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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 771 by ph10, Tue Nov 29 15:34:12 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38        being incorrectly compiled and would have given unpredicatble results.
39    
40    9.  A possessively repeated subpattern with minimum repeat count greater than
41        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43        the first (A) could occur when it should not.
44    
45    10. Add a cast and remove a redundant test from the code.
46    
47    11. JIT should use pcre_malloc/pcre_free for allocation.
48    
49    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51        is anything other than /usr, -L is still shown).
52    
53    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54    
55    14. Perl does not support \N without a following name in a [] class; PCRE now
56        also gives an error.
57    
58    15. If a forward reference was repeated with an upper limit of around 2000,
59        it caused the error "internal error: overran compiling workspace". This
60        is now checked, and causes "too many forward references" instead.
61    
62    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
63        incorrectly expecting the subject to contain another "a" after the start.
64    
65    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
66        in the match, the SKIP should be ignored. This was not happening; instead
67        the SKIP was being treated as NOMATCH. For patterns such as
68        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
69        tested.
70    
71    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
72        now much more compatible with Perl, in particular in cases where the result
73        is a non-match for a non-anchored pattern. For example, if
74        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
75        "m", where previously it did not return a name. A side effect of this
76        change is that for partial matches, the last encountered mark name is
77        returned, as for non matches. A number of tests that were previously not
78        Perl-compatible have been moved into the Perl-compatible test files. The
79        refactoring has had the pleasing side effect of removing one argument from
80        the match() function, thus reducing its stack requirements.
81    
82    
83    Version 8.20 21-Oct-2011
84    ------------------------
85    
86    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
87        a POSIX class. After further experiments with Perl, which convinced me that
88        Perl has bugs and confusions, a closing square bracket is no longer allowed
89        in a POSIX name. This bug also affected patterns with classes that started
90        with full stops.
91    
92    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
93        captured substring, but while checking the failing first alternative,
94        substring 1 is temporarily captured. If the output vector supplied to
95        pcre_exec() was not big enough for this capture, the yield of the function
96        was still zero ("insufficient space for captured substrings"). This cannot
97        be totally fixed without adding another stack variable, which seems a lot
98        of expense for a edge case. However, I have improved the situation in cases
99        such as /(a)(b)x|abc/ matched against "abc", where the return code
100        indicates that fewer than the maximum number of slots in the ovector have
101        been set.
102    
103    3.  Related to (2) above: when there are more back references in a pattern than
104        slots in the output vector, pcre_exec() uses temporary memory during
105        matching, and copies in the captures as far as possible afterwards. It was
106        using the entire output vector, but this conflicts with the specification
107        that only 2/3 is used for passing back captured substrings. Now it uses
108        only the first 2/3, for compatibility. This is, of course, another edge
109        case.
110    
111    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
112        main code base, and can be used by building with --enable-jit. When this is
113        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
114        runtime --no-jit option is given.
115    
116    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
117        ovector, the return from the function was zero, implying that there were
118        other matches that did not fit. The correct "exactly full" value is now
119        returned.
120    
121    6.  If a subpattern that was called recursively or as a subroutine contained
122        (*PRUNE) or any other control that caused it to give a non-standard return,
123        invalid errors such as "Error -26 (nested recursion at the same subject
124        position)" or even infinite loops could occur.
125    
126    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
127        computing the minimum length on reaching *ACCEPT, and so ended up with the
128        wrong value of 1 rather than 0. Further investigation indicates that
129        computing a minimum subject length in the presence of *ACCEPT is difficult
130        (think back references, subroutine calls), and so I have changed the code
131        so that no minimum is registered for a pattern that contains *ACCEPT.
132    
133    8.  If (*THEN) was present in the first (true) branch of a conditional group,
134        it was not handled as intended. [But see 16 below.]
135    
136    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
137        Sheri Pierce.
138    
139    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
140        the first byte in a match must be "a".
141    
142    11. Change 17 for 8.13 increased the recursion depth for patterns like
143        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
144        pattern contains any instances of (*THEN). If it does not, the old
145        optimizations are restored. It would be nice to do this on a per-group
146        basis, but at the moment that is not feasible.
147    
148    12. In some environments, the output of pcretest -C is CRLF terminated. This
149        broke RunTest's code that checks for the link size. A single white space
150        character after the value is now allowed for.
151    
152    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
153        For "fr", it uses the Windows-specific input and output files.
154    
155    14. If (*THEN) appeared in a group that was called recursively or as a
156        subroutine, it did not work as intended. [But see next item.]
157    
158    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
159        pattern fragments (but not containing any | characters). If A and B are
160        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
161        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
162        D. In other words, Perl considers parentheses that do not contain any |
163        characters to be part of a surrounding alternative, whereas PCRE was
164        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
165        differently. PCRE now behaves in the same way as Perl, except in the case
166        of subroutine/recursion calls such as (?1) which have in any case always
167        been different (but PCRE had them first :-).
168    
169    16. Related to 15 above: Perl does not treat the | in a conditional group as
170        creating alternatives. Such a group is treated in the same way as an
171        ordinary group without any | characters when processing (*THEN). PCRE has
172        been changed to match Perl's behaviour.
173    
174    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
175        RunGrepTest script failed.
176    
177    18. Change 22 for version 13 caused atomic groups to use more stack. This is
178        inevitable for groups that contain captures, but it can lead to a lot of
179        stack use in large patterns. The old behaviour has been restored for atomic
180        groups that do not contain any capturing parentheses.
181    
182    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
183        suppress the check for a minimum subject length at run time. (If it was
184        given to pcre_exec() or pcre_dfa_exec() it did work.)
185    
186    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
187        fail to work when decoding hex characters in data strings in EBCDIC
188        environments.
189    
190    21. It appears that in at least one Mac OS environment, the isxdigit() function
191        is implemented as a macro that evaluates to its argument more than once,
192        contravening the C 90 Standard (I haven't checked a later standard). There
193        was an instance in pcretest which caused it to go wrong when processing
194        \x{...} escapes in subject strings. The has been rewritten to avoid using
195        things like p++ in the argument of isxdigit().
196    
197    
198    Version 8.13 16-Aug-2011
199    ------------------------
200    
201    1.  The Unicode data tables have been updated to Unicode 6.0.0.
202    
203    2.  Two minor typos in pcre_internal.h have been fixed.
204    
205    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
206        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
207        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
208    
209    4.  There were a number of related bugs in the code for matching backrefences
210        caselessly in UTF-8 mode when codes for the characters concerned were
211        different numbers of bytes. For example, U+023A and U+2C65 are an upper
212        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
213        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
214        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
215        2-byte code at the end of the subject (it thought there wasn't enough data
216        left).
217    
218    5.  Comprehensive information about what went wrong is now returned by
219        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
220        as the output vector has at least 2 elements. The offset of the start of
221        the failing character and a reason code are placed in the vector.
222    
223    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
224        now returned is for the first byte of the failing character, instead of the
225        last byte inspected. This is an incompatible change, but I hope it is small
226        enough not to be a problem. It makes the returned offset consistent with
227        pcre_exec() and pcre_dfa_exec().
228    
229    7.  pcretest now gives a text phrase as well as the error number when
230        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
231        failure, the offset and reason code are output.
232    
233    8.  When \R was used with a maximizing quantifier it failed to skip backwards
234        over a \r\n pair if the subsequent match failed. Instead, it just skipped
235        back over a single character (\n). This seems wrong (because it treated the
236        two characters as a single entity when going forwards), conflicts with the
237        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
238        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
239        has been changed.
240    
241    9.  Some internal refactoring has changed the processing so that the handling
242        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
243        time (the PCRE_DOTALL option was changed this way some time ago: version
244        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
245        which was always a bit of a fudge. It also means that there is one less
246        argument for the match() function, which reduces its stack requirements
247        slightly. This change also fixes an incompatibility with Perl: the pattern
248        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
249    
250    10. More internal refactoring has drastically reduced the number of recursive
251        calls to match() for possessively repeated groups such as (abc)++ when
252        using pcre_exec().
253    
254    11. While implementing 10, a number of bugs in the handling of groups were
255        discovered and fixed:
256    
257        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
258        (a|)*(?1) gave a compile-time internal error.
259        ((a|)+)+  did not notice that the outer group could match an empty string.
260        (^a|^)+   was not marked as anchored.
261        (.*a|.*)+ was not marked as matching at start or after a newline.
262    
263    12. Yet more internal refactoring has removed another argument from the match()
264        function. Special calls to this function are now indicated by setting a
265        value in a variable in the "match data" data block.
266    
267    13. Be more explicit in pcre_study() instead of relying on "default" for
268        opcodes that mean there is no starting character; this means that when new
269        ones are added and accidentally left out of pcre_study(), testing should
270        pick them up.
271    
272    14. The -s option of pcretest has been documented for ages as being an old
273        synonym of -m (show memory usage). I have changed it to mean "force study
274        for every regex", that is, assume /S for every regex. This is similar to -i
275        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
276        using it. It makes it easier to run collections of tests with and without
277        study enabled, and thereby test pcre_study() more easily. All the standard
278        tests are now run with and without -s (but some patterns can be marked as
279        "never study" - see 20 below).
280    
281    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
282        restoration of the capturing data to the outer values was not happening
283        correctly.
284    
285    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
286        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
287        pattern had matched an empty string, and so incorrectly returned a no
288        match.
289    
290    17. There was optimizing code for the last branch of non-capturing parentheses,
291        and also for the obeyed branch of a conditional subexpression, which used
292        tail recursion to cut down on stack usage. Unfortunately, now that there is
293        the possibility of (*THEN) occurring in these branches, tail recursion is
294        no longer possible because the return has to be checked for (*THEN). These
295        two optimizations have therefore been removed. [But see 8.20/11 above.]
296    
297    18. If a pattern containing \R was studied, it was assumed that \R always
298        matched two bytes, thus causing the minimum subject length to be
299        incorrectly computed because \R can also match just one byte.
300    
301    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
302        was incorrectly computed.
303    
304    20. If /S is present twice on a test pattern in pcretest input, it now
305        *disables* studying, thereby overriding the use of -s on the command line
306        (see 14 above). This is necessary for one or two tests to keep the output
307        identical in both cases.
308    
309    21. When (*ACCEPT) was used in an assertion that matched an empty string and
310        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
311    
312    22. When an atomic group that contained a capturing parenthesis was
313        successfully matched, but the branch in which it appeared failed, the
314        capturing was not being forgotten if a higher numbered group was later
315        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
316        group 1 to "a", when in fact it should be unset. This applied to multi-
317        branched capturing and non-capturing groups, repeated or not, and also to
318        positive assertions (capturing in negative assertions does not happen
319        in PCRE) and also to nested atomic groups.
320    
321    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
322        subject after a captured substring, to make it easier to tell which of a
323        number of identical substrings has been captured.
324    
325    24. The way atomic groups are processed by pcre_exec() has been changed so that
326        if they are repeated, backtracking one repetition now resets captured
327        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
328        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
329        "aaa". Previously, it would have been "a". As part of this code
330        refactoring, the way recursive calls are handled has also been changed.
331    
332    25. If an assertion condition captured any substrings, they were not passed
333        back unless some other capturing happened later. For example, if
334        (?(?=(a))a) was matched against "a", no capturing was returned.
335    
336    26. When studying a pattern that contained subroutine calls or assertions,
337        the code for finding the minimum length of a possible match was handling
338        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
339        group 1 called group 2 while simultaneously a separate group 2 called group
340        1). A stack overflow occurred in this case. I have fixed this by limiting
341        the recursion depth to 10.
342    
343    27. Updated RunTest.bat in the distribution to the version supplied by Tom
344        Fortmann. This supports explicit test numbers on the command line, and has
345        argument validation and error reporting.
346    
347    28. An instance of \X with an unlimited repeat could fail if at any point the
348        first character it looked at was a mark character.
349    
350    29. Some minor code refactoring concerning Unicode properties and scripts
351        should reduce the stack requirement of match() slightly.
352    
353    30. Added the '=' option to pcretest to check the setting of unused capturing
354        slots at the end of the pattern, which are documented as being -1, but are
355        not included in the return count.
356    
357    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
358        compiled something random. Now it gives a compile-time error (as does
359        Perl).
360    
361    32. A *MARK encountered during the processing of a positive assertion is now
362        recorded and passed back (compatible with Perl).
363    
364    33. If --only-matching or --colour was set on a pcregrep call whose pattern
365        had alternative anchored branches, the search for a second match in a line
366        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
367        matched the line "0102" twice. The same bug affected patterns that started
368        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
369        twice.
370    
371    34. Previously, PCRE did not allow quantification of assertions. However, Perl
372        does, and because of capturing effects, quantifying parenthesized
373        assertions may at times be useful. Quantifiers are now allowed for
374        parenthesized assertions.
375    
376    35. A minor code tidy in pcre_compile() when checking options for \R usage.
377    
378    36. \g was being checked for fancy things in a character class, when it should
379        just be a literal "g".
380    
381    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
382        appearance of a nested POSIX class supersedes an apparent external class.
383        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
384        unescaped square brackets may also appear as part of class names. For
385        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
386        more like Perl. (But see 8.20/1 above.)
387    
388    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
389        was because it thought it was \N{name}, which is not supported).
390    
391    39. Add minix to OS list not supporting the -S option in pcretest.
392    
393    40. PCRE tries to detect cases of infinite recursion at compile time, but it
394        cannot analyze patterns in sufficient detail to catch mutual recursions
395        such as ((?1))((?2)). There is now a runtime test that gives an error if a
396        subgroup is called recursively as a subpattern for a second time at the
397        same position in the subject string. In previous releases this might have
398        been caught by the recursion limit, or it might have run out of stack.
399    
400    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
401        happen only once. PCRE was, however incorrectly giving a compile time error
402        "recursive call could loop indefinitely" because it cannot analyze the
403        pattern in sufficient detail. The compile time test no longer happens when
404        PCRE is compiling a conditional subpattern, but actual runaway loops are
405        now caught at runtime (see 40 above).
406    
407    42. It seems that Perl allows any characters other than a closing parenthesis
408        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
409        has been changed to be the same.
410    
411    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
412        as not to get warnings when autogen.sh is called. Also changed
413        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
414    
415    44. To help people who use pcregrep to scan files containing exceedingly long
416        lines, the following changes have been made:
417    
418        (a) The default value of the buffer size parameter has been increased from
419            8K to 20K. (The actual buffer used is three times this size.)
420    
421        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
422            PCRE is built.
423    
424        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
425            to be set at run time.
426    
427        (d) Numerical values in pcregrep options can be followed by K or M, for
428            example --buffer-size=50K.
429    
430        (e) If a line being scanned overflows pcregrep's buffer, an error is now
431            given and the return code is set to 2.
432    
433    45. Add a pointer to the latest mark to the callout data block.
434    
435    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
436        partial match of an empty string instead of no match. This was specific to
437        the use of ".".
438    
439    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
440        complete match instead of a partial match. This bug was dependent on both
441        the PCRE_UTF8 and PCRE_DOTALL options being set.
442    
443    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
444        starting byte set, because \b was not being ignored.
445    
446    
447    Version 8.12 15-Jan-2011
448    ------------------------
449    
450    1.  Fixed some typos in the markup of the man pages, and wrote a script that
451        checks for such things as part of the documentation building process.
452    
453    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
454        --match-limit and --recursion-limit options (added for 8.11). In
455        particular, this made one of the standard tests fail. (The integer value
456        went into the wrong half of a long int.)
457    
458    3.  If the --colour option was given to pcregrep with -v (invert match), it
459        did strange things, either producing crazy output, or crashing. It should,
460        of course, ignore a request for colour when reporting lines that do not
461        match.
462    
463    4.  Another pcregrep bug caused similar problems if --colour was specified with
464        -M (multiline) and the pattern match finished with a line ending.
465    
466    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
467        matched in multiline mode, the following line was shown as part of the
468        match. This seems wrong, so I have changed it.
469    
470    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
471        the check for further matches in the same line (so they could be coloured)
472        to overrun the end of the current line. If another match was found, it was
473        incorrectly shown (and then shown again when found in the next line).
474    
475    7.  If pcregrep was compiled under Windows, there was a reference to the
476        function pcregrep_exit() before it was defined. I am assuming this was
477        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
478        reported by a user. I've moved the definition above the reference.
