripts.
2168    
2169    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2170        There were also incompatibilities regarding the handling of \Q..\E inside
2171        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2172        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2173    
2174    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2175        matches an empty string, and forcibly breaks the loop. There were bugs in
2176        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2177        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2178        separate and independent bugs (that affected different cases) have been
2179        fixed.
2180    
2181    39. Refactored the code to abolish the use of different opcodes for small
2182        capturing bracket numbers. This is a tidy that I avoided doing when I
2183        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2184        The new approach is not only tidier, it makes it possible to reduce the
2185        memory needed to fix the previous bug (38).
2186    
2187    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2188        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2189        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2190        mode.
2191    
2192    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2193        report.
2194    
2195    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2196        copy construction and assignment in the C++ wrapper.
2197    
2198    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2199        couple of bugs that could have given wrong results in the "--newline=crlf"
2200        case.
2201    
2202    44. Added a number of casts and did some reorganization of signed/unsigned int
2203        variables following suggestions from Dair Grant. Also renamed the variable
2204        "this" as "item" because it is a C++ keyword.
2205    
2206    45. Arranged for dftables to add
2207    
2208          #include "pcre_internal.h"
2209    
2210        to pcre_chartables.c because without it, gcc 4.x may remove the array
2211        definition from the final binary if PCRE is built into a static library and
2212        dead code stripping is activated.
2213    
2214    46. For an unanchored pattern, if a match attempt fails at the start of a
2215        newline sequence, and the newline setting is CRLF or ANY, and the next two
2216        characters are CRLF, advance by two characters instead of one.
2217    
2218    
2219    Version 6.7 04-Jul-06
2220    ---------------------
2221    
2222     1. In order to handle tests when input lines are enormously long, pcretest has
2223        been re-factored so that it automatically extends its buffers when
2224        necessary. The code is crude, but this _is_ just a test program. The
2225        default size has been increased from 32K to 50K.
2226    
2227     2. The code in pcre_study() was using the value of the re argument before
2228        testing it for NULL. (Of course, in any sensible call of the function, it
2229        won't be NULL.)
2230    
2231     3. The memmove() emulation function in pcre_internal.h, which is used on
2232        systems that lack both memmove() and bcopy() - that is, hardly ever -
2233        was missing a "static" storage class specifier.
2234    
2235     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2236        containing an extended class (one that cannot be represented by a bitmap
2237        because it contains high-valued characters or Unicode property items, e.g.
2238        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2239        pattern, but PCRE should not loop if you do not (it no longer does).
2240        [Detail: two cases were found: (a) a repeated subpattern containing an
2241        extended class; (b) a recursive reference to a subpattern that followed a
2242        previous extended class. It wasn't skipping over the extended class
2243        correctly when UTF-8 mode was not set.]
2244    
2245     5. A negated single-character class was not being recognized as fixed-length
2246        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2247        compile error "lookbehind assertion is not fixed length".
2248    
2249     6. The RunPerlTest auxiliary script was showing an unexpected difference
2250        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2251        write a Perl script that can interpret lines of an input file either as
2252        byte characters or as UTF-8, which is what "perltest" was being required to
2253        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2254        can't do is switch easily at run time between having the "use utf8;" pragma
2255        or not. In the end, I fudged it by using the RunPerlTest script to insert
2256        "use utf8;" explicitly for the UTF-8 tests.
2257    
2258     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2259        the end of the subject string, contrary to the documentation and to what
2260        Perl does. This was true of both matching functions. Now it matches only at
2261        the start of the subject and immediately after *internal* newlines.
2262    
2263     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2264        a pointer to an int instead of a pointer to an unsigned long int. This
2265        caused problems on 64-bit systems.
2266    
2267     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2268        instance of the 'standard' template library not being so standard".
2269    
2270    10. There was no check on the number of named subpatterns nor the maximum
2271        length of a subpattern name. The product of these values is used to compute
2272        the size of the memory block for a compiled pattern. By supplying a very
2273        long subpattern name and a large number of named subpatterns, the size
2274        computation could be caused to overflow. This is now prevented by limiting
2275        the length of names to 32 characters, and the number of named subpatterns
2276        to 10,000.
2277    
2278    11. Subpatterns that are repeated with specific counts have to be replicated in
2279        the compiled pattern. The size of memory for this was computed from the
2280        length of the subpattern and the repeat count. The latter is limited to
2281        65535, but there was no limit on the former, meaning that integer overflow
2282        could in principle occur. The compiled length of a repeated subpattern is
2283        now limited to 30,000 bytes in order to prevent this.
2284    
2285    12. Added the optional facility to have named substrings with the same name.
2286    
2287    13. Added the ability to use a named substring as a condition, using the
2288        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2289        are numbers (not recommended). Forward references are permitted.
2290    
2291    14. Added forward references in named backreferences (if you see what I mean).
2292    
2293    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2294        pattern could run off the end of the subject. For example, the pattern
2295        "(?s)(.{1,5})"8 did this with the subject "ab".
2296    
2297    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2298        PCRE_CASELESS was set when matching characters that were quantified with ?
2299        or *.
2300    
2301    17. A character class other than a single negated character that had a minimum
2302        but no maximum quantifier - for example [ab]{6,} - was not handled
2303        correctly by pce_dfa_exec(). It would match only one character.