479    
480    
481    Version 8.11 10-Dec-2010
482    ------------------------
483    
484    1.  (*THEN) was not working properly if there were untried alternatives prior
485        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
486        backtracked to try for "b" instead of moving to the next alternative branch
487        at the same level (in this case, to look for "c"). The Perl documentation
488        is clear that when (*THEN) is backtracked onto, it goes to the "next
489        alternative in the innermost enclosing group".
490    
491    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
492        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
493        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
494        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
495        (*THEN).
496    
497    3.  If \s appeared in a character class, it removed the VT character from
498        the class, even if it had been included by some previous item, for example
499        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
500        of \s, but is part of the POSIX "space" class.)
501    
502    4.  A partial match never returns an empty string (because you can always
503        match an empty string at the end of the subject); however the checking for
504        an empty string was starting at the "start of match" point. This has been
505        changed to the "earliest inspected character" point, because the returned
506        data for a partial match starts at this character. This means that, for
507        example, /(?<=abc)def/ gives a partial match for the subject "abc"
508        (previously it gave "no match").
509    
510    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
511        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
512        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
513        has an implication that the given string is incomplete (because a partial
514        match is preferred over a full match). For this reason, these items now
515        give a partial match in this situation. [Aside: previously, the one case
516        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
517        match rather than a full match, which was wrong by the old rules, but is
518        now correct.]
519    
520    6.  There was a bug in the handling of #-introduced comments, recognized when
521        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
522        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
523        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
524        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
525        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
526        places in pcre_compile().
527    
528    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
529        comments when looking ahead for named forward references to subpatterns,
530        the only newline sequence it recognized was NL. It now handles newlines
531        according to the set newline convention.
532    
533    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
534        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
535        cater for a lack of strerror(). These oversights have been fixed.
536    
537    9.  Added --match-limit and --recursion-limit to pcregrep.
538    
539    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
540    
541    11. When the -o option was used, pcregrep was setting a return code of 1, even
542        when matches were found, and --line-buffered was not being honoured.
543    
544    12. Added an optional parentheses number to the -o and --only-matching options
545        of pcregrep.
546    
547    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
548        can match an empty string. The code to do it in pcretest and pcredemo
549        needed fixing:
550    
551        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
552            only one byte after an empty string match just before CRLF (this case
553            just got forgotten; "any" and "anycrlf" were OK).
554    
555        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
556            mode when an empty string match preceded an ASCII character followed by
557            a non-ASCII character. (The code for advancing by one character rather
558            than one byte was nonsense.)
559    
560        (c) The pcredemo.c sample program did not have any code at all to handle
561            the cases when CRLF is a valid newline sequence.
562    
563    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
564        as a starting offset was within the subject string. There is now a new
565        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
566        negative or greater than the length of the string. In order to test this,
567        pcretest is extended to allow the setting of negative starting offsets.
568    
569    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
570        starting offset points to the beginning of a UTF-8 character was
571        unnecessarily clumsy. I tidied it up.
572    
573    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
574        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
575    
576    17. Nobody had reported that the --include_dir option, which was added in
577        release 7.7 should have been called --include-dir (hyphen, not underscore)
578        for compatibility with GNU grep. I have changed it to --include-dir, but
579        left --include_dir as an undocumented synonym, and the same for
580        --exclude-dir, though that is not available in GNU grep, at least as of
581        release 2.5.4.
582    
583    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
584        characters from a string of bytes) have been redefined so as not to use
585        loops, in order to improve performance in some environments. At the same
586        time, I abstracted some of the common code into auxiliary macros to save
587        repetition (this should not affect the compiled code).
588    
589    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
590        compile-time error is now given if \c is not followed by an ASCII
591        character, that is, a byte less than 128. (In EBCDIC mode, the code is
592        different, and any byte value is allowed.)
593    
594    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
595        START_OPTIMIZE option, which is now allowed at compile time - but just
596        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
597        to pcregrep and other applications that have no direct access to PCRE
598        options. The new /Y option in pcretest sets this option when calling
599        pcre_compile().
600    
601    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
602        back references. Groups containing recursive back references were forced to
603        be atomic by that change, but in the case of named groups, the amount of
604        memory required was incorrectly computed, leading to "Failed: internal
605        error: code overflow". This has been fixed.
606    
607    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
608        pcretest.c, to avoid build problems in some Borland environments.
609    
610    
611    Version 8.10 25-Jun-2010
612    ------------------------
613    
614    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
615        THEN.
616    
617    2.  (*ACCEPT) was not working when inside an atomic group.
618    
619    3.  Inside a character class, \B is treated as a literal by default, but
620        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
621        causes the error). The code is unchanged, but I tidied the documentation.
622    
623    4.  Inside a character class, PCRE always treated \R and \X as literals,
624        whereas Perl faults them if its -w option is set. I have changed PCRE so
625        that it faults them when PCRE_EXTRA is set.
626    
627    5.  Added support for \N, which always matches any character other than
628        newline. (It is the same as "." when PCRE_DOTALL is not set.)
629    
630    6.  When compiling pcregrep with newer versions of gcc which may have
631        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
632        declared with attribute warn_unused_result" were given. Just casting the
633        result to (void) does not stop the warnings; a more elaborate fudge is
634        needed. I've used a macro to implement this.
635    
636    7.  Minor change to pcretest.c to avoid a compiler warning.
637    
638    8.  Added four artifical Unicode properties to help with an option to make
639        \s etc use properties (see next item). The new properties are: Xan
640        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
641    
642    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
643        use Unicode properties. (*UCP) at the start of a pattern can be used to set
644        this option. Modified pcretest to add /W to test this facility. Added
645        REG_UCP to make it available via the POSIX interface.
646    
647    10. Added --line-buffered to pcregrep.
648    
649    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
650        studied, and the match started with a letter with a code point greater than
651        127 whose first byte was different to the first byte of the other case of
652        the letter, the other case of this starting letter was not recognized
653        (#976).
654    
655    12. If a pattern that was studied started with a repeated Unicode property
656        test, for example, \p{Nd}+, there was the theoretical possibility of
657        setting up an incorrect bitmap of starting bytes, but fortunately it could
658        not have actually happened in practice until change 8 above was made (it
659        added property types that matched character-matching opcodes).
660    
661    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
662        possible starting bytes for non-anchored patterns.
663    
664    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
665        \R, and also a number of cases that involve Unicode properties, both
666        explicit and implicit when PCRE_UCP is set.
667    
668    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
669        input, it could crash or give wrong results if characters with values
670        greater than 0xc0 were present in the subject string. (Detail: it assumed
671        UTF-8 input when processing these items.)
672    
673    16. Added a lot of (int) casts to avoid compiler warnings in systems where
674        size_t is 64-bit (#991).
675    
676    17. Added a check for running out of memory when PCRE is compiled with
677        --disable-stack-for-recursion (#990).
678    
679    18. If the last data line in a file for pcretest does not have a newline on
680        the end, a newline was missing in the output.
681    
682    19. The default pcre_chartables.c file recognizes only ASCII characters (values
683        less than 128) in its various bitmaps. However, there is a facility for
684        generating tables according to the current locale when PCRE is compiled. It
685        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
686        characters, are recognized by isspace() and therefore were getting set in
687        these tables, and indeed these tables seem to approximate to ISO 8859. This
688        caused a problem in UTF-8 mode when pcre_study() was used to create a list
689        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
690        which of course cannot start UTF-8 characters. I have changed the code so
691        that only real ASCII characters (less than 128) and the correct starting
692        bytes for UTF-8 encodings are set for characters greater than 127 when in
693        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
694        altogether.)
695    
696    20. Added the /T option to pcretest so as to be able to run tests with non-
697        standard character tables, thus making it possible to include the tests
698        used for 19 above in the standard set of tests.
699    
700    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
701        reference to a subpattern the other side of a comment that contains an
702        opening parenthesis caused either an internal compiling error, or a
703        reference to the wrong subpattern.
704    
705    
706    Version 8.02 19-Mar-2010
707    ------------------------
708    
709    1.  The Unicode data tables have been updated to Unicode 5.2.0.
710    
711    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
712        configured.
713    
714    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
715        original author of that file, following a query about its status.
716    
717    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
718        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
719    
720    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
721        quantifier applied to a forward-referencing subroutine call, could compile
722        incorrect code or give the error "internal error: previously-checked
723        referenced subpattern not found".
724    
725    6.  Both MS Visual Studio and Symbian OS have problems with initializing
726        variables to point to external functions. For these systems, therefore,
727        pcre_malloc etc. are now initialized to local functions that call the
728        relevant global functions.
729    
730    7.  There were two entries missing in the vectors called coptable and poptable
731        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
732        I've fixed the data, and added a kludgy way of testing at compile time that
733        the lengths are correct (equal to the number of opcodes).
734    
735    8.  Following on from 7, I added a similar kludge to check the length of the
736        eint vector in pcreposix.c.
737    
738    9.  Error texts for pcre_compile() are held as one long string to avoid too
739        much relocation at load time. To find a text, the string is searched,
740        counting zeros. There was no check for running off the end of the string,
741        which could happen if a new error number was added without updating the
742        string.
743    
744    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
745    
746    11. \K was not working if it appeared in an atomic group or in a group that
747        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
748        \K is "not well defined" if used in an assertion. PCRE now accepts it if
749        the assertion is positive, but not if it is negative.
750    
751    12. Change 11 fortuitously reduced the size of the stack frame used in the
752        "match()" function of pcre_exec.c by one pointer. Forthcoming
753        implementation of support for (*MARK) will need an extra pointer on the
754        stack; I have reserved it now, so that the stack frame size does not
755        decrease.
756    
757    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
758        item in branch that calls a recursion is a subroutine call - as in the
759        second branch in the above example - was incorrectly given the compile-
760        time error "recursive call could loop indefinitely" because pcre_compile()
761        was not correctly checking the subroutine for matching a non-empty string.
762    
763    14. The checks for overrunning compiling workspace could trigger after an
764        overrun had occurred. This is a "should never occur" error, but it can be
765        triggered by pathological patterns such as hundreds of nested parentheses.
766        The checks now trigger 100 bytes before the end of the workspace.
767    
768    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
769    
770    
771    Version 8.01 19-Jan-2010
772    ------------------------
773    
774    1.  If a pattern contained a conditional subpattern with only one branch (in
775        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
776        computed the wrong minimum data length (which is of course zero for such
777        subpatterns). This could cause incorrect "no match" results.
778    
779    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
780        the pattern is reset in the first branch, pcre_compile() failed with
781        "internal error: code overflow at offset...". This happened only when
782        the reset was to the original external option setting. (An optimization
783        abstracts leading options settings into an external setting, which was the
784        cause of this.)
785    
786    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
787        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
788        assertion pattern did not match (meaning that the assertion was true), it
789        was incorrectly treated as false if the SKIP had been reached during the
790        matching. This also applied to assertions used as conditions.
791    
792    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
793        assertion subpattern, including such a pattern used as a condition,
794        unpredictable results occurred, instead of the error return
795        PCRE_ERROR_DFA_UITEM.
796    
797    5.  The C++ GlobalReplace function was not working like Perl for the special
798        situation when an empty string is matched. It now does the fancy magic
799        stuff that is necessary.
800    
801    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
802        removed. (These were left over from very, very early versions of PCRE.)
803    
804    7.  Some cosmetic changes to the code to make life easier when compiling it
805        as part of something else:
806    
807        (a) Change DEBUG to PCRE_DEBUG.
808    
809        (b) In pcre_compile(), rename the member of the "branch_chain" structure
810            called "current" as "current_branch", to prevent a collision with the
811            Linux macro when compiled as a kernel module.
812    
813        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
814            prevent a collision with the Linux macro when compiled as a kernel
815            module.
816    
817    8.  In pcre_compile() there are some checks for integer overflows that used to
818        cast potentially large values to (double). This has been changed to that
819        when building, a check for int64_t is made, and if it is found, it is used
820        instead, thus avoiding the use of floating point arithmetic. (There is no
821        other use of FP in PCRE.) If int64_t is not found, the fallback is to
822        double.
823    
824    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
825        2005 (difference between two addresses compared to an unsigned value).
826    
827    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
828        custom one, because of the following reported problem in Windows:
829    
830          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
831              under Win32.
832          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
833              therefore missing the function definition.
834          - The compiler thus generates a "C" signature for the test function.
835          - The linker fails to find the "C" function.
836          - PCRE fails to configure if asked to do so against libbz2.
837    
838    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
839        messages were output:
840    
841          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
842          rerunning libtoolize, to keep the correct libtool macros in-tree.
843          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
844    
845        I have done both of these things.
846    
847    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
848        most of the time, it *can* run out if it is given a pattern that contains a
849        runaway infinite recursion. I updated the discussion in the pcrestack man
850        page.
851    
852    13. Now that we have gone to the x.xx style of version numbers, the minor
853        version may start with zero. Using 08 or 09 is a bad idea because users
854        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
855        interpreted as invalid octal numbers. I've updated the previous comment in
856        configure.ac, and also added a check that gives an error if 08 or 09 are
857        used.
858    
859    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
860        causing partial matching to fail when the end of the subject matched \W
861        in a UTF-8 pattern where \W was quantified with a minimum of 3.
862    
863    15. There were some discrepancies between the declarations in pcre_internal.h
864        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
865        their definitions. The declarations used "const uschar *" and the
866        definitions used USPTR. Even though USPTR is normally defined as "const
867        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
868        reported that: "This difference in casting confuses some C++ compilers, for
869        example, SunCC recognizes above declarations as different functions and
870        generates broken code for hbpcre." I have changed the declarations to use
871        USPTR.
872    
873    16. GNU libtool is named differently on some systems. The autogen.sh script now
874        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
875        (FreeBSD).
876    
877    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
878        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
879        comment: "Figure out how to create a longlong from a string: strtoll and
880        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
881        instance, but it only takes 2 args instead of 3!"
882    
883    18. A subtle bug concerned with back references has been fixed by a change of
884        specification, with a corresponding code fix. A pattern such as
885        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
886        refers, was giving matches when it shouldn't. For example, xa=xaaa would
887        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
888        same bug. Such groups have to be quantified to be useful, or contained
889        inside another quantified group. (If there's no repetition, the reference
890        can never match.) The problem arises because, having left the group and
891        moved on to the rest of the pattern, a later failure that backtracks into
892        the group uses the captured value from the final iteration of the group
893        rather than the correct earlier one. I have fixed this in PCRE by forcing
894        any group that contains a reference to itself to be an atomic group; that
895        is, there cannot be any backtracking into it once it has completed. This is
896        similar to recursive and subroutine calls.
897    
898    
899    Version 8.00 19-Oct-09
900    ----------------------
901    
902    1.  The table for translating pcre_compile() error codes into POSIX error codes
903        was out-of-date, and there was no check on the pcre_compile() error code
904        being within the table. This could lead to an OK return being given in
905        error.
906    
907    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
908        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
909        in a Windows environment.
910    
911    3.  The pcregrep --count option prints the count for each file even when it is
912        zero, as does GNU grep. However, pcregrep was also printing all files when
913        --files-with-matches was added. Now, when both options are given, it prints
914        counts only for those files that have at least one match. (GNU grep just
915        prints the file name in this circumstance, but including the count seems
916        more useful - otherwise, why use --count?) Also ensured that the
917        combination -clh just lists non-zero counts, with no names.
918    
919    4.  The long form of the pcregrep -F option was incorrectly implemented as
920        --fixed_strings instead of --fixed-strings. This is an incompatible change,
921        but it seems right to fix it, and I didn't think it was worth preserving
922        the old behaviour.
923    
924    5.  The command line items --regex=pattern and --regexp=pattern were not
925        recognized by pcregrep, which required --regex pattern or --regexp pattern
926        (with a space rather than an '='). The man page documented the '=' forms,
927        which are compatible with GNU grep; these now work.
928    
929    6.  No libpcreposix.pc file was created for pkg-config; there was just
930        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
931    
932    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
933        when UCP support is not needed, by modifying the Python script that
934        generates it from Unicode data files. This should not matter if the module
935        is correctly used as a library, but I received one complaint about 50K of
936        unwanted data. My guess is that the person linked everything into his
937        program rather than using a library. Anyway, it does no harm.
938    
939    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
940        was a minimum greater than 1 for a wide character in a possessive
941        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
942        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
943        character. Chaos in the form of incorrect output or a compiling loop could
944        result.
945    
946    9.  The restrictions on what a pattern can contain when partial matching is
947        requested for pcre_exec() have been removed. All patterns can now be
948        partially matched by this function. In addition, if there are at least two
949        slots in the offset vector, the offset of the earliest inspected character
950        for the match and the offset of the end of the subject are set in them when
951        PCRE_ERROR_PARTIAL is returned.
952    
953    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
954        synonymous with PCRE_PARTIAL, for backwards compatibility, and
955        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
956        and may be more useful for multi-segment matching.
957    
958    11. Partial matching with pcre_exec() is now more intuitive. A partial match
959        used to be given if ever the end of the subject was reached; now it is
960        given only if matching could not proceed because another character was
961        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
962        string "Z", which now yields "no match" instead of "partial match". In the
963        case of pcre_dfa_exec(), "no match" is given if every matching path for the
964        final character ended with (*FAIL).
965    
966    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
967        if the pattern had a "must contain" character that was already found in the
968        earlier partial match, unless partial matching was again requested. For
969        example, with the pattern /dog.(body)?/, the "must contain" character is
970        "g". If the first part-match was for the string "dog", restarting with
971        "sbody" failed. This bug has been fixed.
972    
973    13. The string returned by pcre_dfa_exec() after a partial match has been
974        changed so that it starts at the first inspected character rather than the
975        first character of the match. This makes a difference only if the pattern
976        starts with a lookbehind assertion or \b or \B (\K is not supported by
977        pcre_dfa_exec()). It's an incompatible change, but it makes the two
978        matching functions compatible, and I think it's the right thing to do.
979    
980    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
981        so that the demonstration program is easily available in environments where
982        PCRE has not been installed from source.
983    
984    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
985        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
986        library.
987    
988    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
989        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
990        is not the first non-POSIX option to be added. Clearly some people find
991        these options useful.
992    
993    17. If a caller to the POSIX matching function regexec() passes a non-zero
994        value for nmatch with a NULL value for pmatch, the value of
995        nmatch is forced to zero.
996    
997    18. RunGrepTest did not have a test for the availability of the -u option of
998        the diff command, as RunTest does. It now checks in the same way as
999        RunTest, and also checks for the -b option.
1000    
1001    19. If an odd number of negated classes containing just a single character
1002        interposed, within parentheses, between a forward reference to a named
1003        subpattern and the definition of the subpattern, compilation crashed with
1004        an internal error, complaining that it could not find the referenced
1005        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1006        [The bug was that it was starting one character too far in when skipping
1007        over the character class, thus treating the ] as data rather than
1008        terminating the class. This meant it could skip too much.]
1009    
1010    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1011        /g option in pcretest when the pattern contains \K, which makes it possible
1012        to have an empty string match not at the start, even when the pattern is
1013        anchored. Updated pcretest and pcredemo to use this option.
1014    
1015    21. If the maximum number of capturing subpatterns in a recursion was greater
1016        than the maximum at the outer level, the higher number was returned, but
1017        with unset values at the outer level. The correct (outer level) value is
1018        now given.
1019    
1020    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1021        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1022        make it do so. The string so far is captured, making this feature
1023        compatible with Perl.
1024    
1025    23. The tests have been re-organized, adding tests 11 and 12, to make it
1026        possible to check the Perl 5.10 features against Perl 5.10.
1027    
1028    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1029        pattern matches a fixed length string. PCRE did not allow this; now it
1030        does. Neither allows recursion.
1031    
1032    25. I finally figured out how to implement a request to provide the minimum
1033        length of subject string that was needed in order to match a given pattern.
1034        (It was back references and recursion that I had previously got hung up
1035        on.) This code has now been added to pcre_study(); it finds a lower bound
1036        to the length of subject needed. It is not necessarily the greatest lower
1037        bound, but using it to avoid searching strings that are too short does give
1038        some useful speed-ups. The value is available to calling programs via
1039        pcre_fullinfo().
1040    
1041    26. While implementing 25, I discovered to my embarrassment that pcretest had
1042        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1043        study optimizations had never been tested with that matching function.
1044        Oops. What is worse, even when it was passed study data, there was a bug in
1045        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1046        were also very few tests of studied patterns with pcre_dfa_exec().
1047    
1048    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1049        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1050        on the other side of the coin, they are no longer allowed to have different
1051        names, because these cannot be distinguished in PCRE, and this has caused
1052        confusion. (This is a difference from Perl.)
1053    
1054    28. When duplicate subpattern names are present (necessarily with different
1055        numbers, as required by 27 above), and a test is made by name in a
1056        conditional pattern, either for a subpattern having been matched, or for
1057        recursion in such a pattern, all the associated numbered subpatterns are
1058        tested, and the overall condition is true if the condition is true for any
1059        one of them. This is the way Perl works, and is also more like the way
1060        testing by number works.
1061    
1062    
1063    Version 7.9 11-Apr-09
1064    ---------------------
1065    
1066    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1067        (pcretest), all targets were linked against these libraries. This included
1068        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1069        libraries. This caused unwanted dependencies to be created. This problem
1070        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1071        pcretest is linked with readline.
1072    
1073    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1074        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1075        moved outside it again, because FALSE and TRUE are already defined in AIX,
1076        but BOOL is not.
1077    
1078    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1079        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1080    
1081    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1082        hyphens) following file names and line numbers when outputting matching
1083        lines. This is not true; no spaces are inserted. I have also clarified the
1084        wording for the --colour (or --color) option.
1085    
1086    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1087        was not coloured; this is different to GNU grep, so I have changed it to be
1088        the same.
1089    
1090    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1091        each matching line was coloured. Now it goes on to look for further matches
1092        of any of the test patterns, which is the same behaviour as GNU grep.
1093    
1094    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1095        doesn't make sense to accept an empty string match in pcregrep, so I have
1096        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1097        seems to be how GNU grep behaves.
1098    
1099    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1100        start or after a newline", because the conditional assertion was not being
1101        correctly handled. The rule now is that both the assertion and what follows
1102        in the first alternative must satisfy the test.
1103    
1104    9.  If auto-callout was enabled in a pattern with a conditional group whose
1105        condition was an assertion, PCRE could crash during matching, both with
1106        pcre_exec() and pcre_dfa_exec().
1107    
1108    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1109        used for matching.
1110    
1111    11. Unicode property support in character classes was not working for
1112        characters (bytes) greater than 127 when not in UTF-8 mode.
1113    
1114    12. Added the -M command line option to pcretest.
1115    
1116    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1117    
1118    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1119    
1120    16. Added comments and documentation about mis-use of no_arg in the C++
1121        wrapper.
1122    
1123    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1124        from Martin Jerabek that uses macro names for all relevant character and
1125        string constants.
1126    
1127    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1128        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1129        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1130        these, but not everybody uses configure.
1131    
1132    19. A conditional group that had only one branch was not being correctly
1133        recognized as an item that could match an empty string. This meant that an
1134        enclosing group might also not be so recognized, causing infinite looping
1135        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1136        with the subject "ab", where knowledge that the repeated group can match
1137        nothing is needed in order to break the loop.
1138    
1139    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1140        exec(), but without supplying a callout function, matching went wrong.
1141    
1142    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1143        leak if the size of the offset vector was greater than 30. When the vector
1144        is smaller, the saved offsets during recursion go onto a local stack
1145        vector, but for larger vectors malloc() is used. It was failing to free
1146        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1147        error, in fact).
1148    
1149    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1150        heapframe that is used only when UTF-8 support is enabled. This caused no
1151        problem, but was untidy.
1152    
1153    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1154        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1155        included within another project.
1156    
1157    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1158        slightly modified by me:
1159    
1160          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1161              not building pcregrep.
1162    
1163          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1164              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1165    
1166    25. Forward references, both numeric and by name, in patterns that made use of
1167        duplicate group numbers, could behave incorrectly or give incorrect errors,
1168        because when scanning forward to find the reference group, PCRE was not
1169        taking into account the duplicate group numbers. A pattern such as
1170        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1171    
1172    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1173        the feature of a custom pointer more persuasive (as requested by a user).
1174    
1175    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1176        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1177        pre-defined.
1178    
1179    28. Added support for (*UTF8) at the start of a pattern.
1180    
1181    29. Arrange for flags added by the "release type" setting in CMake to be shown
1182        in the configuration summary.
1183    
1184    
1185    Version 7.8 05-Sep-08
1186    ---------------------
1187    
1188    1.  Replaced UCP searching code with optimized version as implemented for Ad
1189        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1190        stage table and inline lookup instead of a function, giving speed ups of 2
1191        to 5 times on some simple patterns that I tested. Permission was given to
1192        distribute the MultiStage2.py script that generates the tables (it's not in
1193        the tarball, but is in the Subversion repository).
1194    
1195    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1196        scripts.
1197    
1198    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1199        a group with a zero qualifier. The result of the study could be incorrect,
1200        or the function might crash, depending on the pattern.
1201    
1202    4.  Caseless matching was not working for non-ASCII characters in back
1203        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1204        It now works when Unicode Property Support is available.
1205    
1206    5.  In pcretest, an escape such as \x{de} in the data was always generating
1207        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1208        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1209        truncation.
1210    
1211    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1212    
1213    7.  Added two (int) casts to pcregrep when printing the difference of two
1214        pointers, in case they are 64-bit values.
1215    
1216    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1217        test 2 if it fails.
1218    
1219    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1220        and a #define of that name to empty if it is not externally set. This is to
1221        allow users of MSVC to set it if necessary.
1222    
1223    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1224        the convenience functions in the pcre_get.c source file.
1225    
1226    11. An option change at the start of a pattern that had top-level alternatives
1227        could cause overwriting and/or a crash. This command provoked a crash in
1228        some environments:
1229    
1230          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1231    
1232        This potential security problem was recorded as CVE-2008-2371.
1233    
1234    12. For a pattern where the match had to start at the beginning or immediately
1235        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1236        pcre_dfa_exec() could read past the end of the passed subject if there was
1237        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1238        pcretest so that it places the subject at the end of its malloc-ed buffer.
1239    
1240    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1241        exec() might read past the end of the data buffer in UTF-8 mode.
1242    
1243    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1244        the data contained the byte 0x85 as part of a UTF-8 character within its
1245        first line. This applied both to normal and DFA matching.
1246    
1247    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1248        /^[^d]*?$/8 failed to match "abc".
1249    
1250    16. Added a missing copyright notice to pcrecpp_internal.h.
1251    
1252    17. Make it more clear in the documentation that values returned from
1253        pcre_exec() in ovector are byte offsets, not character counts.
1254    
1255    18. Tidied a few places to stop certain compilers from issuing warnings.
1256    
1257    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1258        supplied by Stefan Weber. I made a further small update for 7.8 because
1259        there is a change of source arrangements: the pcre_searchfuncs.c module is
1260        replaced by pcre_ucd.c.
1261    
1262    
1263    Version 7.7 07-May-08
1264    ---------------------
1265    
1266    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1267        a string to a long long, pretend we don't even have a long long." This is
1268        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1269    
1270    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1271        pre-7.6 versions, which defined a global no_arg variable instead of putting
1272        it in the RE class. (See also #8 below.)
1273    
1274    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1275        Lopes.
1276    
1277    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1278    
1279        (1) The include/exclude patterns were being applied to the whole pathnames
1280            of files, instead of just to the final components.
1281    
1282        (2) If there was more than one level of directory, the subdirectories were
1283            skipped unless they satisfied the include/exclude conditions. This is
1284            inconsistent with GNU grep (and could even be seen as contrary to the
1285            pcregrep specification - which I improved to make it absolutely clear).
1286            The action now is always to scan all levels of directory, and just
1287            apply the include/exclude patterns to regular files.
1288    
1289    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1290        --exclude_dir in the tests to avoid scanning .svn directories.
1291    
1292    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1293        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1294        doesn't support NULs in patterns.
1295    
1296    7.  Added some missing "const"s to declarations of static tables in
1297        pcre_compile.c and pcre_dfa_exec.c.
1298    
1299    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1300        caused by fix #2  above. (Subsequently also a second patch to fix the
1301        first patch. And a third patch - this was a messy problem.)
1302    
1303    9.  Applied Craig's patch to remove the use of push_back().
1304    
1305    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1306        matching function regexec().
1307    
1308    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1309        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1310        references. PCRE supports relative numbers with this syntax (I don't think
1311        Oniguruma does).
1312    
1313    12. Previously, a group with a zero repeat such as (...){0} was completely
1314        omitted from the compiled regex. However, this means that if the group
1315        was called as a subroutine from elsewhere in the pattern, things went wrong
1316        (an internal error was given). Such groups are now left in the compiled
1317        pattern, with a new opcode that causes them to be skipped at execution
1318        time.
1319    
1320    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1321        to the way PCRE behaves:
1322    
1323        (a) A lone ] character is dis-allowed (Perl treats it as data).
1324    
1325        (b) A back reference to an unmatched subpattern matches an empty string
1326            (Perl fails the current match path).
1327    
1328        (c) A data ] in a character class must be notated as \] because if the
1329            first data character in a class is ], it defines an empty class. (In
1330            Perl it is not possible to have an empty class.) The empty class []
1331            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1332            The negative empty class [^] matches any one character, independently
1333            of the DOTALL setting.
1334    
1335    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1336        non-existent subpattern following a character class starting with ']' and
1337        containing () gave an internal compiling error instead of "reference to
1338        non-existent subpattern". Fortunately, when the pattern did exist, the
1339        compiled code was correct. (When scanning forwards to check for the
1340        existencd of the subpattern, it was treating the data ']' as terminating
1341        the class, so got the count wrong. When actually compiling, the reference
1342        was subsequently set up correctly.)
1343    
1344    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1345        it was being rejected as not supported by pcre_dfa_exec(), even though
1346        other assertions are supported. I have made pcre_dfa_exec() support
1347        (*FAIL).
1348    
1349    16. The implementation of 13c above involved the invention of a new opcode,
1350        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1351        cannot be changed at match time, I realized I could make a small
1352        improvement to matching performance by compiling OP_ALLANY instead of
1353        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1354        on the OP_ANY path.
1355    
1356    17. Compiling pcretest on Windows with readline support failed without the
1357        following two fixes: (1) Make the unistd.h include conditional on
1358        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1359    
1360    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1361        ncurses library to be included for pcretest when ReadLine support is
1362        requested, but also to allow for it to be overridden. This patch came from
1363        Daniel Bergström.
1364    
1365    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1366        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1367        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1368        spotting this.
1369    
1370    
1371    Version 7.6 28-Jan-08
1372    ---------------------
1373    
1374    1.  A character class containing a very large number of characters with
1375        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1376        overflow.
1377    
1378    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1379        HAVE_LONG_LONG is not defined.
1380    
1381    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1382        bring them up to date and include new features. This patch includes:
1383    
1384        - Fixed PH's badly added libz and libbz2 support.
1385        - Fixed a problem with static linking.
1386        - Added pcredemo. [But later removed - see 7 below.]
1387        - Fixed dftables problem and added an option.
1388        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1389            HAVE_LONG_LONG.
1390        - Added readline support for pcretest.
1391        - Added an listing of the option settings after cmake has run.
1392    
1393    4.  A user submitted a patch to Makefile that makes it easy to create
1394        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1395        Makefile.am that cause it to include this special target, without
1396        affecting anything else. Note that the same mingw target plus all
1397        the other distribution libraries and programs are now supported
1398        when configuring with CMake (see 6 below) instead of with
1399        Configure/Make.
1400    
1401    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1402        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1403        exported in the Windows port". It has not yet been confirmed that the patch
1404        solves the problem, but it does no harm.
1405    
1406    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1407        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1408        with CMake, and also correct the comment about stack recursion.
1409    
1410    7.  Remove the automatic building of pcredemo from the ./configure system and
1411        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1412        of a program that users should build themselves after PCRE is installed, so
1413        building it automatically is not really right. What is more, it gave
1414        trouble in some build environments.
1415    
1416    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1417    
1418    
1419    Version 7.5 10-Jan-08
1420    ---------------------
1421    
1422    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1423        values in parens when parsing an RE using the C++ wrapper."
1424    
1425    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1426        Characters greater than 255 were excluded from the class instead of being
1427        included.
1428    
1429    3.  The same bug as (2) above applied to negated POSIX classes such as
1430        [:^space:].
1431    
1432    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1433        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1434        I have changed it.