2304    
2305    18. A valid (though odd) pattern that looked like a POSIX character
2306        class but used an invalid character after [ (for example [[,abc,]]) caused
2307        pcre_compile() to give the error "Failed: internal error: code overflow" or
2308        in some cases to crash with a glibc free() error. This could even happen if
2309        the pattern terminated after [[ but there just happened to be a sequence of
2310        letters, a binary zero, and a closing ] in the memory that followed.
2311    
2312    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2313        over the years. Originally (before any Unicode support), just the bottom 8
2314        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2315        output from "man perlunicode" includes this:
2316    
2317          The regular expression compiler produces polymorphic opcodes.  That
2318          is, the pattern adapts to the data and automatically switches to
2319          the Unicode character scheme when presented with Unicode data--or
2320          instead uses a traditional byte scheme when presented with byte
2321          data.
2322    
2323        Sadly, a wide octal escape does not cause a switch, and in a string with
2324        no other multibyte characters, these octal escapes are treated as before.
2325        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2326        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2327        Unicode string.
2328    
2329        I have not perpetrated such confusion in PCRE. Up till now, it took just
2330        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2331        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2332        translate to the appropriate multibyte character.
2333    
2334    29. Applied some refactoring to reduce the number of warnings from Microsoft
2335        and Borland compilers. This has included removing the fudge introduced
2336        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2337        a warning about an unused variable.
2338    
2339    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2340        characters since release 4.0, because Perl (from release 5.004) does not.
2341        [Or at least, is documented not to: some releases seem to be in conflict
2342        with the documentation.] However, when a pattern was studied with
2343        pcre_study() and all its branches started with \s, PCRE still included VT
2344        as a possible starting character. Of course, this did no harm; it just
2345        caused an unnecessary match attempt.
2346    
2347    22. Removed a now-redundant internal flag bit that recorded the fact that case
2348        dependency changed within the pattern. This was once needed for "required
2349        byte" processing, but is no longer used. This recovers a now-scarce options
2350        bit. Also moved the least significant internal flag bit to the most-
2351        significant bit of the word, which was not previously used (hangover from
2352        the days when it was an int rather than a uint) to free up another bit for
2353        the future.
2354    
2355    23. Added support for CRLF line endings as well as CR and LF. As well as the
2356        default being selectable at build time, it can now be changed at runtime
2357        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2358        specify that it is scanning data with non-default line endings.
2359    
2360    24. Changed the definition of CXXLINK to make it agree with the definition of
2361        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2362    
2363    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2364        recursions. This makes a big different to stack usage for some patterns.
2365    
2366    26. If a subpattern containing a named recursion or subroutine reference such
2367        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2368        the space required for the compiled pattern went wrong and gave too small a
2369        value. Depending on the environment, this could lead to "Failed: internal
2370        error: code overflow at offset 49" or "glibc detected double free or
2371        corruption" errors.
2372    
2373    27. Applied patches from Google (a) to support the new newline modes and (b) to
2374        advance over multibyte UTF-8 characters in GlobalReplace.
2375    
2376    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2377        difference for some implementation of PCRE in some Windows version.
2378    
2379    29. Added some extra testing facilities to pcretest:
2380    
2381        \q<number>   in a data line sets the "match limit" value
2382        \Q<number>   in a data line sets the "match recursion limt" value
2383        -S <number>  sets the stack size, where <number> is in megabytes
2384    
2385        The -S option isn't available for Windows.
2386    
2387    
2388    Version 6.6 06-Feb-06
2389    ---------------------
2390    
2391     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2392        in pcreposix.h. I have copied the definition from pcre.h.
2393    
2394     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2395        because pcre.h is no longer a built file.
2396    
2397     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2398        not normally included in the compiled code.
2399    
2400    
2401    Version 6.5 01-Feb-06
2402    ---------------------
2403    
2404     1. When using the partial match feature with pcre_dfa_exec(), it was not
2405        anchoring the second and subsequent partial matches at the new starting
2406        point. This could lead to incorrect results. For example, with the pattern
2407        /1234/, partially matching against "123" and then "a4" gave a match.
2408    
2409     2. Changes to pcregrep:
2410    
2411        (a) All non-match returns from pcre_exec() were being treated as failures
2412            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2413            error message is output. Some extra information is given for the
2414            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2415            probably the only errors that are likely to be caused by users (by
2416            specifying a regex that has nested indefinite repeats, for instance).
2417            If there are more than 20 of these errors, pcregrep is abandoned.
2418    
2419        (b) A binary zero was treated as data while matching, but terminated the
2420            output line if it was written out. This has been fixed: binary zeroes
2421            are now no different to any other data bytes.
2422    
2423        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2424            used to set a locale for matching. The --locale=xxxx long option has
2425            been added (no short equivalent) to specify a locale explicitly on the
2426            pcregrep command, overriding the environment variables.
2427    
2428        (d) When -B was used with -n, some line numbers in the output were one less
2429            than they should have been.
2430    
2431        (e) Added the -o (--only-matching) option.
2432    
2433        (f) If -A or -C was used with -c (count only), some lines of context were
2434            accidentally printed for the final match.
2435    
2436        (g) Added the -H (--with-filename) option.
2437    
2438        (h) The combination of options -rh failed to suppress file names for files
2439            that were found from directory arguments.
2440    
2441        (i) Added the -D (--devices) and -d (--directories) options.
2442    
2443        (j) Added the -F (--fixed-strings) option.
2444    
2445        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2446    
2447        (l) Added the --colo(u)r option.
2448    
2449        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2450            is not present by default.
2451    
2452     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2453        items such as (?R) or (?1), when the recursion could match a number of
2454        alternatives. If it matched one of the alternatives, but subsequently,
2455        outside the recursion, there was a failure, the code tried to back up into
2456        the recursion. However, because of the way PCRE is implemented, this is not
2457        possible, and the result was an incorrect result from the match.
2458    
2459        In order to prevent this happening, the specification of recursion has
2460        been changed so that all such subpatterns are automatically treated as
2461        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2462    
2463     4. I had overlooked the fact that, in some locales, there are characters for
2464        which isalpha() is true but neither isupper() nor islower() are true. In
2465        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2466        and ordfeminine) are like this. This affected the treatment of \w and \W
2467        when they appeared in character classes, but not when they appeared outside
2468        a character class. The bit map for "word" characters is now created
2469        separately from the results of isalnum() instead of just taking it from the
2470        upper, lower, and digit maps. (Plus the underscore character, of course.)
2471    
2472     5. The above bug also affected the handling of POSIX character classes such as
2473        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2474        permanent tables. Instead, the bit maps for such a class were previously
2475        created as the appropriate unions of the upper, lower, and digit bitmaps.
2476        Now they are created by subtraction from the [[:word:]] class, which has
2477        its own bitmap.
2478    
2479     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2480        It is created by subtracting the vertical space characters (\x09, \x0a,
2481        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2482        subtraction was done in the overall bitmap for a character class, meaning
2483        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2484        be recognized. This bug has been fixed.
2485    
2486     7. Patches from the folks at Google:
2487    
2488          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2489          real life, but is still worth protecting against".
2490    
2491          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2492          regular expressions".
2493    
2494          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2495          have it.
2496    
2497          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2498          "configure" and the latter not, in order to fix a problem somebody had
2499          with compiling the Arg class on HP-UX.
2500    
2501          (e) Improve the error-handling of the C++ wrapper a little bit.
2502    
2503          (f) New tests for checking recursion limiting.
2504    
2505     8. The pcre_memmove() function, which is used only if the environment does not
2506        have a standard memmove() function (and is therefore rarely compiled),
2507        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2508        returning a result (though PCRE never actually uses the result).
2509    
2510     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2511        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2512        returned instead of calling malloc() with an overflowing number that would
2513        most likely cause subsequent chaos.
2514    
2515    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2516    
2517    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2518        with this option is matched, the nmatch and pmatch options of regexec() are
2519        ignored.
2520    
2521    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2522        provided in case anyone wants to the the POSIX interface with UTF-8
2523        strings.
2524    
2525    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2526        C++ linking (needed for some HP-UX environments).
2527    
2528    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2529        (unused parameter) and in the pcre_printint() function (omitted "default"
2530        switch label when the default is to do nothing).
2531    
2532    15. Added some code to make it possible, when PCRE is compiled as a C++
2533        library, to replace subject pointers for pcre_exec() with a smart pointer
2534        class, thus making it possible to process discontinuous strings.
2535    
2536    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2537        much the same function. They were added by different people who were trying
2538        to make PCRE easy to compile on non-Unix systems. It has been suggested
2539        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2540        for compiling on Windows systems. I have therefore replaced it with
2541        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2542        defaults to "extern" for C or "extern C" for C++, which works fine on
2543        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2544        SCOPE with something explicit in config.h. In addition:
2545    
2546        (a) pcreposix.h still had just "extern" instead of either of these macros;
2547            I have replaced it with PCRE_DATA_SCOPE.
2548    
2549        (b) Functions such as _pcre_xclass(), which are internal to the library,
2550            but external in the C sense, all had PCRE_EXPORT in their definitions.
2551            This is apparently wrong for the Windows case, so I have removed it.
2552            (It makes no difference on Unix-like systems.)
2553    
2554    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2555        of recursive calls to match(). This is different to MATCH_LIMIT because
2556        that limits the total number of calls to match(), not all of which increase
2557        the depth of recursion. Limiting the recursion depth limits the amount of
2558        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2559        when PCRE is compiled, and changed at run time. A patch from Google adds
2560        this functionality to the C++ interface.
2561    
2562    18. Changes to the handling of Unicode character properties:
2563    
2564        (a) Updated the table to Unicode 4.1.0.
2565    
2566        (b) Recognize characters that are not in the table as "Cn" (undefined).
2567    
2568        (c) I revised the way the table is implemented to a much improved format
2569            which includes recognition of ranges. It now supports the ranges that
2570            are defined in UnicodeData.txt, and it also amalgamates other
2571            characters into ranges. This has reduced the number of entries in the
2572            table from around 16,000 to around 3,000, thus reducing its size
2573            considerably. I realized I did not need to use a tree structure after
2574            all - a binary chop search is just as efficient. Having reduced the
2575            number of entries, I extended their size from 6 bytes to 8 bytes to
2576            allow for more data.
2577    
2578        (d) Added support for Unicode script names via properties such as \p{Han}.
2579    
2580    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2581        matching that character.