1435    
1436    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1437        first named subpattern) and a construct such as (?&a) would reference the
1438        first named subpattern whose name started with "a" (in other words, the
1439        length check was missing). Both these problems are fixed. "Subpattern name
1440        expected" is now given for (?&) (a zero-length name), and this patch also
1441        makes it give the same error for \k'' (previously it complained that that
1442        was a reference to a non-existent subpattern).
1443    
1444    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1445        this is right because (?- can be followed by option settings as well as by
1446        digits. I have, however, made the messages clearer.
1447    
1448    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1449        than the number used in the conditional) now cause a compile-time error.
1450        This is actually not compatible with Perl, which accepts such patterns, but
1451        treats the conditional as always being FALSE (as PCRE used to), but it
1452        seems to me that giving a diagnostic is better.
1453    
1454    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1455        and messages.
1456    
1457    9.  Fix two occurrences of "backslash" in comments that should have been
1458        "backspace".
1459    
1460    10. Remove two redundant lines of code that can never be obeyed (their function
1461        was moved elsewhere).
1462    
1463    11. The program that makes PCRE's Unicode character property table had a bug
1464        which caused it to generate incorrect table entries for sequences of
1465        characters that have the same character type, but are in different scripts.
1466        It amalgamated them into a single range, with the script of the first of
1467        them. In other words, some characters were in the wrong script. There were
1468        thirteen such cases, affecting characters in the following ranges:
1469    
1470          U+002b0 - U+002c1
1471          U+0060c - U+0060d
1472          U+0061e - U+00612
1473          U+0064b - U+0065e
1474          U+0074d - U+0076d
1475          U+01800 - U+01805
1476          U+01d00 - U+01d77
1477          U+01d9b - U+01dbf
1478          U+0200b - U+0200f
1479          U+030fc - U+030fe
1480          U+03260 - U+0327f
1481          U+0fb46 - U+0fbb1
1482          U+10450 - U+1049d
1483    
1484    12. The -o option (show only the matching part of a line) for pcregrep was not
1485        compatible with GNU grep in that, if there was more than one match in a
1486        line, it showed only the first of them. It now behaves in the same way as
1487        GNU grep.
1488    
1489    13. If the -o and -v options were combined for pcregrep, it printed a blank
1490        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1491        does the same. The return code can be used to tell if there were any
1492        non-matching lines.
1493    
1494    14. Added --file-offsets and --line-offsets to pcregrep.
1495    
1496    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1497        infinitely looping recursion. The bug was that positive lookaheads were not
1498        being skipped when checking for a possible empty match (negative lookaheads
1499        and both kinds of lookbehind were skipped).
1500    
1501    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1502        inclusion of <windows.h> to before rather than after the definition of
1503        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1504    
1505    17. Specifying a possessive quantifier with a specific limit for a Unicode
1506        character property caused pcre_compile() to compile bad code, which led at
1507        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1508        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1509        caused the error; without that there was no problem.
1510    
1511    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1512    
1513    19. Added --enable-pcretest-libreadline.
1514    
1515    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1516        RE::GlobalReplace(). As a result, the number of replacements returned was
1517        double what it should be. I removed one of the increments, but Craig sent a
1518        later patch that removed the other one (the right fix) and added unit tests
1519        that check the return values (which was not done before).
1520    
1521    21. Several CMake things:
1522    
1523        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1524            the names libpcre and libpcreposix, not just pcre and pcreposix.
1525    
1526        (2) The above change means that pcretest and pcregrep are now correctly
1527            linked with the newly-built libraries, not previously installed ones.
1528    
1529        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1530    
1531    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1532        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1533        UTF-8 newline character). The key issue is that the pattern starts .*;
1534        this means that the match must be either at the beginning, or after a
1535        newline. The bug was in the code for advancing after a failed match and
1536        checking that the new position followed a newline. It was not taking
1537        account of UTF-8 characters correctly.
1538    
1539    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1540        character classes. PCRE was not treating the sequence [:...:] as a
1541        character class unless the ... were all letters. Perl, however, seems to
1542        allow any characters between [: and :], though of course it rejects as
1543        unknown any "names" that contain non-letters, because all the known class
1544        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1545        for example, whereas PCRE did not - it did not recognize a POSIX character
1546        class. This seemed a bit dangerous, so the code has been changed to be
1547        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1548        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1549        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1550        Perl does, and where it didn't before.
1551    
1552    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1553        Windows environments %n is disabled by default.
1554    
1555    
1556    Version 7.4 21-Sep-07
1557    ---------------------
1558    
1559    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1560        means that a class such as [\s] counted as "explicit reference to CR or
1561        LF". That isn't really right - the whole point of the change was to try to
1562        help when there was an actual mention of one of the two characters. So now
1563        the change happens only if \r or \n (or a literal CR or LF) character is
1564        encountered.
1565    
1566    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1567        of both had grown to the point where there were only 3 bits left.
1568        Fortunately, there was spare space in the data structure, and so I have
1569        moved the internal flags into a new 16-bit field to free up more option
1570        bits.
1571    
1572    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1573        but did not set the internal JCHANGED flag - either of these is enough to
1574        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1575        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1576        start sets both bits.
1577    
1578    4.  Added options (at build time, compile time, exec time) to change \R from
1579        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1580    
1581    5.  doc/pcresyntax.html was missing from the distribution.
1582    
1583    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1584        compatibility, even though it is no longer used.
1585    
1586    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1587        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1588        windows.h file is present (where different names are used). [This was
1589        reversed later after testing - see 16 below.]
1590    
1591    8.  Changed all #include <config.h> to #include "config.h". There were also
1592        some further <pcre.h> cases that I changed to "pcre.h".
1593    
1594    9.  When pcregrep was used with the --colour option, it missed the line ending
1595        sequence off the lines that it output.
1596    
1597    10. It was pointed out to me that arrays of string pointers cause lots of
1598        relocations when a shared library is dynamically loaded. A technique of
1599        using a single long string with a table of offsets can drastically reduce
1600        these. I have refactored PCRE in four places to do this. The result is
1601        dramatic:
1602    
1603          Originally:                          290
1604          After changing UCP table:            187
1605          After changing error message table:   43
1606          After changing table of "verbs"       36
1607          After changing table of Posix names   22
1608    
1609        Thanks to the folks working on Gregex for glib for this insight.
1610    
1611    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1612        unicode-properties was also set.
1613    
1614    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1615    
1616    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1617        checked only for CRLF.
1618    
1619    14. Added casts to pcretest.c to avoid compiler warnings.
1620    
1621    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1622    
1623    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1624        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1625        entirely. This removes changes made in 7 above.
1626    
1627    17. The CMake files have been updated, and there is now more information about
1628        building with CMake in the NON-UNIX-USE document.
1629    
1630    
1631    Version 7.3 28-Aug-07
1632    ---------------------
1633    
1634     1. In the rejigging of the build system that eventually resulted in 7.1, the
1635        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1636        brackets there is not right, since it causes compilers to look for an
1637        installed pcre.h, not the version that is in the source that is being
1638        compiled (which of course may be different). I have changed it back to:
1639    
1640          #include "pcre.h"
1641    
1642        I have a vague recollection that the change was concerned with compiling in
1643        different directories, but in the new build system, that is taken care of
1644        by the VPATH setting the Makefile.
1645    
1646     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1647        when the subject happened to end in the byte 0x85 (e.g. if the last
1648        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1649        characters but of course it shouldn't be taken as a newline when it is part
1650        of another character. The bug was that, for an unlimited repeat of . in
1651        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1652        characters when looking for a newline.
1653    
1654     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1655    
1656     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1657        in debug output.
1658    
1659     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1660        long printing in the pcrecpp unittest when running under MinGW.
1661    
1662     6. ESC_K was left out of the EBCDIC table.
1663    
1664     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1665        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1666        limit also applies to "virtual nesting" when a pattern is recursive, and in
1667        this case 1000 isn't so big. I have been able to remove this limit at the
1668        expense of backing off one optimization in certain circumstances. Normally,
1669        when pcre_exec() would call its internal match() function recursively and
1670        immediately return the result unconditionally, it uses a "tail recursion"
1671        feature to save stack. However, when a subpattern that can match an empty
1672        string has an unlimited repetition quantifier, it no longer makes this
1673        optimization. That gives it a stack frame in which to save the data for
1674        checking that an empty string has been matched. Previously this was taken
1675        from the 1000-entry workspace that had been reserved. So now there is no
1676        explicit limit, but more stack is used.
1677    
1678     8. Applied Daniel's patches to solve problems with the import/export magic
1679        syntax that is required for Windows, and which was going wrong for the
1680        pcreposix and pcrecpp parts of the library. These were overlooked when this
1681        problem was solved for the main library.
1682    
1683     9. There were some crude static tests to avoid integer overflow when computing
1684        the size of patterns that contain repeated groups with explicit upper
1685        limits. As the maximum quantifier is 65535, the maximum group length was
1686        set at 30,000 so that the product of these two numbers did not overflow a
1687        32-bit integer. However, it turns out that people want to use groups that
1688        are longer than 30,000 bytes (though not repeat them that many times).
1689        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1690        made it possible to implement the integer overflow checks in a much more
1691        dynamic way, which I have now done. The artificial limitation on group
1692        length has been removed - we now have only the limit on the total length of
1693        the compiled pattern, which depends on the LINK_SIZE setting.
1694    
1695    10. Fixed a bug in the documentation for get/copy named substring when
1696        duplicate names are permitted. If none of the named substrings are set, the
1697        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1698        empty string.
1699    
1700    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1701        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1702        because the ] is interpreted as the first data character and the
1703        terminating ] is not found. PCRE has been made compatible with Perl in this
1704        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1705        cause memory overwriting.
1706    
1707    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1708        string has been matched (to stop an infinite loop). It was not recognizing
1709        a conditional subpattern that could match an empty string if that
1710        subpattern was within another subpattern. For example, it looped when
1711        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1712        condition was not nested. This bug has been fixed.
1713    
1714    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1715        past the start of the subject in the presence of bytes with the top bit
1716        set, for example "\x8aBCD".
1717    
1718    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1719        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1720    
1721    14. Optimized (?!) to (*FAIL).
1722    
1723    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1724        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1725        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1726        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1727        does: it's just the validity check that is more restrictive.
1728    
1729    16. Inserted checks for integer overflows during escape sequence (backslash)
1730        processing, and also fixed erroneous offset values for syntax errors during
1731        backslash processing.
1732    
1733    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1734        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1735    
1736    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1737        caused an overrun.
1738    
1739    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1740        something other than just ASCII characters) inside a group that had an
1741        unlimited repeat caused a loop at compile time (while checking to see
1742        whether the group could match an empty string).
1743    
1744    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1745        [\P{Any}] did so. (Error in the code for printing property names.)
1746    
1747    21. An orphan \E inside a character class could cause a crash.
1748    
1749    22. A repeated capturing bracket such as (A)? could cause a wild memory
1750        reference during compilation.
1751    
1752    23. There are several functions in pcre_compile() that scan along a compiled
1753        expression for various reasons (e.g. to see if it's fixed length for look
1754        behind). There were bugs in these functions when a repeated \p or \P was
1755        present in the pattern. These operators have additional parameters compared
1756        with \d, etc, and these were not being taken into account when moving along
1757        the compiled data. Specifically:
1758    
1759        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1760            length.
1761    
1762        (b) An item such as \pL+ within a repeated group could cause crashes or
1763            loops.
1764    
1765        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1766            "reference to non-existent subpattern" error.
1767    
1768        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1769    
1770    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1771        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1772    
1773    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1774    
1775    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1776        character were causing crashes (broken optimization).
1777    
1778    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1779        \p or \P) caused a compile-time loop.
1780    
1781    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1782        break. For example, the unstudied pattern [\r\n]A does not match the string
1783        "\r\nA" because change 7.0/46 below moves the current point on by two
1784        characters after failing to match at the start. However, the pattern \nA
1785        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1786        the same is true. There doesn't seem any very clean way out of this, but
1787        what I have chosen to do makes the common cases work: PCRE now takes note
1788        of whether there can be an explicit match for \r or \n anywhere in the
1789        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1790        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1791        pattern has explicit CR or LF references.
1792    
1793    29. Added (*CR) etc for changing newline setting at start of pattern.
1794    
1795    
1796    Version 7.2 19-Jun-07
1797    ---------------------
1798    
1799     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1800        which is apparently normally available under Windows.
1801    
1802     2. Re-jig the pcregrep tests with different newline settings in an attempt
1803        to make them independent of the local environment's newline setting.
1804    
1805     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1806    
1807     4. Some of the "internals" tests were previously cut out when the link size
1808        was not 2, because the output contained actual offsets. The recent new
1809        "Z" feature of pcretest means that these can be cut out, making the tests
1810        usable with all link sizes.
1811    
1812     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1813        stack recursion. This gives a massive performance boost under BSD, but just
1814        a small improvement under Linux. However, it saves one field in the frame
1815        in all cases.
1816    
1817     6. Added more features from the forthcoming Perl 5.10:
1818    
1819        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1820            recursion call. It refers to the nth most recently opened parentheses.
1821    
1822        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1823            to be opened parentheses.
1824    
1825        (c) Conditions that refer to capturing parentheses can be specified
1826            relatively, for example, (?(-2)... or (?(+3)...
1827    
1828        (d) \K resets the start of the current match so that everything before
1829            is not part of it.
1830    
1831        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1832    
1833        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1834            reference syntax.
1835    
1836        (g) (?| introduces a group in which the numbering of parentheses in each
1837            alternative starts with the same number.
1838    
1839        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1840    
1841     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1842        PCRE_INFO_JCHANGED.
1843    
1844     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1845        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1846        for detecting groups that can match an empty string.
1847    
1848     9. A pattern with a very large number of alternatives (more than several
1849        hundred) was running out of internal workspace during the pre-compile
1850        phase, where pcre_compile() figures out how much memory will be needed. A
1851        bit of new cunning has reduced the workspace needed for groups with
1852        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1853        workspace instead of running out of the 4096 that are available.
1854    
1855    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1856    
1857    11. Applied patch from Google to remove an optimization that didn't quite work.
1858        The report of the bug said:
1859    
1860          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1861          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1862          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1863    
1864    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1865        it matched the wrong number of bytes.
1866    
1867    
1868    Version 7.1 24-Apr-07
1869    ---------------------
1870    
1871     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1872        that is more "standard", making use of automake and other Autotools. There
1873        is some re-arrangement of the files and adjustment of comments consequent
1874        on this.
1875    
1876     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1877        for recursive directory scanning broke on some systems because the files
1878        are not scanned in any specific order and on different systems the order
1879        was different. A call to "sort" has been inserted into RunGrepTest for the
1880        approprate test as a short-term fix. In the longer term there may be an
1881        alternative.
1882    
1883     3. I had an email from Eric Raymond about problems translating some of PCRE's
1884        man pages to HTML (despite the fact that I distribute HTML pages, some
1885        people do their own conversions for various reasons). The problems
1886        concerned the use of low-level troff macros .br and .in. I have therefore
1887        removed all such uses from the man pages (some were redundant, some could
1888        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1889        HTML has been updated to handle .nf/.fi and to complain if it encounters
1890        .br or .in.
1891    
1892     4. Updated comments in configure.ac that get placed in config.h.in and also
1893        arranged for config.h to be included in the distribution, with the name
1894        config.h.generic, for the benefit of those who have to compile without
1895        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1896    
1897     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1898        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1899        makevp.bat for the current PCRE, using the additional files
1900        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1901    
1902     6. A Windows user reported a minor discrepancy with test 2, which turned out
1903        to be caused by a trailing space on an input line that had got lost in his
1904        copy. The trailing space was an accident, so I've just removed it.
1905    
1906     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1907        that is needed.
1908    
1909     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1910        as "const" (a) because they are and (b) because it helps the PHP
1911        maintainers who have recently made a script to detect big data structures
1912        in the php code that should be moved to the .rodata section. I remembered
1913        to update Builducptable as well, so it won't revert if ucptable.h is ever
1914        re-created.
1915    
1916     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1917        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1918        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1919        support is not required. This saves 1.5-2K of code, which is important in
1920        some applications.
1921    
1922        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1923        so as not to refer to the tables, even though these functions will never be
1924        called when UTF-8 support is disabled. Otherwise there are problems with a
1925        shared library.
1926    
1927    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1928    
1929        (a) It was defining its arguments as char * instead of void *.
1930    
1931        (b) It was assuming that all moves were upwards in memory; this was true
1932            a long time ago when I wrote it, but is no longer the case.