2582    
2583    20. When matching a repeated Unicode property with a minimum greater than zero,
2584        (for example \pL{2,}), PCRE could look past the end of the subject if it
2585        reached it while seeking the minimum number of characters. This could
2586        happen only if some of the characters were more than one byte long, because
2587        there is a check for at least the minimum number of bytes.
2588    
2589    21. Refactored the implementation of \p and \P so as to be more general, to
2590        allow for more different types of property in future. This has changed the
2591        compiled form incompatibly. Anybody with saved compiled patterns that use
2592        \p or \P will have to recompile them.
2593    
2594    22. Added "Any" and "L&" to the supported property types.
2595    
2596    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2597        but give a compile time error if the value is greater than 0xff.
2598    
2599    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2600        accidentally not being installed or uninstalled.
2601    
2602    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2603        made were to insert the current release number. This seemed silly, because
2604        it made things harder for people building PCRE on systems that don't run
2605        "configure". I have turned pcre.h into a distributed file, no longer built
2606        by "configure", with the version identification directly included. There is
2607        no longer a pcre.h.in file.
2608    
2609        However, this change necessitated a change to the pcre-config script as
2610        well. It is built from pcre-config.in, and one of the substitutions was the
2611        release number. I have updated configure.ac so that ./configure now finds
2612        the release number by grepping pcre.h.
2613    
2614    26. Added the ability to run the tests under valgrind.
2615    
2616    
2617    Version 6.4 05-Sep-05
2618    ---------------------
2619    
2620     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2621        "--" to be printed when multiple files were scanned, even when none of the
2622        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2623        consider it to be a bug, and have restored the previous behaviour.
2624    
2625     2. A couple of code tidies to get rid of compiler warnings.
2626    
2627     3. The pcretest program used to cheat by referring to symbols in the library
2628        whose names begin with _pcre_. These are internal symbols that are not
2629        really supposed to be visible externally, and in some environments it is
2630        possible to suppress them. The cheating is now confined to including
2631        certain files from the library's source, which is a bit cleaner.
2632    
2633     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2634        file's purpose clearer.
2635    
2636     5. Reorganized pcre_ucp_findchar().
2637    
2638    
2639    Version 6.3 15-Aug-05
2640    ---------------------
2641    
2642     1. The file libpcre.pc.in did not have general read permission in the tarball.
2643    
2644     2. There were some problems when building without C++ support:
2645    
2646        (a) If C++ support was not built, "make install" and "make test" still
2647            tried to test it.
2648    
2649        (b) There were problems when the value of CXX was explicitly set. Some
2650            changes have been made to try to fix these, and ...
2651    
2652        (c) --disable-cpp can now be used to explicitly disable C++ support.
2653    
2654        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2655            backslash in a target when C++ was disabled. This confuses some
2656            versions of "make", apparently. Using an intermediate variable solves
2657            this. (Same for CPP_LOBJ.)
2658    
2659     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2660        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2661        necessary on certain architectures.
2662    
2663     4. Added a setting of -export-symbols-regex to the link command to remove
2664        those symbols that are exported in the C sense, but actually are local
2665        within the library, and not documented. Their names all begin with
2666        "_pcre_". This is not a perfect job, because (a) we have to except some
2667        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2668        available (and never for static libraries). I have made a note to try to
2669        find a way round (a) in the future.
2670    
2671    
2672    Version 6.2 01-Aug-05
2673    ---------------------
2674    
2675     1. There was no test for integer overflow of quantifier values. A construction
2676        such as {1111111111111111} would give undefined results. What is worse, if
2677        a minimum quantifier for a parenthesized subpattern overflowed and became
2678        negative, the calculation of the memory size went wrong. This could have
2679        led to memory overwriting.
2680    
2681     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2682    
2683     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2684        operating environments where this matters.
2685    
2686     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2687        PCRE options from within the C++ wrapper.
2688    
2689     5. Named capturing subpatterns were not being correctly counted when a pattern
2690        was compiled. This caused two problems: (a) If there were more than 100
2691        such subpatterns, the calculation of the memory needed for the whole
2692        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2693        back references of the form \12, where the number was greater than 9, were
2694        not recognized as back references, even though there were sufficient
2695        previous subpatterns.
2696    
2697     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2698        versions of gcc, e.g. 2.95.4.
2699    
2700    
2701    Version 6.1 21-Jun-05
2702    ---------------------
2703    
2704     1. There was one reference to the variable "posix" in pcretest.c that was not
2705        surrounded by "#if !defined NOPOSIX".
2706    
2707     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2708        the cross-check on the old pcre_info() function, for the benefit of the
2709        cut-down version of PCRE that is currently imported into Exim.
2710    
2711     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2712        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2713        patterns that start (?i) but I don't think that matters. The use of (?i) is
2714        just an example; this all applies to the other options as well.
2715    
2716     4. Since libtool seems to echo the compile commands it is issuing, the output
2717        from "make" can be reduced a bit by putting "@" in front of each libtool
2718        compile command.
2719    
2720     5. Patch from the folks at Google for configure.in to be a bit more thorough
2721        in checking for a suitable C++ installation before trying to compile the
2722        C++ stuff. This should fix a reported problem when a compiler was present,
2723        but no suitable headers.
2724    
2725     6. The man pages all had just "PCRE" as their title. I have changed them to
2726        be the relevant file name. I have also arranged that these names are
2727        retained in the file doc/pcre.txt, which is a concatenation in text format
2728        of all the man pages except the little individual ones for each function.
2729    
2730     7. The NON-UNIX-USE file had not been updated for the different set of source
2731        files that come with release 6. I also added a few comments about the C++
2732        wrapper.
2733    
2734    
2735    Version 6.0 07-Jun-05
2736    ---------------------
2737    
2738     1. Some minor internal re-organization to help with my DFA experiments.
2739    
2740     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2741        didn't matter for the library itself when fully configured, but did matter
2742        when compiling without UCP support, or within Exim, where the ucp files are
2743        not imported.
2744    
2745     3. Refactoring of the library code to split up the various functions into
2746        different source modules. The addition of the new DFA matching code (see
2747        below) to a single monolithic source would have made it really too
2748        unwieldy, quite apart from causing all the code to be include in a
2749        statically linked application, when only some functions are used. This is
2750        relevant even without the DFA addition now that patterns can be compiled in
2751        one application and matched in another.
2752    
2753        The downside of splitting up is that there have to be some external
2754        functions and data tables that are used internally in different modules of
2755        the library but which are not part of the API. These have all had their
2756        names changed to start with "_pcre_" so that they are unlikely to clash
2757        with other external names.
2758    
2759     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2760        a different (DFA) algorithm. Although it is slower than the original
2761        function, it does have some advantages for certain types of matching
2762        problem.
2763    
2764     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2765        including restarting after a partial match.
2766    
2767     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2768        defined when compiling for Windows was sent to me. I have put it into the
2769        code, though I have no means of testing or verifying it.
2770    
2771     7. Added the pcre_refcount() auxiliary function.
2772    
2773     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2774        match before or at the first newline in the subject string. In pcretest,
2775        the /f option on a pattern can be used to set this.
2776    
2777     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2778        would behave wrongly. This has been present in PCRE since release 4.0.
2779    
2780    10. A number of changes to the pcregrep command:
2781    
2782        (a) Refactored how -x works; insert ^(...)$ instead of setting
2783            PCRE_ANCHORED and checking the length, in preparation for adding
2784            something similar for -w.
2785    
2786        (b) Added the -w (match as a word) option.
2787    
2788        (c) Refactored the way lines are read and buffered so as to have more
2789            than one at a time available.
2790    
2791        (d) Implemented a pcregrep test script.
2792    
2793        (e) Added the -M (multiline match) option. This allows patterns to match
2794            over several lines of the subject. The buffering ensures that at least
2795            8K, or the rest of the document (whichever is the shorter) is available
2796            for matching (and similarly the previous 8K for lookbehind assertions).
2797    
2798        (f) Changed the --help output so that it now says
2799    
2800              -w, --word-regex(p)
2801    
2802            instead of two lines, one with "regex" and the other with "regexp"
2803            because that confused at least one person since the short forms are the
2804            same. (This required a bit of code, as the output is generated
2805            automatically from a table. It wasn't just a text change.)
2806    
2807        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2808            option but starts with a hyphen. Could be a pattern or a path name
2809            starting with a hyphen, for instance.
2810    
2811        (h) "-" can be given as a file name to represent stdin.
2812    
2813        (i) When file names are being printed, "(standard input)" is used for
2814            the standard input, for compatibility with GNU grep. Previously
2815            "<stdin>" was used.
2816    
2817        (j) The option --label=xxx can be used to supply a name to be used for
2818            stdin when file names are being printed. There is no short form.
2819    
2820        (k) Re-factored the options decoding logic because we are going to add
2821            two more options that take data. Such options can now be given in four
2822            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2823    
2824        (l) Added the -A, -B, and -C options for requesting that lines of context
2825            around matches be printed.
2826    
2827        (m) Added the -L option to print the names of files that do not contain
2828            any matching lines, that is, the complement of -l.
2829    
2830        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2831            continue to scan other files.
2832    
2833        (o) The -s option was incorrectly implemented. For compatibility with other
2834            greps, it now suppresses the error message for a non-existent or non-
2835            accessible file (but not the return code). There is a new option called
2836            -q that suppresses the output of matching lines, which was what -s was
2837            previously doing.
2838    
2839        (p) Added --include and --exclude options to specify files for inclusion
2840            and exclusion when recursing.
2841    
2842    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2843        Hopefully, it now does.
2844    
2845    12. Missing cast in pcre_study().
2846    
2847    13. Added an "uninstall" target to the makefile.
2848    
2849    14. Replaced "extern" in the function prototypes in Makefile.in with
2850        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2851        world, but is set differently for Windows.
2852    
2853    15. Added a second compiling function called pcre_compile2(). The only
2854        difference is that it has an extra argument, which is a pointer to an
2855        integer error code. When there is a compile-time failure, this is set
2856        non-zero, in addition to the error test pointer being set to point to an
2857        error message. The new argument may be NULL if no error number is required
2858        (but then you may as well call pcre_compile(), which is now just a
2859        wrapper). This facility is provided because some applications need a
2860        numeric error indication, but it has also enabled me to tidy up the way
2861        compile-time errors are handled in the POSIX wrapper.
2862    
2863    16. Added VPATH=.libs to the makefile; this should help when building with one
2864        prefix path and installing with another. (Or so I'm told by someone who
2865        knows more about this stuff than I do.)