1933    
1934        The emulated memove() is provided for those environments that have neither
1935        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1936        is clearly not the case, as these two bugs were recently reported.
1937    
1938    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1939        and Detrail to create the HTML documentation, the .txt form of the man
1940        pages, and it removes trailing spaces from listed files. It also creates
1941        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1942        case, it wraps all the #defines with #ifndefs. This script should be run
1943        before "make dist".
1944    
1945    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1946        with Unicode property support.
1947    
1948        (a) For a maximizing quantifier, if the two different cases of the
1949            character were of different lengths in their UTF-8 codings (there are
1950            some cases like this - I found 11), and the matching function had to
1951            back up over a mixture of the two cases, it incorrectly assumed they
1952            were both the same length.
1953    
1954        (b) When PCRE was configured to use the heap rather than the stack for
1955            recursion during matching, it was not correctly preserving the data for
1956            the other case of a UTF-8 character when checking ahead for a match
1957            while processing a minimizing repeat. If the check also involved
1958            matching a wide character, but failed, corruption could cause an
1959            erroneous result when trying to check for a repeat of the original
1960            character.
1961    
1962    13. Some tidying changes to the testing mechanism:
1963    
1964        (a) The RunTest script now detects the internal link size and whether there
1965            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1966            values substituted by "configure". (The RunGrepTest script already did
1967            this for UTF-8.) The configure.ac script no longer substitutes the
1968            relevant variables.
1969    
1970        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1971            with length and offset values. This means that the output is different
1972            for different internal link sizes. Test 2 is skipped for link sizes
1973            other than 2 because of this, bypassing the problem. Unfortunately,
1974            there was also a test in test 3 (the locale tests) that used /B and
1975            failed for link sizes other than 2. Rather than cut the whole test out,
1976            I have added a new /Z option to pcretest that replaces the length and
1977            offset values with spaces. This is now used to make test 3 independent
1978            of link size. (Test 2 will be tidied up later.)
1979    
1980    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1981        segmentation fault instead of returning the appropriate error message.
1982    
1983    15. In multiline mode when the newline sequence was set to "any", the pattern
1984        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1985        This doesn't seem right; it now treats the CRLF combination as the line
1986        ending, and so does not match in that case. It's only a pattern such as ^$
1987        that would hit this one: something like ^ABC$ would have failed after \r
1988        and then tried again after \r\n.
1989    
1990    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1991        in an attempt to make files that differ only in their line terminators
1992        compare equal. This works on Linux.
1993    
1994    17. Under certain error circumstances pcregrep might try to free random memory
1995        as it exited. This is now fixed, thanks to valgrind.
1996    
1997    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1998        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1999        was because its rules for how to advance for /g after matching an empty
2000        string at the end of a line did not allow for this case. They now check for
2001        it specially.
2002    
2003    20. pcretest is supposed to handle patterns and data of any length, by
2004        extending its buffers when necessary. It was getting this wrong when the
2005        buffer for a data line had to be extended.
2006    
2007    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2008        CRLF as a newline sequence.
2009    
2010    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2011        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2012        I have nevertheless tidied it up.
2013    
2014    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2015    
2016    24. Added a man page for pcre-config.
2017    
2018    
2019    Version 7.0 19-Dec-06
2020    ---------------------
2021    
2022     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2023        moving to gcc 4.1.1.
2024    
2025     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2026        sys/time.h, which is documented as needed for this function. It doesn't
2027        seem to matter on Linux, but it showed up on some releases of OS X.
2028    
2029     3. It seems that there are systems where bytes whose values are greater than
2030        127 match isprint() in the "C" locale. The "C" locale should be the
2031        default when a C program starts up. In most systems, only ASCII printing
2032        characters match isprint(). This difference caused the output from pcretest
2033        to vary, making some of the tests fail. I have changed pcretest so that:
2034    
2035        (a) When it is outputting text in the compiled version of a pattern, bytes
2036            other than 32-126 are always shown as hex escapes.
2037    
2038        (b) When it is outputting text that is a matched part of a subject string,
2039            it does the same, unless a different locale has been set for the match
2040            (using the /L modifier). In this case, it uses isprint() to decide.
2041    
2042     4. Fixed a major bug that caused incorrect computation of the amount of memory
2043        required for a compiled pattern when options that changed within the
2044        pattern affected the logic of the preliminary scan that determines the
2045        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2046        that the computed length was too small. The symptoms of this bug were
2047        either the PCRE error "internal error: code overflow" from pcre_compile(),
2048        or a glibc crash with a message such as "pcretest: free(): invalid next
2049        size (fast)". Examples of patterns that provoked this bug (shown in
2050        pcretest format) are:
2051    
2052          /(?-x: )/x
2053          /(?x)(?-x: \s*#\s*)/
2054          /((?i)[\x{c0}])/8
2055          /(?i:[\x{c0}])/8
2056    
2057        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2058        is now done differently.
2059    
2060     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2061        wrapper classes; (b) implement a new function in the C++ scanner that is
2062        more efficient than the old way of doing things because it avoids levels of
2063        recursion in the regex matching; (c) add a paragraph to the documentation
2064        for the FullMatch() function.
2065    
2066     6. The escape sequence \n was being treated as whatever was defined as
2067        "newline". Not only was this contrary to the documentation, which states
2068        that \n is character 10 (hex 0A), but it also went horribly wrong when
2069        "newline" was defined as CRLF. This has been fixed.
2070    
2071     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2072        was being set to -1 for the "end of line" case (supposedly a value that no
2073        character can have). Though this value is never used (the check for end of
2074        line is "zero bytes in current character"), it caused compiler complaints.
2075        I've changed it to 0xffffffff.
2076    
2077     8. In pcre_version.c, the version string was being built by a sequence of
2078        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2079        string (as it is for production releases) called a macro with an empty
2080        argument. The C standard says the result of this is undefined. The gcc
2081        compiler treats it as an empty string (which was what was wanted) but it is
2082        reported that Visual C gives an error. The source has been hacked around to
2083        avoid this problem.
2084    
2085     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2086        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2087        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2088        of them did).
2089    
2090    10. Originally, pcretest opened its input and output without "b"; then I was
2091        told that "b" was needed in some environments, so it was added for release
2092        5.0 to both the input and output. (It makes no difference on Unix-like
2093        systems.) Later I was told that it is wrong for the input on Windows. I've
2094        now abstracted the modes into two macros, to make it easier to fiddle with
2095        them, and removed "b" from the input mode under Windows.
2096    
2097    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2098    
2099    12. Added -help and --help to pcretest as an official way of being reminded
2100        of the options.
2101    
2102    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2103        and pcrecpp.cc because they annoy compilers at high warning levels.
2104    
2105    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2106    
2107    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2108        scripts are not C programs :-) and which was not noticed because it works
2109        on Linux.
2110    
2111    16. pcretest is supposed to handle any length of pattern and data line (as one
2112        line or as a continued sequence of lines) by extending its input buffer if
2113        necessary. This feature was broken for very long pattern lines, leading to
2114        a string of junk being passed to pcre_compile() if the pattern was longer
2115        than about 50K.
2116    
2117    17. I have done a major re-factoring of the way pcre_compile() computes the
2118        amount of memory needed for a compiled pattern. Previously, there was code
2119        that made a preliminary scan of the pattern in order to do this. That was
2120        OK when PCRE was new, but as the facilities have expanded, it has become
2121        harder and harder to keep it in step with the real compile phase, and there
2122        have been a number of bugs (see for example, 4 above). I have now found a
2123        cunning way of running the real compile function in a "fake" mode that
2124        enables it to compute how much memory it would need, while actually only
2125        ever using a few hundred bytes of working memory and without too many
2126        tests of the mode. This should make future maintenance and development
2127        easier. A side effect of this work is that the limit of 200 on the nesting
2128        depth of parentheses has been removed (though this was never a serious
2129        limitation, I suspect). However, there is a downside: pcre_compile() now
2130        runs more slowly than before (30% or more, depending on the pattern). I
2131        hope this isn't a big issue. There is no effect on runtime performance.
2132    
2133    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2134        newline (only possible for the last line of a file) and it was a
2135        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2136    
2137    19. Added additional timing features to pcretest. (1) The -tm option now times
2138        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2139        separate command line item, by a number that specifies the number of
2140        repeats to use when timing. The default is 50000; this gives better
2141        precision, but takes uncomfortably long for very large patterns.
2142    
2143    20. Extended pcre_study() to be more clever in cases where a branch of a
2144        subpattern has no definite first character. For example, (a*|b*)[cd] would
2145        previously give no result from pcre_study(). Now it recognizes that the
2146        first character must be a, b, c, or d.
2147    
2148    21. There was an incorrect error "recursive call could loop indefinitely" if
2149        a subpattern (or the entire pattern) that was being tested for matching an
2150        empty string contained only one non-empty item after a nested subpattern.
2151        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2152        incorrectly, because the \d was being skipped in the check.
2153    
2154    22. The pcretest program now has a new pattern option /B and a command line
2155        option -b, which is equivalent to adding /B to every pattern. This causes
2156        it to show the compiled bytecode, without the additional information that
2157        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2158        is the same as /B/I).
2159    
2160    23. A new optimization is now able automatically to treat some sequences such
2161        as a*b as a*+b. More specifically, if something simple (such as a character
2162        or a simple class like \d) has an unlimited quantifier, and is followed by
2163        something that cannot possibly match the quantified thing, the quantifier
2164        is automatically "possessified".
2165    
2166    24. A recursive reference to a subpattern whose number was greater than 39
2167        went wrong under certain circumstances in UTF-8 mode. This bug could also
2168        have affected the operation of pcre_study().
2169    
2170    25. Realized that a little bit of performance could be had by replacing
2171        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2172    
2173    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2174    
2175    27. Possessive quantifiers such as a++ were previously implemented by turning
2176        them into atomic groups such as ($>a+). Now they have their own opcodes,
2177        which improves performance. This includes the automatically created ones
2178        from 23 above.
2179    
2180    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2181        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2182        the first matched character to be a colon. This applied both to named and
2183        numbered groups.
2184    
2185    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2186    
2187    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2188        building PCRE on Windows easier, so I have included it in the distribution.
2189    
2190    31. There is now a check in pcretest against a ridiculously large number being
2191        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2192        loop, the loop is abandoned.
2193    
2194    32. Forward references to subpatterns in conditions such as (?(2)...) where
2195        subpattern 2 is defined later cause pcre_compile() to search forwards in
2196        the pattern for the relevant set of parentheses. This search went wrong
2197        when there were unescaped parentheses in a character class, parentheses
2198        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2199    
2200    33. "Subroutine" calls and backreferences were previously restricted to
2201        referencing subpatterns earlier in the regex. This restriction has now
2202        been removed.
2203    
2204    34. Added a number of extra features that are going to be in Perl 5.10. On the
2205        whole, these are just syntactic alternatives for features that PCRE had
2206        previously implemented using the Python syntax or my own invention. The
2207        other formats are all retained for compatibility.
2208    
2209        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2210            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2211            also .NET compatible.
2212    
2213        (b) A recursion or subroutine call to a named group can now be defined as
2214            (?&name) as well as (?P>name).
2215    
2216        (c) A backreference to a named group can now be defined as \k<name> or
2217            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2218            5.10, are also .NET compatible.
2219    
2220        (d) A conditional reference to a named group can now use the syntax
2221            (?(<name>) or (?('name') as well as (?(name).
2222    
2223        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2224            groups (named and numbered) that are never evaluated inline, but can be
2225            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2226            is always false. There may be only one alternative in such a group.
2227    
2228        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2229            as the simple (?(R). The condition is true only if the most recent
2230            recursion is that of the given number or name. It does not search out
2231            through the entire recursion stack.
2232    
2233        (g) The escape \gN or \g{N} has been added, where N is a positive or
2234            negative number, specifying an absolute or relative reference.
2235    
2236    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2237        some "unreachable code" warnings.
2238    
2239    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2240        things, this adds five new scripts.
2241    
2242    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2243        There were also incompatibilities regarding the handling of \Q..\E inside
2244        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2245        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2246    
2247    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2248        matches an empty string, and forcibly breaks the loop. There were bugs in
2249        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2250        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2251        separate and independent bugs (that affected different cases) have been
2252        fixed.
2253    
2254    39. Refactored the code to abolish the use of different opcodes for small
2255        capturing bracket numbers. This is a tidy that I avoided doing when I
2256        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2257        The new approach is not only tidier, it makes it possible to reduce the
2258        memory needed to fix the previous bug (38).
2259    
2260    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2261        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2262        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2263        mode.
2264    
2265    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2266        report.
2267    
2268    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2269        copy construction and assignment in the C++ wrapper.
2270    
2271    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2272        couple of bugs that could have given wrong results in the "--newline=crlf"
2273        case.
2274    
2275    44. Added a number of casts and did some reorganization of signed/unsigned int
2276        variables following suggestions from Dair Grant. Also renamed the variable
2277        "this" as "item" because it is a C++ keyword.
2278    
2279    45. Arranged for dftables to add
2280    
2281          #include "pcre_internal.h"
2282    
2283        to pcre_chartables.c because without it, gcc 4.x may remove the array
2284        definition from the final binary if PCRE is built into a static library and
2285        dead code stripping is activated.
2286    
2287    46. For an unanchored pattern, if a match attempt fails at the start of a
2288        newline sequence, and the newline setting is CRLF or ANY, and the next two
2289        characters are CRLF, advance by two characters instead of one.
2290    
2291    
2292    Version 6.7 04-Jul-06
2293    ---------------------
2294    
2295     1. In order to handle tests when input lines are enormously long, pcretest has
2296        been re-factored so that it automatically extends its buffers when
2297        necessary. The code is crude, but this _is_ just a test program. The
2298        default size has been increased from 32K to 50K.
2299    
2300     2. The code in pcre_study() was using the value of the re argument before
2301        testing it for NULL. (Of course, in any sensible call of the function, it
2302        won't be NULL.)
2303    
2304     3. The memmove() emulation function in pcre_internal.h, which is used on
2305        systems that lack both memmove() and bcopy() - that is, hardly ever -
2306        was missing a "static" storage class specifier.
2307    
2308     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2309        containing an extended class (one that cannot be represented by a bitmap
2310        because it contains high-valued characters or Unicode property items, e.g.
2311        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2312        pattern, but PCRE should not loop if you do not (it no longer does).
2313        [Detail: two cases were found: (a) a repeated subpattern containing an
2314        extended class; (b) a recursive reference to a subpattern that followed a
2315        previous extended class. It wasn't skipping over the extended class
2316        correctly when UTF-8 mode was not set.]
2317    
2318     5. A negated single-character class was not being recognized as fixed-length
2319        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2320        compile error "lookbehind assertion is not fixed length".
2321    
2322     6. The RunPerlTest auxiliary script was showing an unexpected difference
2323        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2324        write a Perl script that can interpret lines of an input file either as
2325        byte characters or as UTF-8, which is what "perltest" was being required to
2326        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2327        can't do is switch easily at run time between having the "use utf8;" pragma
2328        or not. In the end, I fudged it by using the RunPerlTest script to insert
2329        "use utf8;" explicitly for the UTF-8 tests.
2330    
2331     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2332        the end of the subject string, contrary to the documentation and to what
2333        Perl does. This was true of both matching functions. Now it matches only at
2334        the start of the subject and immediately after *internal* newlines.
2335    
2336     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2337        a pointer to an int instead of a pointer to an unsigned long int. This
2338        caused problems on 64-bit systems.
2339    
2340     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2341        instance of the 'standard' template library not being so standard".
2342    
2343    10. There was no check on the number of named subpatterns nor the maximum
2344        length of a subpattern name. The product of these values is used to compute
2345        the size of the memory block for a compiled pattern. By supplying a very
2346        long subpattern name and a large number of named subpatterns, the size
2347        computation could be caused to overflow. This is now prevented by limiting
2348        the length of names to 32 characters, and the number of named subpatterns
2349        to 10,000.
2350    
2351    11. Subpatterns that are repeated with specific counts have to be replicated in
2352        the compiled pattern. The size of memory for this was computed from the
2353        length of the subpattern and the repeat count. The latter is limited to
2354        65535, but there was no limit on the former, meaning that integer overflow
2355        could in principle occur. The compiled length of a repeated subpattern is
2356        now limited to 30,000 bytes in order to prevent this.
2357    
2358    12. Added the optional facility to have named substrings with the same name.