2866    
2867    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2868        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2869        match everything, including newlines. This is not POSIX-compatible, but
2870        somebody wanted the feature. From pcretest it can be activated by using
2871        both the P and the s flags.
2872    
2873    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2874    
2875    19. libpcre.pc was being incorrectly installed as executable.
2876    
2877    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2878        it now also looks for '\r' so that it will work unmodified on Windows.
2879    
2880    21. Added Google's contributed C++ wrapper to the distribution.
2881    
2882    22. Added some untidy missing memory free() calls in pcretest, to keep
2883        Electric Fence happy when testing.
2884    
2885    
2886    
2887    Version 5.0 13-Sep-04
2888    ---------------------
2889    
2890     1. Internal change: literal characters are no longer packed up into items
2891        containing multiple characters in a single byte-string. Each character
2892        is now matched using a separate opcode. However, there may be more than one
2893        byte in the character in UTF-8 mode.
2894    
2895     2. The pcre_callout_block structure has two new fields: pattern_position and
2896        next_item_length. These contain the offset in the pattern to the next match
2897        item, and its length, respectively.
2898    
2899     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2900        insertion of callouts before each pattern item. Added the /C option to
2901        pcretest to make use of this.
2902    
2903     4. On the advice of a Windows user, the lines
2904    
2905          #if defined(_WIN32) || defined(WIN32)
2906          _setmode( _fileno( stdout ), 0x8000 );
2907          #endif  /* defined(_WIN32) || defined(WIN32) */
2908    
2909        have been added to the source of pcretest. This apparently does useful
2910        magic in relation to line terminators.
2911    
2912     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2913        for the benefit of those environments where the "b" makes a difference.
2914    
2915     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2916        to know about it. I have put a hack into configure.in that adds in code
2917        to set GCC=yes if CC=icc. This seems to end up at a point in the
2918        generated configure script that is early enough to affect the setting of
2919        compiler options, which is what is needed, but I have no means of testing
2920        whether it really works. (The user who reported this had patched the
2921        generated configure script, which of course I cannot do.)
2922    
2923        LATER: After change 22 below (new libtool files), the configure script
2924        seems to know about icc (and also ecc). Therefore, I have commented out
2925        this hack in configure.in.
2926    
2927     7. Added support for pkg-config (2 patches were sent in).
2928    
2929     8. Negated POSIX character classes that used a combination of internal tables
2930        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2931        [[:^ascii]]. Typically, they would match almost any characters. The other
2932        POSIX classes were not broken in this way.
2933    
2934     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2935        to find the match, as PCRE was deluded into thinking that the match had to
2936        start at the start point or following a newline. The same bug applied to
2937        patterns with negative forward assertions or any backward assertions
2938        preceding ".*" at the start, unless the pattern required a fixed first
2939        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2940    
2941    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2942        starting at the last subject character, bytes beyond the end of the subject
2943        string were read.
2944    
2945    11. Renamed the variable "class" as "classbits" to make life easier for C++
2946        users. (Previously there was a macro definition, but it apparently wasn't
2947        enough.)
2948    
2949    12. Added the new field "tables" to the extra data so that tables can be passed
2950        in at exec time, or the internal tables can be re-selected. This allows
2951        a compiled regex to be saved and re-used at a later time by a different
2952        program that might have everything at different addresses.
2953    
2954    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2955        -R library as well as a -L library.
2956    
2957    14. The debugging options of pcretest (-d on the command line or D on a
2958        pattern) showed incorrect output for anything following an extended class
2959        that contained multibyte characters and which was followed by a quantifier.
2960    
2961    15. Added optional support for general category Unicode character properties
2962        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2963        support. It adds about 90K to the size of the library. The meanings of the
2964        inbuilt class escapes such as \d and \s have NOT been changed.
2965    
2966    16. Updated pcredemo.c to include calls to free() to release the memory for the
2967        compiled pattern.
2968    
2969    17. The generated file chartables.c was being created in the source directory
2970        instead of in the building directory. This caused the build to fail if the
2971        source directory was different from the building directory, and was
2972        read-only.
2973    
2974    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2975        file. No doubt somebody will tell me if they don't make sense... Also added
2976        Dan Mooney's comments about building on OpenVMS.
2977    
2978    19. Added support for partial matching via the PCRE_PARTIAL option for
2979        pcre_exec() and the \P data escape in pcretest.
2980    
2981    20. Extended pcretest with 3 new pattern features:
2982    
2983        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2984              write the compiled pattern to the file whose name is "rest-of-line".
2985              This is a straight binary dump of the data, with the saved pointer to
2986              the character tables forced to be NULL. The study data, if any, is
2987              written too. After writing, pcretest reads a new pattern.
2988    
2989        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2990              compiled pattern from the given file. There must not be any
2991              occurrences of "<" in the file name (pretty unlikely); if there are,
2992              pcretest will instead treat the initial "<" as a pattern delimiter.
2993              After reading in the pattern, pcretest goes on to read data lines as
2994              usual.
2995    
2996        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2997              and 16-bit fields in a compiled pattern, to simulate a pattern that
2998              was compiled on a host of opposite endianness.
2999    
3000    21. The pcre-exec() function can now cope with patterns that were compiled on
3001        hosts of opposite endianness, with this restriction:
3002    
3003          As for any compiled expression that is saved and used later, the tables
3004          pointer field cannot be preserved; the extra_data field in the arguments
3005          to pcre_exec() should be used to pass in a tables address if a value
3006          other than the default internal tables were used at compile time.
3007    
3008    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3009        now diagnosed as an error. Previously, most of the time, a negative number
3010        would have been treated as zero, but if in addition "ovector" was passed as
3011        NULL, a crash could occur.
3012    
3013    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3014        new versions from the libtool 1.5 distribution (the last one is a copy of
3015        a file called libtool.m4). This seems to have fixed the need to patch
3016        "configure" to support Darwin 1.3 (which I used to do). However, I still
3017        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3018        workstation).
3019    
3020    24. Changed the PCRE licence to be the more standard "BSD" licence.
3021    
3022    
3023    Version 4.5 01-Dec-03
3024    ---------------------
3025    
3026     1. There has been some re-arrangement of the code for the match() function so
3027        that it can be compiled in a version that does not call itself recursively.
3028        Instead, it keeps those local variables that need separate instances for
3029        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3030        needs to "recurse". Keeping track of where control must go is done by means
3031        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3032        hide most of the details from the main code, and operates only if
3033        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3034        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3035        operating.
3036    
3037        To make it easier for callers to provide specially tailored get/free
3038        functions for this usage, two new functions, pcre_stack_malloc, and
3039        pcre_stack_free, are used. They are always called in strict stacking order,
3040        and the size of block requested is always the same.
3041    
3042        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3043        PCRE has been compiled to use the stack or the heap for recursion. The
3044        -C option of pcretest uses this to show which version is compiled.
3045    
3046        A new data escape \S, is added to pcretest; it causes the amounts of store
3047        obtained and freed by both kinds of malloc/free at match time to be added
3048        to the output.
3049    
3050     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3051        what's available on my current Linux desktop machine.
3052    
3053     3. When matching a UTF-8 string, the test for a valid string at the start has
3054        been extended. If start_offset is not zero, PCRE now checks that it points
3055        to a byte that is the start of a UTF-8 character. If not, it returns
3056        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3057        this is necessary because there may be backward assertions in the pattern.
3058        When matching the same subject several times, it may save resources to use
3059        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3060    
3061     4. The code for checking the validity of UTF-8 strings has been tightened so
3062        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3063        containing "overlong sequences".
3064    
3065     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3066        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3067        should have been "&", but it just so happened that all the cases this let
3068        through by mistake were picked up later in the function.
3069    
3070     6. I had used a variable called "isblank" - this is a C99 function, causing
3071        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3072    
3073     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3074        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3075        so the version is shown; (c) stop "make test" echoing "./RunTest".
3076    
3077     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3078    
3079     9. The prototype for memmove() for systems that don't have it was using
3080        size_t, but the inclusion of the header that defines size_t was later. I've
3081        moved the #includes for the C headers earlier to avoid this.
3082    
3083    10. Added some adjustments to the code to make it easier to compiler on certain
3084        special systems:
3085    
3086          (a) Some "const" qualifiers were missing.
3087          (b) Added the macro EXPORT before all exported functions; by default this
3088              is defined to be empty.
3089          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3090              that it reads its output file name as an argument instead of writing
3091              to the standard output and assuming this can be redirected.
3092    
3093    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3094        class containing characters with values greater than 255, PCRE compilation
3095        went into a loop.
3096    
3097    12. A recursive reference to a subpattern that was within another subpattern
3098        that had a minimum quantifier of zero caused PCRE to crash. For example,
3099        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3100        recursion. If the recursively-called subpattern itself had a zero repeat,
3101        that was OK.
3102    
3103    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3104        buffer into which it was copied (for escape processing) was still set at
3105        1024, so long lines caused crashes.
3106    
3107    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3108        "internal error: code overflow...". This applied to any character class
3109        that was followed by a possessive quantifier.
3110    
3111    15. Modified the Makefile to add libpcre.la as a prerequisite for
3112        libpcreposix.la because I was told this is needed for a parallel build to
3113        work.
3114    
3115    16. If a pattern that contained .* following optional items at the start was
3116        studied, the wrong optimizing data was generated, leading to matching
3117        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3118        matching string must start with a or b or c. The correct conclusion for
3119        this pattern is that a match can start with any character.
3120    
3121    
3122    Version 4.4 13-Aug-03
3123    ---------------------
3124    
3125     1. In UTF-8 mode, a character class containing characters with values between
3126        127 and 255 was not handled correctly if the compiled pattern was studied.
3127        In fixing this, I have also improved the studying algorithm for such
3128        classes (slightly).
3129    
3130     2. Three internal functions had redundant arguments passed to them. Removal
3131        might give a very teeny performance improvement.
3132    
3133     3. Documentation bug: the value of the capture_top field in a callout is *one
3134        more than* the number of the hightest numbered captured substring.
3135    
3136     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3137        in incorrectly linking with a previously installed version. They now link
3138        explicitly with libpcre.la.
3139    
3140     5. configure.in no longer needs to recognize Cygwin specially.
3141    
3142     6. A problem in pcre.in for Windows platforms is fixed.