2359    
2360    13. Added the ability to use a named substring as a condition, using the
2361        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2362        are numbers (not recommended). Forward references are permitted.
2363    
2364    14. Added forward references in named backreferences (if you see what I mean).
2365    
2366    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2367        pattern could run off the end of the subject. For example, the pattern
2368        "(?s)(.{1,5})"8 did this with the subject "ab".
2369    
2370    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2371        PCRE_CASELESS was set when matching characters that were quantified with ?
2372        or *.
2373    
2374    17. A character class other than a single negated character that had a minimum
2375        but no maximum quantifier - for example [ab]{6,} - was not handled
2376        correctly by pce_dfa_exec(). It would match only one character.
2377    
2378    18. A valid (though odd) pattern that looked like a POSIX character
2379        class but used an invalid character after [ (for example [[,abc,]]) caused
2380        pcre_compile() to give the error "Failed: internal error: code overflow" or
2381        in some cases to crash with a glibc free() error. This could even happen if
2382        the pattern terminated after [[ but there just happened to be a sequence of
2383        letters, a binary zero, and a closing ] in the memory that followed.
2384    
2385    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2386        over the years. Originally (before any Unicode support), just the bottom 8
2387        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2388        output from "man perlunicode" includes this:
2389    
2390          The regular expression compiler produces polymorphic opcodes.  That
2391          is, the pattern adapts to the data and automatically switches to
2392          the Unicode character scheme when presented with Unicode data--or
2393          instead uses a traditional byte scheme when presented with byte
2394          data.
2395    
2396        Sadly, a wide octal escape does not cause a switch, and in a string with
2397        no other multibyte characters, these octal escapes are treated as before.
2398        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2399        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2400        Unicode string.
2401    
2402        I have not perpetrated such confusion in PCRE. Up till now, it took just
2403        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2404        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2405        translate to the appropriate multibyte character.
2406    
2407    29. Applied some refactoring to reduce the number of warnings from Microsoft
2408        and Borland compilers. This has included removing the fudge introduced
2409        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2410        a warning about an unused variable.
2411    
2412    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2413        characters since release 4.0, because Perl (from release 5.004) does not.
2414        [Or at least, is documented not to: some releases seem to be in conflict
2415        with the documentation.] However, when a pattern was studied with
2416        pcre_study() and all its branches started with \s, PCRE still included VT
2417        as a possible starting character. Of course, this did no harm; it just
2418        caused an unnecessary match attempt.
2419    
2420    22. Removed a now-redundant internal flag bit that recorded the fact that case
2421        dependency changed within the pattern. This was once needed for "required
2422        byte" processing, but is no longer used. This recovers a now-scarce options
2423        bit. Also moved the least significant internal flag bit to the most-
2424        significant bit of the word, which was not previously used (hangover from
2425        the days when it was an int rather than a uint) to free up another bit for
2426        the future.
2427    
2428    23. Added support for CRLF line endings as well as CR and LF. As well as the
2429        default being selectable at build time, it can now be changed at runtime
2430        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2431        specify that it is scanning data with non-default line endings.
2432    
2433    24. Changed the definition of CXXLINK to make it agree with the definition of
2434        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2435    
2436    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2437        recursions. This makes a big different to stack usage for some patterns.
2438    
2439    26. If a subpattern containing a named recursion or subroutine reference such
2440        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2441        the space required for the compiled pattern went wrong and gave too small a
2442        value. Depending on the environment, this could lead to "Failed: internal
2443        error: code overflow at offset 49" or "glibc detected double free or
2444        corruption" errors.
2445    
2446    27. Applied patches from Google (a) to support the new newline modes and (b) to
2447        advance over multibyte UTF-8 characters in GlobalReplace.
2448    
2449    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2450        difference for some implementation of PCRE in some Windows version.
2451    
2452    29. Added some extra testing facilities to pcretest:
2453    
2454        \q<number>   in a data line sets the "match limit" value
2455        \Q<number>   in a data line sets the "match recursion limt" value
2456        -S <number>  sets the stack size, where <number> is in megabytes
2457    
2458        The -S option isn't available for Windows.
2459    
2460    
2461    Version 6.6 06-Feb-06
2462    ---------------------
2463    
2464     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2465        in pcreposix.h. I have copied the definition from pcre.h.
2466    
2467     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2468        because pcre.h is no longer a built file.
2469    
2470     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2471        not normally included in the compiled code.
2472    
2473    
2474    Version 6.5 01-Feb-06
2475    ---------------------
2476    
2477     1. When using the partial match feature with pcre_dfa_exec(), it was not
2478        anchoring the second and subsequent partial matches at the new starting
2479        point. This could lead to incorrect results. For example, with the pattern
2480        /1234/, partially matching against "123" and then "a4" gave a match.
2481    
2482     2. Changes to pcregrep:
2483    
2484        (a) All non-match returns from pcre_exec() were being treated as failures
2485            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2486            error message is output. Some extra information is given for the
2487            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2488            probably the only errors that are likely to be caused by users (by
2489            specifying a regex that has nested indefinite repeats, for instance).
2490            If there are more than 20 of these errors, pcregrep is abandoned.
2491    
2492        (b) A binary zero was treated as data while matching, but terminated the
2493            output line if it was written out. This has been fixed: binary zeroes
2494            are now no different to any other data bytes.
2495    
2496        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2497            used to set a locale for matching. The --locale=xxxx long option has
2498            been added (no short equivalent) to specify a locale explicitly on the
2499            pcregrep command, overriding the environment variables.
2500    
2501        (d) When -B was used with -n, some line numbers in the output were one less
2502            than they should have been.
2503    
2504        (e) Added the -o (--only-matching) option.
2505    
2506        (f) If -A or -C was used with -c (count only), some lines of context were
2507            accidentally printed for the final match.
2508    
2509        (g) Added the -H (--with-filename) option.
2510    
2511        (h) The combination of options -rh failed to suppress file names for files
2512            that were found from directory arguments.
2513    
2514        (i) Added the -D (--devices) and -d (--directories) options.
2515    
2516        (j) Added the -F (--fixed-strings) option.
2517    
2518        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2519    
2520        (l) Added the --colo(u)r option.
2521    
2522        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2523            is not present by default.
2524    
2525     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2526        items such as (?R) or (?1), when the recursion could match a number of
2527        alternatives. If it matched one of the alternatives, but subsequently,
2528        outside the recursion, there was a failure, the code tried to back up into
2529        the recursion. However, because of the way PCRE is implemented, this is not
2530        possible, and the result was an incorrect result from the match.
2531    
2532        In order to prevent this happening, the specification of recursion has
2533        been changed so that all such subpatterns are automatically treated as
2534        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2535    
2536     4. I had overlooked the fact that, in some locales, there are characters for
2537        which isalpha() is true but neither isupper() nor islower() are true. In
2538        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2539        and ordfeminine) are like this. This affected the treatment of \w and \W
2540        when they appeared in character classes, but not when they appeared outside
2541        a character class. The bit map for "word" characters is now created
2542        separately from the results of isalnum() instead of just taking it from the
2543        upper, lower, and digit maps. (Plus the underscore character, of course.)
2544    
2545     5. The above bug also affected the handling of POSIX character classes such as
2546        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2547        permanent tables. Instead, the bit maps for such a class were previously
2548        created as the appropriate unions of the upper, lower, and digit bitmaps.
2549        Now they are created by subtraction from the [[:word:]] class, which has
2550        its own bitmap.
2551    
2552     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2553        It is created by subtracting the vertical space characters (\x09, \x0a,
2554        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2555        subtraction was done in the overall bitmap for a character class, meaning
2556        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2557        be recognized. This bug has been fixed.
2558    
2559     7. Patches from the folks at Google:
2560    
2561          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2562          real life, but is still worth protecting against".
2563    
2564          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2565          regular expressions".
2566    
2567          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2568          have it.
2569    
2570          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2571          "configure" and the latter not, in order to fix a problem somebody had
2572          with compiling the Arg class on HP-UX.
2573    
2574          (e) Improve the error-handling of the C++ wrapper a little bit.
2575    
2576          (f) New tests for checking recursion limiting.
2577    
2578     8. The pcre_memmove() function, which is used only if the environment does not
2579        have a standard memmove() function (and is therefore rarely compiled),
2580        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2581        returning a result (though PCRE never actually uses the result).
2582    
2583     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2584        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2585        returned instead of calling malloc() with an overflowing number that would
2586        most likely cause subsequent chaos.
2587    
2588    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2589    
2590    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2591        with this option is matched, the nmatch and pmatch options of regexec() are
2592        ignored.
2593    
2594    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2595        provided in case anyone wants to the the POSIX interface with UTF-8
2596        strings.
2597    
2598    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2599        C++ linking (needed for some HP-UX environments).
2600    
2601    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2602        (unused parameter) and in the pcre_printint() function (omitted "default"
2603        switch label when the default is to do nothing).
2604    
2605    15. Added some code to make it possible, when PCRE is compiled as a C++
2606        library, to replace subject pointers for pcre_exec() with a smart pointer
2607        class, thus making it possible to process discontinuous strings.
2608    
2609    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2610        much the same function. They were added by different people who were trying
2611        to make PCRE easy to compile on non-Unix systems. It has been suggested
2612        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2613        for compiling on Windows systems. I have therefore replaced it with
2614        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2615        defaults to "extern" for C or "extern C" for C++, which works fine on
2616        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2617        SCOPE with something explicit in config.h. In addition:
2618    
2619        (a) pcreposix.h still had just "extern" instead of either of these macros;
2620            I have replaced it with PCRE_DATA_SCOPE.
2621    
2622        (b) Functions such as _pcre_xclass(), which are internal to the library,
2623            but external in the C sense, all had PCRE_EXPORT in their definitions.
2624            This is apparently wrong for the Windows case, so I have removed it.
2625            (It makes no difference on Unix-like systems.)
2626    
2627    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2628        of recursive calls to match(). This is different to MATCH_LIMIT because
2629        that limits the total number of calls to match(), not all of which increase
2630        the depth of recursion. Limiting the recursion depth limits the amount of
2631        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2632        when PCRE is compiled, and changed at run time. A patch from Google adds
2633        this functionality to the C++ interface.
2634    
2635    18. Changes to the handling of Unicode character properties:
2636    
2637        (a) Updated the table to Unicode 4.1.0.
2638    
2639        (b) Recognize characters that are not in the table as "Cn" (undefined).
2640    
2641        (c) I revised the way the table is implemented to a much improved format
2642            which includes recognition of ranges. It now supports the ranges that
2643            are defined in UnicodeData.txt, and it also amalgamates other
2644            characters into ranges. This has reduced the number of entries in the
2645            table from around 16,000 to around 3,000, thus reducing its size
2646            considerably. I realized I did not need to use a tree structure after
2647            all - a binary chop search is just as efficient. Having reduced the
2648            number of entries, I extended their size from 6 bytes to 8 bytes to
2649            allow for more data.
2650    
2651        (d) Added support for Unicode script names via properties such as \p{Han}.
2652    
2653    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2654        matching that character.
2655    
2656    20. When matching a repeated Unicode property with a minimum greater than zero,
2657        (for example \pL{2,}), PCRE could look past the end of the subject if it
2658        reached it while seeking the minimum number of characters. This could
2659        happen only if some of the characters were more than one byte long, because
2660        there is a check for at least the minimum number of bytes.
2661    
2662    21. Refactored the implementation of \p and \P so as to be more general, to
2663        allow for more different types of property in future. This has changed the
2664        compiled form incompatibly. Anybody with saved compiled patterns that use
2665        \p or \P will have to recompile them.
2666    
2667    22. Added "Any" and "L&" to the supported property types.
2668    
2669    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2670        but give a compile time error if the value is greater than 0xff.
2671    
2672    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2673        accidentally not being installed or uninstalled.
2674    
2675    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2676        made were to insert the current release number. This seemed silly, because
2677        it made things harder for people building PCRE on systems that don't run
2678        "configure". I have turned pcre.h into a distributed file, no longer built
2679        by "configure", with the version identification directly included. There is
2680        no longer a pcre.h.in file.
2681    
2682        However, this change necessitated a change to the pcre-config script as
2683        well. It is built from pcre-config.in, and one of the substitutions was the
2684        release number. I have updated configure.ac so that ./configure now finds
2685        the release number by grepping pcre.h.
2686    
2687    26. Added the ability to run the tests under valgrind.
2688    
2689    
2690    Version 6.4 05-Sep-05
2691    ---------------------
2692    
2693     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2694        "--" to be printed when multiple files were scanned, even when none of the
2695        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2696        consider it to be a bug, and have restored the previous behaviour.
2697    
2698     2. A couple of code tidies to get rid of compiler warnings.
2699    
2700     3. The pcretest program used to cheat by referring to symbols in the library
2701        whose names begin with _pcre_. These are internal symbols that are not
2702        really supposed to be visible externally, and in some environments it is
2703        possible to suppress them. The cheating is now confined to including
2704        certain files from the library's source, which is a bit cleaner.
2705    
2706     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2707        file's purpose clearer.
2708    
2709     5. Reorganized pcre_ucp_findchar().
2710    
2711    
2712    Version 6.3 15-Aug-05
2713    ---------------------
2714    
2715     1. The file libpcre.pc.in did not have general read permission in the tarball.
2716    
2717     2. There were some problems when building without C++ support:
2718    
2719        (a) If C++ support was not built, "make install" and "make test" still
2720            tried to test it.
2721    
2722        (b) There were problems when the value of CXX was explicitly set. Some
2723            changes have been made to try to fix these, and ...
2724    
2725        (c) --disable-cpp can now be used to explicitly disable C++ support.
2726    
2727        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2728            backslash in a target when C++ was disabled. This confuses some
2729            versions of "make", apparently. Using an intermediate variable solves
2730            this. (Same for CPP_LOBJ.)
2731    
2732     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2733        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2734        necessary on certain architectures.
2735    
2736     4. Added a setting of -export-symbols-regex to the link command to remove
2737        those symbols that are exported in the C sense, but actually are local
2738        within the library, and not documented. Their names all begin with
2739        "_pcre_". This is not a perfect job, because (a) we have to except some
2740        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2741        available (and never for static libraries). I have made a note to try to
2742        find a way round (a) in the future.
2743    
2744    
2745    Version 6.2 01-Aug-05
2746    ---------------------
2747    
2748     1. There was no test for integer overflow of quantifier values. A construction
2749        such as {1111111111111111} would give undefined results. What is worse, if
2750        a minimum quantifier for a parenthesized subpattern overflowed and became
2751        negative, the calculation of the memory size went wrong. This could have
2752        led to memory overwriting.
2753    
2754     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2755    
2756     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2757        operating environments where this matters.
2758    
2759     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2760        PCRE options from within the C++ wrapper.
2761    
2762     5. Named capturing subpatterns were not being correctly counted when a pattern
2763        was compiled. This caused two problems: (a) If there were more than 100
2764        such subpatterns, the calculation of the memory needed for the whole
2765        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2766        back references of the form \12, where the number was greater than 9, were
2767        not recognized as back references, even though there were sufficient
2768        previous subpatterns.
2769    
2770     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2771        versions of gcc, e.g. 2.95.4.
2772    
2773    
2774    Version 6.1 21-Jun-05
2775    ---------------------
2776    
2777     1. There was one reference to the variable "posix" in pcretest.c that was not
2778        surrounded by "#if !defined NOPOSIX".
2779    
2780     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2781        the cross-check on the old pcre_info() function, for the benefit of the
2782        cut-down version of PCRE that is currently imported into Exim.
2783    
2784     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2785        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2786        patterns that start (?i) but I don't think that matters. The use of (?i) is
2787        just an example; this all applies to the other options as well.
2788    
2789     4. Since libtool seems to echo the compile commands it is issuing, the output
2790        from "make" can be reduced a bit by putting "@" in front of each libtool
2791        compile command.
2792    
2793     5. Patch from the folks at Google for configure.in to be a bit more thorough
2794        in checking for a suitable C++ installation before trying to compile the
2795        C++ stuff. This should fix a reported problem when a compiler was present,
2796        but no suitable headers.
2797    
2798     6. The man pages all had just "PCRE" as their title. I have changed them to
2799        be the relevant file name. I have also arranged that these names are
2800        retained in the file doc/pcre.txt, which is a concatenation in text format
2801        of all the man pages except the little individual ones for each function.
2802    
2803     7. The NON-UNIX-USE file had not been updated for the different set of source
2804        files that come with release 6. I also added a few comments about the C++
2805        wrapper.