3143    
3144     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3145        pcretest, it used to include the size of the study block as part of its
3146        output. Unfortunately, the structure contains a field that has a different
3147        size on different hardware architectures. This meant that the tests that
3148        showed this size failed. As the block is currently always of a fixed size,
3149        this information isn't actually particularly useful in pcretest output, so
3150        I have just removed it.
3151    
3152     8. Three pre-processor statements accidentally did not start in column 1.
3153        Sadly, there are *still* compilers around that complain, even though
3154        standard C has not required this for well over a decade. Sigh.
3155    
3156     9. In pcretest, the code for checking callouts passed small integers in the
3157        callout_data field, which is a void * field. However, some picky compilers
3158        complained about the casts involved for this on 64-bit systems. Now
3159        pcretest passes the address of the small integer instead, which should get
3160        rid of the warnings.
3161    
3162    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3163        both compile and run time, and gives an error if an invalid UTF-8 sequence
3164        is found. There is a option for disabling this check in cases where the
3165        string is known to be correct and/or the maximum performance is wanted.
3166    
3167    11. In response to a bug report, I changed one line in Makefile.in from
3168    
3169            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3170        to
3171            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3172    
3173        to look similar to other lines, but I have no way of telling whether this
3174        is the right thing to do, as I do not use Windows. No doubt I'll get told
3175        if it's wrong...
3176    
3177    
3178    Version 4.3 21-May-03
3179    ---------------------
3180    
3181    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3182       Makefile.
3183    
3184    2. Some refactoring to improve the quality of the code:
3185    
3186       (i)   The utf8_table... variables are now declared "const".
3187    
3188       (ii)  The code for \cx, which used the "case flipping" table to upper case
3189             lower case letters, now just substracts 32. This is ASCII-specific,
3190             but the whole concept of \cx is ASCII-specific, so it seems
3191             reasonable.
3192    
3193       (iii) PCRE was using its character types table to recognize decimal and
3194             hexadecimal digits in the pattern. This is silly, because it handles
3195             only 0-9, a-f, and A-F, but the character types table is locale-
3196             specific, which means strange things might happen. A private
3197             table is now used for this - though it costs 256 bytes, a table is
3198             much faster than multiple explicit tests. Of course, the standard
3199             character types table is still used for matching digits in subject
3200             strings against \d.
3201    
3202       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3203             ending in _t are). So I've renamed it as ESC_tee.
3204    
3205    3. The first argument for regexec() in the POSIX wrapper should have been
3206       defined as "const".
3207    
3208    4. Changed pcretest to use malloc() for its buffers so that they can be
3209       Electric Fenced for debugging.
3210    
3211    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3212       to read one or more bytes before the start of the subject string. Often this
3213       had no effect on PCRE's behaviour, but in some circumstances it could
3214       provoke a segmentation fault.
3215    
3216    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3217       to try to read one or more bytes before the start of the subject string.
3218    
3219    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3220       UTF-8 support could misbehave in various ways if the subject string
3221       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3222       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3223       back over UTF-8 characters.)
3224    
3225    
3226    Version 4.2 14-Apr-03
3227    ---------------------
3228    
3229    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3230    
3231    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3232         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3233         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3234         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3235         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3236           and BUILD_EXEEXT
3237         Note: automatic setting of the BUILD variables is not yet working
3238         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3239           compile-time but not at link-time
3240         [LINK]: use for linking executables only
3241         make different versions for Windows and non-Windows
3242         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3243           libraries
3244         [LINK_FOR_BUILD]: new variable
3245         [OBJEXT]: use throughout
3246         [EXEEXT]: use throughout
3247         <winshared>: new target
3248         <wininstall>: new target
3249         <dftables.o>: use native compiler
3250         <dftables>: use native linker
3251         <install>: handle Windows platform correctly
3252         <clean>: ditto
3253         <check>: ditto
3254         copy DLL to top builddir before testing
3255    
3256       As part of these changes, -no-undefined was removed again. This was reported
3257       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3258       in any case.
3259    
3260    3. Some tidies to get rid of compiler warnings:
3261    
3262       . In the match_data structure, match_limit was an unsigned long int, whereas
3263         match_call_count was an int. I've made them both unsigned long ints.
3264    
3265       . In pcretest the fact that a const uschar * doesn't automatically cast to
3266         a void * provoked a warning.
3267    
3268       . Turning on some more compiler warnings threw up some "shadow" variables
3269         and a few more missing casts.
3270    
3271    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3272       option, a class that contained a single character with a value between 128
3273       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3274    
3275    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3276       option, a class that contained several characters, but with at least one
3277       whose value was between 128 and 255 caused PCRE to crash.
3278    
3279    
3280    Version 4.1 12-Mar-03
3281    ---------------------
3282    
3283    1. Compiling with gcc -pedantic found a couple of places where casts were
3284    needed, and a string in dftables.c that was longer than standard compilers are
3285    required to support.
3286    
3287    2. Compiling with Sun's compiler found a few more places where the code could
3288    be tidied up in order to avoid warnings.
3289    
3290    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3291    first of these names is deprecated in the latest Autoconf in favour of the name
3292    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3293    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3294    analogy I have changed it to CFLAGS_FOR_BUILD.
3295    
3296    4. Added -no-undefined to the linking command in the Makefile, because this is
3297    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3298    linking step for the pcreposix library.
3299