2806    
2807    
2808    Version 6.0 07-Jun-05
2809    ---------------------
2810    
2811     1. Some minor internal re-organization to help with my DFA experiments.
2812    
2813     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2814        didn't matter for the library itself when fully configured, but did matter
2815        when compiling without UCP support, or within Exim, where the ucp files are
2816        not imported.
2817    
2818     3. Refactoring of the library code to split up the various functions into
2819        different source modules. The addition of the new DFA matching code (see
2820        below) to a single monolithic source would have made it really too
2821        unwieldy, quite apart from causing all the code to be include in a
2822        statically linked application, when only some functions are used. This is
2823        relevant even without the DFA addition now that patterns can be compiled in
2824        one application and matched in another.
2825    
2826        The downside of splitting up is that there have to be some external
2827        functions and data tables that are used internally in different modules of
2828        the library but which are not part of the API. These have all had their
2829        names changed to start with "_pcre_" so that they are unlikely to clash
2830        with other external names.
2831    
2832     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2833        a different (DFA) algorithm. Although it is slower than the original
2834        function, it does have some advantages for certain types of matching
2835        problem.
2836    
2837     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2838        including restarting after a partial match.
2839    
2840     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2841        defined when compiling for Windows was sent to me. I have put it into the
2842        code, though I have no means of testing or verifying it.
2843    
2844     7. Added the pcre_refcount() auxiliary function.
2845    
2846     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2847        match before or at the first newline in the subject string. In pcretest,
2848        the /f option on a pattern can be used to set this.
2849    
2850     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2851        would behave wrongly. This has been present in PCRE since release 4.0.
2852    
2853    10. A number of changes to the pcregrep command:
2854    
2855        (a) Refactored how -x works; insert ^(...)$ instead of setting
2856            PCRE_ANCHORED and checking the length, in preparation for adding
2857            something similar for -w.
2858    
2859        (b) Added the -w (match as a word) option.
2860    
2861        (c) Refactored the way lines are read and buffered so as to have more
2862            than one at a time available.
2863    
2864        (d) Implemented a pcregrep test script.
2865    
2866        (e) Added the -M (multiline match) option. This allows patterns to match
2867            over several lines of the subject. The buffering ensures that at least
2868            8K, or the rest of the document (whichever is the shorter) is available
2869            for matching (and similarly the previous 8K for lookbehind assertions).
2870    
2871        (f) Changed the --help output so that it now says
2872    
2873              -w, --word-regex(p)
2874    
2875            instead of two lines, one with "regex" and the other with "regexp"
2876            because that confused at least one person since the short forms are the
2877            same. (This required a bit of code, as the output is generated
2878            automatically from a table. It wasn't just a text change.)
2879    
2880        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2881            option but starts with a hyphen. Could be a pattern or a path name
2882            starting with a hyphen, for instance.
2883    
2884        (h) "-" can be given as a file name to represent stdin.
2885    
2886        (i) When file names are being printed, "(standard input)" is used for
2887            the standard input, for compatibility with GNU grep. Previously
2888            "<stdin>" was used.
2889    
2890        (j) The option --label=xxx can be used to supply a name to be used for
2891            stdin when file names are being printed. There is no short form.
2892    
2893        (k) Re-factored the options decoding logic because we are going to add
2894            two more options that take data. Such options can now be given in four
2895            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2896    
2897        (l) Added the -A, -B, and -C options for requesting that lines of context
2898            around matches be printed.
2899    
2900        (m) Added the -L option to print the names of files that do not contain
2901            any matching lines, that is, the complement of -l.
2902    
2903        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2904            continue to scan other files.
2905    
2906        (o) The -s option was incorrectly implemented. For compatibility with other
2907            greps, it now suppresses the error message for a non-existent or non-
2908            accessible file (but not the return code). There is a new option called
2909            -q that suppresses the output of matching lines, which was what -s was
2910            previously doing.
2911    
2912        (p) Added --include and --exclude options to specify files for inclusion
2913            and exclusion when recursing.
2914    
2915    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2916        Hopefully, it now does.
2917    
2918    12. Missing cast in pcre_study().
2919    
2920    13. Added an "uninstall" target to the makefile.
2921    
2922    14. Replaced "extern" in the function prototypes in Makefile.in with
2923        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2924        world, but is set differently for Windows.
2925    
2926    15. Added a second compiling function called pcre_compile2(). The only
2927        difference is that it has an extra argument, which is a pointer to an
2928        integer error code. When there is a compile-time failure, this is set
2929        non-zero, in addition to the error test pointer being set to point to an
2930        error message. The new argument may be NULL if no error number is required
2931        (but then you may as well call pcre_compile(), which is now just a
2932        wrapper). This facility is provided because some applications need a
2933        numeric error indication, but it has also enabled me to tidy up the way
2934        compile-time errors are handled in the POSIX wrapper.
2935    
2936    16. Added VPATH=.libs to the makefile; this should help when building with one
2937        prefix path and installing with another. (Or so I'm told by someone who
2938        knows more about this stuff than I do.)
2939    
2940    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2941        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2942        match everything, including newlines. This is not POSIX-compatible, but
2943        somebody wanted the feature. From pcretest it can be activated by using
2944        both the P and the s flags.
2945    
2946    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2947    
2948    19. libpcre.pc was being incorrectly installed as executable.
2949    
2950    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2951        it now also looks for '\r' so that it will work unmodified on Windows.
2952    
2953    21. Added Google's contributed C++ wrapper to the distribution.
2954    
2955    22. Added some untidy missing memory free() calls in pcretest, to keep
2956        Electric Fence happy when testing.
2957    
2958    
2959    
2960    Version 5.0 13-Sep-04
2961    ---------------------
2962    
2963     1. Internal change: literal characters are no longer packed up into items
2964        containing multiple characters in a single byte-string. Each character
2965        is now matched using a separate opcode. However, there may be more than one
2966        byte in the character in UTF-8 mode.
2967    
2968     2. The pcre_callout_block structure has two new fields: pattern_position and
2969        next_item_length. These contain the offset in the pattern to the next match
2970        item, and its length, respectively.
2971    
2972     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2973        insertion of callouts before each pattern item. Added the /C option to
2974        pcretest to make use of this.
2975    
2976     4. On the advice of a Windows user, the lines
2977    
2978          #if defined(_WIN32) || defined(WIN32)
2979          _setmode( _fileno( stdout ), 0x8000 );
2980          #endif  /* defined(_WIN32) || defined(WIN32) */
2981    
2982        have been added to the source of pcretest. This apparently does useful
2983        magic in relation to line terminators.
2984    
2985     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2986        for the benefit of those environments where the "b" makes a difference.
2987    
2988     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2989        to know about it. I have put a hack into configure.in that adds in code
2990        to set GCC=yes if CC=icc. This seems to end up at a point in the
2991        generated configure script that is early enough to affect the setting of
2992        compiler options, which is what is needed, but I have no means of testing
2993        whether it really works. (The user who reported this had patched the
2994        generated configure script, which of course I cannot do.)
2995    
2996        LATER: After change 22 below (new libtool files), the configure script
2997        seems to know about icc (and also ecc). Therefore, I have commented out
2998        this hack in configure.in.
2999    
3000     7. Added support for pkg-config (2 patches were sent in).
3001    
3002     8. Negated POSIX character classes that used a combination of internal tables
3003        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3004        [[:^ascii]]. Typically, they would match almost any characters. The other
3005        POSIX classes were not broken in this way.
3006    
3007     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3008        to find the match, as PCRE was deluded into thinking that the match had to
3009        start at the start point or following a newline. The same bug applied to
3010        patterns with negative forward assertions or any backward assertions
3011        preceding ".*" at the start, unless the pattern required a fixed first
3012        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3013    
3014    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3015        starting at the last subject character, bytes beyond the end of the subject
3016        string were read.
3017    
3018    11. Renamed the variable "class" as "classbits" to make life easier for C++
3019        users. (Previously there was a macro definition, but it apparently wasn't
3020        enough.)
3021    
3022    12. Added the new field "tables" to the extra data so that tables can be passed
3023        in at exec time, or the internal tables can be re-selected. This allows
3024        a compiled regex to be saved and re-used at a later time by a different
3025        program that might have everything at different addresses.
3026    
3027    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3028        -R library as well as a -L library.
3029    
3030    14. The debugging options of pcretest (-d on the command line or D on a
3031        pattern) showed incorrect output for anything following an extended class
3032        that contained multibyte characters and which was followed by a quantifier.
3033    
3034    15. Added optional support for general category Unicode character properties
3035        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3036        support. It adds about 90K to the size of the library. The meanings of the
3037        inbuilt class escapes such as \d and \s have NOT been changed.
3038    
3039    16. Updated pcredemo.c to include calls to free() to release the memory for the
3040        compiled pattern.
3041    
3042    17. The generated file chartables.c was being created in the source directory
3043        instead of in the building directory. This caused the build to fail if the
3044        source directory was different from the building directory, and was
3045        read-only.
3046    
3047    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3048        file. No doubt somebody will tell me if they don't make sense... Also added
3049        Dan Mooney's comments about building on OpenVMS.
3050    
3051    19. Added support for partial matching via the PCRE_PARTIAL option for
3052        pcre_exec() and the \P data escape in pcretest.
3053    
3054    20. Extended pcretest with 3 new pattern features:
3055    
3056        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3057              write the compiled pattern to the file whose name is "rest-of-line".
3058              This is a straight binary dump of the data, with the saved pointer to
3059              the character tables forced to be NULL. The study data, if any, is
3060              written too. After writing, pcretest reads a new pattern.
3061    
3062        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3063              compiled pattern from the given file. There must not be any
3064              occurrences of "<" in the file name (pretty unlikely); if there are,
3065              pcretest will instead treat the initial "<" as a pattern delimiter.
3066              After reading in the pattern, pcretest goes on to read data lines as
3067              usual.
3068    
3069        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3070              and 16-bit fields in a compiled pattern, to simulate a pattern that
3071              was compiled on a host of opposite endianness.
3072    
3073    21. The pcre-exec() function can now cope with patterns that were compiled on
3074        hosts of opposite endianness, with this restriction:
3075    
3076          As for any compiled expression that is saved and used later, the tables
3077          pointer field cannot be preserved; the extra_data field in the arguments
3078          to pcre_exec() should be used to pass in a tables address if a value
3079          other than the default internal tables were used at compile time.
3080    
3081    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3082        now diagnosed as an error. Previously, most of the time, a negative number
3083        would have been treated as zero, but if in addition "ovector" was passed as
3084        NULL, a crash could occur.
3085    
3086    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3087        new versions from the libtool 1.5 distribution (the last one is a copy of
3088        a file called libtool.m4). This seems to have fixed the need to patch
3089        "configure" to support Darwin 1.3 (which I used to do). However, I still
3090        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3091        workstation).
3092    
3093    24. Changed the PCRE licence to be the more standard "BSD" licence.
3094    
3095    
3096    Version 4.5 01-Dec-03
3097    ---------------------
3098    
3099     1. There has been some re-arrangement of the code for the match() function so
3100        that it can be compiled in a version that does not call itself recursively.
3101        Instead, it keeps those local variables that need separate instances for
3102        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3103        needs to "recurse". Keeping track of where control must go is done by means
3104        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3105        hide most of the details from the main code, and operates only if
3106        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3107        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3108        operating.
3109    
3110        To make it easier for callers to provide specially tailored get/free
3111        functions for this usage, two new functions, pcre_stack_malloc, and
3112        pcre_stack_free, are used. They are always called in strict stacking order,
3113        and the size of block requested is always the same.
3114    
3115        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3116        PCRE has been compiled to use the stack or the heap for recursion. The
3117        -C option of pcretest uses this to show which version is compiled.
3118    
3119        A new data escape \S, is added to pcretest; it causes the amounts of store
3120        obtained and freed by both kinds of malloc/free at match time to be added
3121        to the output.
3122    
3123     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3124        what's available on my current Linux desktop machine.
3125    
3126     3. When matching a UTF-8 string, the test for a valid string at the start has
3127        been extended. If start_offset is not zero, PCRE now checks that it points
3128        to a byte that is the start of a UTF-8 character. If not, it returns
3129        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3130        this is necessary because there may be backward assertions in the pattern.
3131        When matching the same subject several times, it may save resources to use
3132        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3133    
3134     4. The code for checking the validity of UTF-8 strings has been tightened so
3135        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3136        containing "overlong sequences".
3137    
3138     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3139        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3140        should have been "&", but it just so happened that all the cases this let
3141        through by mistake were picked up later in the function.
3142    
3143     6. I had used a variable called "isblank" - this is a C99 function, causing
3144        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3145    
3146     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3147        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3148        so the version is shown; (c) stop "make test" echoing "./RunTest".
3149    
3150     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3151    
3152     9. The prototype for memmove() for systems that don't have it was using
3153        size_t, but the inclusion of the header that defines size_t was later. I've
3154        moved the #includes for the C headers earlier to avoid this.
3155    
3156    10. Added some adjustments to the code to make it easier to compiler on certain
3157        special systems:
3158    
3159          (a) Some "const" qualifiers were missing.
3160          (b) Added the macro EXPORT before all exported functions; by default this
3161              is defined to be empty.
3162          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3163              that it reads its output file name as an argument instead of writing
3164              to the standard output and assuming this can be redirected.
3165    
3166    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3167        class containing characters with values greater than 255, PCRE compilation
3168        went into a loop.
3169    
3170    12. A recursive reference to a subpattern that was within another subpattern
3171        that had a minimum quantifier of zero caused PCRE to crash. For example,
3172        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3173        recursion. If the recursively-called subpattern itself had a zero repeat,
3174        that was OK.
3175    
3176    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3177        buffer into which it was copied (for escape processing) was still set at
3178        1024, so long lines caused crashes.
3179    
3180    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3181        "internal error: code overflow...". This applied to any character class
3182        that was followed by a possessive quantifier.
3183    
3184    15. Modified the Makefile to add libpcre.la as a prerequisite for
3185        libpcreposix.la because I was told this is needed for a parallel build to
3186        work.
3187    
3188    16. If a pattern that contained .* following optional items at the start was
3189        studied, the wrong optimizing data was generated, leading to matching
3190        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3191        matching string must start with a or b or c. The correct conclusion for
3192        this pattern is that a match can start with any character.
3193    
3194    
3195    Version 4.4 13-Aug-03
3196    ---------------------
3197    
3198     1. In UTF-8 mode, a character class containing characters with values between
3199        127 and 255 was not handled correctly if the compiled pattern was studied.
3200        In fixing this, I have also improved the studying algorithm for such
3201        classes (slightly).
3202    
3203     2. Three internal functions had redundant arguments passed to them. Removal
3204        might give a very teeny performance improvement.
3205    
3206     3. Documentation bug: the value of the capture_top field in a callout is *one
3207        more than* the number of the hightest numbered captured substring.
3208    
3209     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3210        in incorrectly linking with a previously installed version. They now link
3211        explicitly with libpcre.la.
3212    
3213     5. configure.in no longer needs to recognize Cygwin specially.
3214    
3215     6. A problem in pcre.in for Windows platforms is fixed.
3216    
3217     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3218        pcretest, it used to include the size of the study block as part of its
3219        output. Unfortunately, the structure contains a field that has a different
3220        size on different hardware architectures. This meant that the tests that
3221        showed this size failed. As the block is currently always of a fixed size,
3222        this information isn't actually particularly useful in pcretest output, so
3223        I have just removed it.
3224    
3225     8. Three pre-processor statements accidentally did not start in column 1.
3226        Sadly, there are *still* compilers around that complain, even though
3227        standard C has not required this for well over a decade. Sigh.
3228    
3229     9. In pcretest, the code for checking callouts passed small integers in the
3230        callout_data field, which is a void * field. However, some picky compilers
3231        complained about the casts involved for this on 64-bit systems. Now
3232        pcretest passes the address of the small integer instead, which should get
3233        rid of the warnings.
3234    
3235    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3236        both compile and run time, and gives an error if an invalid UTF-8 sequence
3237        is found. There is a option for disabling this check in cases where the
3238        string is known to be correct and/or the maximum performance is wanted.
3239    
3240    11. In response to a bug report, I changed one line in Makefile.in from
3241    
3242            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3243        to
3244            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3245    
3246        to look similar to other lines, but I have no way of telling whether this
3247        is the right thing to do, as I do not use Windows. No doubt I'll get told
3248        if it's wrong...
3249    
3250    
3251    Version 4.3 21-May-03
3252    ---------------------
3253    
3254    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3255       Makefile.
3256    
3257    2. Some refactoring to improve the quality of the code:
3258    
3259       (i)   The utf8_table... variables are now declared "const".
3260    
3261       (ii)  The code for \cx, which used the "case flipping" table to upper case
3262             lower case letters, now just substracts 32. This is ASCII-specific,
3263             but the whole concept of \cx is ASCII-specific, so it seems
3264             reasonable.
3265    
3266       (iii) PCRE was using its character types table to recognize decimal and
3267             hexadecimal digits in the pattern. This is silly, because it handles
3268             only 0-9, a-f, and A-F, but the character types table is locale-
3269             specific, which means strange things might happen. A private
3270             table is now used for this - though it costs 256 bytes, a table is
3271             much faster than multiple explicit tests. Of course, the standard
3272             character types table is still used for matching digits in subject
3273             strings against \d.
3274    
3275       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3276             ending in _t are). So I've renamed it as ESC_tee.
3277    
3278    3. The first argument for regexec() in the POSIX wrapper should have been
3279       defined as "const".
3280    
3281    4. Changed pcretest to use malloc() for its buffers so that they can be
3282       Electric Fenced for debugging.
3283    
3284    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3285       to read one or more bytes before the start of the subject string. Often this
3286       had no effect on PCRE's behaviour, but in some circumstances it could
3287       provoke a segmentation fault.
3288    
3289    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3290       to try to read one or more bytes before the start of the subject string.
3291    
3292    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3293       UTF-8 support could misbehave in various ways if the subject string
3294       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3295       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3296       back over UTF-8 characters.)
3297    
3298    
3299    Version 4.2 14-Apr-03
3300    ---------------------
3301    
3302    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3303    
3304    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3305         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3306         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3307         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3308         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3309           and BUILD_EXEEXT
3310         Note: automatic setting of the BUILD variables is not yet working
3311         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3312           compile-time but not at link-time
3313         [LINK]: use for linking executables only
3314         make different versions for Windows and non-Windows
3315         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3316           libraries
3317         [LINK_FOR_BUILD]: new variable
3318         [OBJEXT]: use throughout
3319         [EXEEXT]: use throughout
3320         <winshared>: new target
3321         <wininstall>: new target
3322         <dftables.o>: use native compiler
3323         <dftables>: use native linker
3324         <install>: handle Windows platform correctly
3325         <clean>: ditto
3326         <check>: ditto
3327         copy DLL to top builddir before testing
3328    
3329       As part of these changes, -no-undefined was removed again. This was reported
3330       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3331       in any case.
3332    
3333    3. Some tidies to get rid of compiler warnings:
3334    
3335       . In the match_data structure, match_limit was an unsigned long int, whereas
3336         match_call_count was an int. I've made them both unsigned long ints.
3337    
3338       . In pcretest the fact that a const uschar * doesn't automatically cast to
3339         a void * provoked a warning.
3340    
3341       . Turning on some more compiler warnings threw up some "shadow" variables
3342         and a few more missing casts.
3343    
3344    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3345       option, a class that contained a single character with a value between 128
3346       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3347    
3348    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3349       option, a class that contained several characters, but with at least one
3350       whose value was between 128 and 255 caused PCRE to crash.
3351    
3352    
3353    Version 4.1 12-Mar-03
3354    ---------------------
3355    
3356    1. Compiling with gcc -pedantic found a couple of places where casts were
3357    needed, and a string in dftables.c that was longer than standard compilers are
3358    required to support.
3359    
3360    2. Compiling with Sun's compiler found a few more places where the code could
3361    be tidied up in order to avoid warnings.
3362    
3363    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3364    first of these names is deprecated in the latest Autoconf in favour of the name
3365    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3366    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3367    analogy I have changed it to CFLAGS_FOR_BUILD.
3368    
3369    4. Added -no-undefined to the linking command in the Makefile, because this is
3370    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3371    linking step for the pcreposix library.
3372    
3373    5. PCRE was failing to diagnose the case of two named groups with the same
3374    name.
3375    
3376    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3377    literal character that is needed in the subject for a match, and scans along to
3378    ensure that it is present before embarking on the full matching process. This
3379    saves time in cases of nested unlimited repeats that are never going to match.
3380    Problem: the scan can take a lot of time if the subject is very long (e.g.
3381    megabytes), thus penalizing straightforward matches. It is now done only if the
3382    amount of subject to be scanned is less than 1000 bytes.
3383    
3384    7. A lesser problem with the same optimization is that it was recording the
3385    first character of an anchored pattern as "needed", thus provoking a search
3386    right along the subject, even when the first match of the pattern was going to
3387    fail. The "needed" character is now not set for anchored patterns, unless it
3388    follows something in the pattern that is of non-fixed length. Thus, it still
3389    fulfils its original purpose of finding quick non-matches in cases of nested
3390    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3391    
3392    
3393    Version 4.0 17-Feb-03
3394    ---------------------
3395    
3396    1. If a comment in an extended regex that started immediately after a meta-item
3397    extended to the end of string, PCRE compiled incorrect data. This could lead to
3398    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3399    
3400    2. Moved to autoconf 2.53 and libtool 1.4.2.
3401    
3402    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3403    the special perltest8 script is no longer needed - all the tests can be run
3404    from a single perltest script.
3405    
3406    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3407    by \s. It has now been removed in PCRE. This means it isn't recognized as
3408    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3409    class [:space:] *does* include VT, thereby creating a mess.
3410    
3411    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3412    space and tab.
3413    
3414    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3415    its new features into the main test script, reducing the number of scripts.
3416    
3417    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3418    were backward compatible, and made the (?i) apply to the whole pattern, as if
3419    /i were given. Now it behaves more logically, and applies the option setting
3420    only to what follows. PCRE has been changed to follow suit. However, if it
3421    finds options settings right at the start of the pattern, it extracts them into
3422    the global options, as before. Thus, they show up in the info data.
3423    
3424    8. Added support for the \Q...\E escape sequence. Characters in between are
3425    treated as literals. This is slightly different from Perl in that $ and @ are
3426    also handled as literals inside the quotes. In Perl, they will cause variable
3427    interpolation. Note the following examples:
3428    
3429        Pattern            PCRE matches      Perl matches
3430    
3431        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3432        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3433        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3434    
3435    For compatibility with Perl, \Q...\E sequences are recognized inside character
3436    classes as well as outside them.
3437    
3438    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3439    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3440    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3441    signed/unsigned warnings.
3442    
3443    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3444    option for pcretest, so I've replaced it by a simple function that does just
3445    that job.
3446    
3447    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3448    "pcregrep -".
3449    
3450    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3451    Java package. This provides some syntactic sugar for simple cases of what my
3452    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3453    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3454    item, you can use this simplified notation. Note that only makes sense with
3455    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3456    greediness, whatever the setting of the PCRE_UNGREEDY option.
3457    
3458    13. A change of greediness default within a pattern was not taking effect at
3459    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3460    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3461    was abstracted outside.
3462    
3463    14. PCRE now supports the \G assertion. It is true when the current matching
3464    position is at the start point of the match. This differs from \A when the
3465    starting offset is non-zero. Used with the /g option of pcretest (or similar
3466    code), it works in the same way as it does for Perl's /g option. If all
3467    alternatives of a regex begin with \G, the expression is anchored to the start
3468    match position, and the "anchored" flag is set in the compiled expression.
3469    
3470    15. Some bugs concerning the handling of certain option changes within patterns
3471    have been fixed. These applied to options other than (?ims). For example,
3472    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3473    been the other way round. Some of this was related to change 7 above.
3474    
3475    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3476    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3477    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3478    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3479    
3480    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3481    mode. Unlike ".", it always matches newline, whatever the setting of
3482    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3483    assertions. Perl allows it, but it doesn't (in general) work because it can't
3484    calculate the length of the lookbehind. At least, that's the case for Perl
3485    5.8.0 - I've been told they are going to document that it doesn't work in
3486    future.
3487    
3488    18. Added an error diagnosis for escapes that PCRE does not support: these are
3489    \L, \l, \N, \P, \p, \U, \u, and \X.
3490    
3491    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3492    reading past the end of the pattern in cases such as /[abcd/.
3493    
3494    20. PCRE was getting more memory than necessary for patterns with classes that
3495    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3496    
3497    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3498    compiling PCRE for use with Virtual Pascal.
3499    
3500    22. Small fix to the Makefile to make it work properly if the build is done
3501    outside the source tree.
3502    
3503    23. Added a new extension: a condition to go with recursion. If a conditional
3504    subpattern starts with (?(R) the "true" branch is used if recursion has
3505    happened, whereas the "false" branch is used only at the top level.
3506    
3507    24. When there was a very long string of literal characters (over 255 bytes
3508    without UTF support, over 250 bytes with UTF support), the computation of how
3509    much memory was required could be incorrect, leading to segfaults or other
3510    strange effects.
3511    
3512    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3513    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3514    there was a subsequent back reference to those brackets. This meant that, for
3515    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3516    possible to check for precisely this case. All we can do is abandon the
3517    optimization if .* occurs inside capturing brackets when there are any back
3518    references whatsoever. (See below for a better fix that came later.)
3519    
3520    26. The handling of the optimization for finding the first character of a
3521    non-anchored pattern, and for finding a character that is required later in the
3522    match were failing in some cases. This didn't break the matching; it just
3523    failed to optimize when it could. The way this is done has been re-implemented.
3524    
3525    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3526    
3527    28. Added a new feature that provides some of the functionality that Perl
3528    provides with (?{...}). The facility is termed a "callout". The way it is done
3529    in PCRE is for the caller to provide an optional function, by setting
3530    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3531    global variable. By default it is unset, which disables all calling out. To get
3532    the function called, the regex must include (?C) at appropriate points. This
3533    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3534    This provides a means of identifying different callout points. When PCRE
3535    reaches such a point in the regex, if pcre_callout has been set, the external
3536    function is called. It is provided with data in a structure called
3537    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3538    matching continues; if it returns a non-zero value, the match at the current
3539    point fails. However, backtracking will occur if possible. [This was changed
3540    later and other features added - see item 49 below.]
3541    
3542    29. pcretest is upgraded to test the callout functionality. It provides a
3543    callout function that displays information. By default, it shows the start of
3544    the match and the current position in the text. There are some new data escapes
3545    to vary what happens:
3546    
3547        \C+         in addition, show current contents of captured substrings
3548        \C-         do not supply a callout function
3549        \C!n        return 1 when callout number n is reached
3550        \C!n!m      return 1 when callout number n is reached for the mth time
3551    
3552    30. If pcregrep was called with the -l option and just a single file name, it
3553    output "<stdin>" if a match was found, instead of the file name.
3554    
3555    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3556    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3557    pcre_exec(). This saves a malloc/free per call. The default value of
3558    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3559    when configuring.
3560    
3561    32. The default maximum size of a compiled pattern is 64K. There have been a
3562    few cases of people hitting this limit. The code now uses macros to handle the
3563    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3564    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3565    configuring. Tests 2 and 5 work only with 2-byte links because they output
3566    debugging information about compiled patterns.
3567    
3568    33. Internal code re-arrangements:
3569    
3570    (a) Moved the debugging function for printing out a compiled regex into
3571        its own source file (printint.c) and used #include to pull it into
3572        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3573        separate copies.
3574    
3575    (b) Defined the list of op-code names for debugging as a macro in
3576        internal.h so that it is next to the definition of the opcodes.
3577    
3578    (c) Defined a table of op-code lengths for simpler skipping along compiled
3579        code. This is again a macro in internal.h so that it is next to the
3580        definition of the opcodes.
3581    
3582    34. Added support for recursive calls to individual subpatterns, along the
3583    lines of Robin Houston's patch (but implemented somewhat differently).
3584    
3585    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3586    allow it to read and process whole directories in Win32. This code was
3587    contributed by Lionel Fourquaux; it has not been tested by me.
3588    
3589    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3590    used to name a group. Names consist of alphanumerics and underscores, and must
3591    be unique. Back references use the syntax (?P=name) and recursive calls use
3592    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3593    numbers. The function pcre_fullinfo() can be used after compilation to extract
3594    a name/number map. There are three relevant calls:
3595    
3596      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3597      PCRE_INFO_NAMECOUNT            yields the number of entries
3598      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3599    
3600    The map is a vector of fixed-size entries. The size of each entry depends on
3601    the length of the longest name used. The first two bytes of each entry are the
3602    group number, most significant byte first. There follows the corresponding
3603    name, zero terminated. The names are in alphabetical order.
3604    
3605    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3606    case instead of 255. Making it the same both with and without UTF-8 support
3607    means that the same test output works with both.
3608    
3609    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3610    calling malloc() with a zero argument.
3611    
3612    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3613    optimization. I've improved things by keeping a bitmap of backreferences with
3614    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3615    fact referenced, the optimization can be applied. It is unlikely that a
3616    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3617    the match to follow \n) will appear inside brackets with a number greater than
3618    31, but if it does, any back reference > 31 suppresses the optimization.
3619    
3620    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3621    of disabling numbered capturing parentheses. Any opening parenthesis that is
3622    not followed by ? behaves as if it were followed by ?: but named parentheses
3623    can still be used for capturing (and they will acquire numbers in the usual
3624    way).
3625    
3626    41. Redesigned the return codes from the match() function into yes/no/error so
3627    that errors can be passed back from deep inside the nested calls. A malloc
3628    failure while inside a recursive subpattern call now causes the
3629    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3630    
3631    42. It is now possible to set a limit on the number of times the match()
3632    function is called in a call to pcre_exec(). This facility makes it possible to
3633    limit the amount of recursion and backtracking, though not in a directly
3634    obvious way, because the match() function is used in a number of different
3635    circumstances. The count starts from zero for each position in the subject
3636    string (for non-anchored patterns). The default limit is, for compatibility, a
3637    large number, namely 10 000 000. You can change this in two ways:
3638    
3639    (a) When configuring PCRE before making, you can use --with-match-limit=n
3640        to set a default value for the compiled library.
3641    
3642    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3643        a different value is set. See 45 below.
3644    
3645    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3646    
3647    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3648    of things that can be changed at compile time. The first argument specifies
3649    what is wanted and the second points to where the information is to be placed.
3650    The current list of available information is:
3651    
3652      PCRE_CONFIG_UTF8
3653    
3654    The output is an integer that is set to one if UTF-8 support is available;
3655    otherwise it is set to zero.
3656    
3657      PCRE_CONFIG_NEWLINE
3658    
3659    The output is an integer that it set to the value of the code that is used for
3660    newline. It is either LF (10) or CR (13).
3661    
3662      PCRE_CONFIG_LINK_SIZE
3663    
3664    The output is an integer that contains the number of bytes used for internal
3665    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3666    
3667      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3668    
3669    The output is an integer that contains the threshold above which the POSIX
3670    interface uses malloc() for output vectors. See item 31 above.
3671    
3672      PCRE_CONFIG_MATCH_LIMIT
3673    
3674    The output is an unsigned integer that contains the default limit of the number
3675    of match() calls in a pcre_exec() execution. See 42 above.
3676    
3677    44. pcretest has been upgraded by the addition of the -C option. This causes it
3678    to extract all the available output from the new pcre_config() function, and to
3679    output it. The program then exits immediately.
3680    
3681    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3682    order to support additional features. One way would have been to define
3683    pcre_exec2() (for example) with extra arguments, but this would not have been
3684    extensible, and would also have required all calls to the original function to
3685    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3686    is used for passing in "extra" data from pcre_study().
3687    
3688    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3689    contains the following fields:
3690    
3691      flags         a bitmap indicating which of the following fields are set
3692      study_data    opaque data from pcre_study()
3693      match_limit   a way of specifying a limit on match() calls for a specific
3694                      call to pcre_exec()
3695      callout_data  data for callouts (see 49 below)
3696    
3697    The flag bits are also defined in pcre.h, and are
3698    
3699      PCRE_EXTRA_STUDY_DATA
3700      PCRE_EXTRA_MATCH_LIMIT
3701      PCRE_EXTRA_CALLOUT_DATA
3702    
3703    The pcre_study() function now returns one of these new pcre_extra blocks, with
3704    the actual study data pointed to by the study_data field, and the
3705    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3706    before. That is, this change is entirely upwards-compatible and requires n